Taperssection.com
Gear / Technical Help => Recording Gear => Topic started by: dallman on September 30, 2019, 02:48:31 PM
-
This started as a recording thread but that has been moved to a new thread, and I leave this for those that wish to debate defend complain or whatever about their gear or the 32bit float process.
I thought with a few new decks having 32bit float recording and more coming, it might be nice to have a thread dedicated to recording with this process. There are plenty of threads debating the merits and features of the decks but many of us are interested in focusing on how these decks record in 32bit float and what the results are, so I thought this might be useful.
I used my MixPre6 II on Saturday to record Gov't Mule. I ran hotter than I ever had, and I did lower the levels about 2db after the first song which was in the red the entire time. My goal was to run hot, but not super in the red at all times hot. I ran 4 channels, Schoeps MK41's Hypercardioid, and Superlux S502 Cardioid. Here are my thoughts:
I started recording in 1971. It was really hard, really hard to watch levels run regularly in the red after 48 years of avoiding the red. I did it, but I figured how hot to run will be a gradual process.
The results though were excellent! There was nothing even close to distortion. I opened in Soundforge which is what I like to use nest. I transferred the poly file to a stereo file for the Schoeps and a stereo file for the Superlux. On both files I lowered the gain on the first song 2 db and then the entire file another 2 db. The files looked perfect and sounded great. I see obvious advantages , especially when the levels at the start of the show are much hotter than expected.
The learning curve might just be does anything matter? Is any gain too much? Do I care that I have to lower the gain in post if it sound great? Will it at any point with high hot gain sound compressed? These seem to be the details that we will all be looking at, and there probably will be a few differing opinions just like in all of our recording techniques. However there is much more freedom. After 48 years I do not make too many mistakes and I know my mics and most local rooms, but...this was really cool, I totally look forward to using the deck more. Tonight I will record 6 channels at a DSO show with different mics and while I do not plan to run superhot, will push my levels into the red with confidence.
-
Do you have any screen shots of the 'before & after' during Post Edit?
-
Will it at any point with high hot gain sound compressed?
in absence of using limiters it shouldnt
-
Do you have any screen shots of the 'before & after' during Post Edit?
check this out! this is from
https://www.sounddevices.com/noise-in-32-bit-float/?fbclid=IwAR1gJr5l4Bfo1GMNZXvLMqa_hWC2k1Sf9vk2tZoD7tdcz1K0GTsWcgYpRUY
I opened the one that says "all 3" in iZotope and saved the clipped one by itself. As you can see it's +31db and is full of distortion etc. Lowered the gain to just below clipping and it looks and more importantly sounds perfect! I'm sold.....
-
The only things you need to worry about with the MixPre-II in 32-bit mode are:
1) The microphone itself clipping!
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
-
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
^ This.
A lot of hot modern condensers will clip that, then 32 bit float does nothing for you.
-
i suppose for every recorder there is an optimum gain level that will offer maximum S/N
at some point adding extra gain just to remove it later can only add noise unnecessarily, right?
-
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
^ This.
A lot of hot modern condensers will clip that, then 32 bit float does nothing for you.
explain like I'm an idiot please ;)
-
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
^ This.
A lot of hot modern condensers will clip that, then 32 bit float does nothing for you.
DPA MMP-A says maximum output voltage is >5V rms/ >16V peak but youd prob have to be recording jet engines to get it to output that hot
-
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
^ This.
A lot of hot modern condensers will clip that, then 32 bit float does nothing for you.
explain like I'm an idiot please ;)
If it's clipped....it's clipped.....you can turn it down.....but it's still clipped.....
A TLM103 will exceed that input level anywhere close to a drum kit. As will any Sennheiser MKH mic. A loud show with a lot of subs blowing out the bass will make a lot of mics output that kind of voltage if you are near the stacks. You might still need a pad on the mic or the preamp. Wind noise or vibration will definitely make levels like that, and if it clips it's harder to clean in post. You still have to set levels with an eye on not turning up too much, there's still a ceiling, it's just (confusingly) higher than it was, and past a monitor-able level.
The benefit to me seems to be more from the other end, not having to turn things up a lot in quiet ambient scenarios, especially when you might have unexpected hot transients (thunderstorms?), you can leave it low knowing you have a ton of headroom, without paying a digital penalty for that low level. Recording loud sources that are very even in volume like most people here do, there's not as much to be gained. The guy next to you who starts yelling and exceeds 0dBFS can be turned down if your were recording at -6 or -12 dBFS and he pegs the meter, he might not be distorted. He's a different sort of problem....
-
I guess I just don't understand the specs listed to watch out for. I looked up the tech specs of my Gefell's and don't see the same things. I've never been good at the tech aspect of all this!
-
I guess I just don't understand the specs listed to watch out for. I looked up the tech specs of my Gefell's and don't see the same things. I've never been good at the tech aspect of all this!
I wouldn't even bother looking at theoretical possibilities with mic specs, just know this 32 bit thing is not a free pass to ignore gain, it can't save everything. Most mics give you a max SPL for a 1% distortion, but that doesn't tell you what level they will output in a clear manner.
-
From what I have been able to gather, none of these machines actually use 32-bit floating point converters. They use several 24-bit integer converters at different gain levels and then the data is combined and converted to 32-bit float. So I guess it will depend on the implementation to some extent. Sound Devices patented their method, I think, so it will likely be different in other brand's recorders.
I guess I just don't understand the specs listed to watch out for. I looked up the tech specs of my Gefell's and don't see the same things. I've never been good at the tech aspect of all this!
Mostly mic sensitivity and SPL. See the table at Sengpiel (http://www.sengpielaudio.com/calculator-transferfactor.htm). It won't be easy to do with most mics, but still possible!
The guy next to you who starts yelling and exceeds 0dBFS can be turned down if your were recording at -6 or -12 dBFS and he pegs the meter, he might not be distorted. He's a different sort of problem....
That's what analog limiters are for! You'll have to do something to him in post regardless, right?
-
The guy next to you who starts yelling and exceeds 0dBFS can be turned down if your were recording at -6 or -12 dBFS and he pegs the meter, he might not be distorted. He's a different sort of problem....
That's what analog limiters are for! You'll have to do something to him in post regardless, right?
Yeah, or RX7 type cleanup. If it's splattered with a bunch of distortion harmonics, it's a lot harder to clean.
-
Here are a couple of interesting posts from DSatz on this topic from the F6 thread: http://taperssection.com/index.php?topic=190161.msg2304803#msg2304803 (http://taperssection.com/index.php?topic=190161.msg2304803#msg2304803) and http://taperssection.com/index.php?topic=190161.msg2309381#msg2309381 (http://taperssection.com/index.php?topic=190161.msg2309381#msg2309381).
-
Yeah, or RX7 type cleanup. If it's splattered with a bunch of distortion harmonics, it's a lot harder to clean.
This is why I have been using the limiter on my MixPre-6 first version. I set the levels pretty conservatively, so I tend not to hit them often, but occasionally there is that guy doing the "missile whistle" under the mics...
-
This thread, dedicated to 32bit Float recording in itself but not addressing the details of how it is implemented in specific recorders won't really be of much practical interest to tapers.. other than being useful to dispel some academic misunderstandings about what it can and can't do. In other words it will be mostly academic because what really matters is how its implemented in each specific recorder in question.
That's where the rubber meets the road and where all the current confusion lies!
-
The constraint is not 32bit floating point storage (which is ridiculously immense), but real-world bottle-necks prior to it.
One needn't adjust input trim/gain on a recorder if:
1) The total dynamic range of the signal will fit within the available dynamic range of the storage format..
(check, 32-bit floating point does this easily)
AND
2) The total dynamic range of the signal will fit within the available dynamic range through the recorder's signal path, up to the point where the signal is stored in that format..
(this reflects the most critical part of any specific implementation- the preamp and ADC performance of the specific recorder)
AND
3) The actual upper and lower dynamic range values of the signal fit comfortably within the upper and lower dynamic range limits of the recorder's signal path, up to the point where the signal is stored.
(this is about shifting level of the source if necessary to comfortably fit requirement 2, by making sure the signal does not overlap or exceed either end of the available range)
32-bit float storage solves the first problem. The available storage range of 32-bit float is effectively unlimited for audio purposes. Practically, it will depend on how well the other two things are handled in any particular recorder touting the benefit of 32-bit floating point storage.
The problem is accommodating the total output range of a source to the available input range through the recorder (which has been steadily increasing as tech advances, yet remains limited). Most of the hoopla from manufacturers and reviewers thus far has focused on the "wow factor" of the second part of that statement, while not talking much about the "real world" constraints of the first part. Those of us calling for temperance understand that the first part is really what is going to matter. And that will boil down to a performance measure of each individual recorder which is rarely specified clearly by manufacturers.
-
This thread, dedicated to 32bit Float recording in itself but not addressing the details of how it is implemented in specific recorders won't really be of much practical interest to tapers.. other than being useful to dispel some academic misunderstandings about what it can and can't do. In other words it will be mostly academic because what really matters is how its implemented in each specific recorder in question.
That's where the rubber meets the road and where all the current confusion lies!
True. I think that we will just have to figure it out ourselves. I have an F6 on the way, and will get to put it through its paces very soon. One difference in implementation right off the bat has to do with that max level. The MixPre II units have that incredible +14 dBu max mic input level, whereas on the F6 it's only +4 dBu. I don't record anything that would get my mics near that lower spec, but some others here might need to consider it if they don't have inline pads.
-
^ Yes! Exactly the problem I have currently with the F8.
-
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
^ This.
A lot of hot modern condensers will clip that, then 32 bit float does nothing for you.
That's never been an issue for me. I'm recording with the same exact gear that worked at 24 bit, I can just run much hotter without fear. I have never had any of my mics distort and the only variable for me is recording at 32bit float instead of 24bit.
I'm recording right now a little hotter than I did at Mule. I'll try to post some pictures tomorrow.
-
I am pretty sure DSD recording allows for this same behavior as well just not as much of it - The Korg MR-1 could go 2 db over 0 before it clipped - you would just need to lower the levels after the recording to remove the clipping above 0db.
-
Considering the SoundDevices MixPre-II implementation in this light-
Requirement 1) The total dynamic range of the signal will fit within the available dynamic range of the storage format..
(check, 32-bit floating point does this easily)
Requirement 2) The total dynamic range of the signal will fit within the available dynamic range through the recorder's signal path, up to the point where the signal is stored in that format..
Paul reports a total dynamic range through the SoundDevices MixpreII's of 142dB (difference between Mic EIN @ -130dBV and maximum mic input level before clipping @ +12dBV). This is very impressive and reflects the clever design by SD using parallel preamps and ADC's stages of more limited range combined with a way of shifting between their outputs in real time. That will also shift the noise floor simultaneously, so technically the recorder cannot represent the full 142dB range at any particular moment, instead it automatically adjusts constantly to provide that overall range in a way which is presumably audibly transparent (and I expect it is). More on this in panatrope's post at GS. (https://www.gearslutz.com/board/showpost.php?p=14190912&postcount=526) The noise-floor shifting thing is probably not consequential for music recording - we don't hear or care about the noise floor of current recorders when signal is nearing 0dBFS - but it would be consequential if this were a data recorder rather than an audio recorder, which actually needed that full range at any moment. Anyway, this represents more than sufficient dynamic range for me.
Requirement 3) The actual upper and lower dynamic range values of the signal fit comfortably within the upper and lower dynamic range limits of the recorder's signal path, up to the point where the signal is stored.
This is gets to what paul and EmRR posted above:
The only things you need to worry about with the MixPre-II in 32-bit mode are:
1) The microphone itself clipping!
2) Not exceeding the +12dBv (+14dBu or 11 volts p-p) maximum input on the MixPre's mic input. Good luck with that!
^
A lot of hot modern condensers will clip that, then 32 bit float does nothing for you.
One will need to attenuate any input hotter than +12dBv (+14dBu) prior to the recorder. This is probably the biggest real world limitation of this particular implementation.
-
^ Yes! Exactly the problem I have currently with the F8.
Hmm, specs on that say mic input will do +14 dBu but with limiters on. I wonder what it is without limiters that they felt the need to pad that spec. (Ha!)
-
One will need to attenuate any input hotter than +12dBv prior to the recorder.
Do any recorders have metering that accurately shows this? You have to know at what point the meter is tapping the signal also. It could be showing +14, but that could be post-preamp, at the ADC. By then, the input has already been overloaded.
-
Hmm, specs on that say mic input will do +14 dBu but with limiters on. I wonder what it is without limiters that they felt the need to pad that spec. (Ha!)
F8 with Advanced Limiter switched on (implemented post ADC) attenuates input by 10dB prior to the ADC, so presumably +4dBu (that's not enough!)
F8N allows for phantom power on line-input, effectively increasing that to +24dBu with limiters on instead of +14dBu with limiters on (which should be sufficient)
-
One will need to attenuate any input hotter than +12dBv prior to the recorder.
Do any recorders have metering that accurately shows this? You have to know at what point the meter is tapping the signal also. It could be showing +14, but that could be post-preamp, at the ADC. By then, the input has already been overloaded.
Any of them that meter the input stage. Most probably meter the ADC though. Certainly the small hand-helds we use do.
Consider the "brickwalling" problem with many hand-held recorders. We post about the minimum input level setting which can be used with various recorders without clipping distortion occurring from the an overly hot input even though the meters are not indicating clipping. This indicates a design problem within the recorder for which we are finding a work arounds and sharing them with each other. If the recorder was designed correctly, brickwalling could not occur. The recorder's available range of input trim would be set up such that it could not be adjusted to do so. The minimum input-trim setting which could be used without brickwalling would equal "0" on the scale (equating to the maximum available input attenuation).
MixPre-II does appear correctly designed in that way:
This is achieved with multiple ADCs with more than one preamp in front of each one.
The MixPre-II Mic EIN is -130dBV and its maximum mic input level before clipping is +12dBV thus you have a dynamic range of 142dB. The 142dB dynamic range is mapped such that the maximum mic input signal (+12dBV) does not exceed 0dBFS in the multistage ADC circuit. The 32-bit output of the ADC is then converted to 32-bit float.. (https://www.gearslutz.com/board/showpost.php?p=14190646&postcount=523)
-
Here's something that's been nagging me from the start about all 32bit float audio recorders.
Taking the MixPre-II as example- Ignoring the input overload limits, what is important is the wide 142dB dynamic input range without requiring user adjustment. A 24 bit fixed file format can accommodate a 144dB signal which exceeds the throughput of this recorder..
32-bit floating point maybe helpful to accommodate the internal mixing aspects with appropriate processing headroom, but for anyone recording ISO tracks for later mix-down (most tapers), storing the larger 32-bit floating point files is a pointless waste of storage space.. and of bandwidth, which might be a real problem given the recently reported SDcard issues.
We should be able to harness the full power of that 142dB dynamic range when writing 24bit ISO files.
-
so what yore saying is with proper circuit design, the clipping point of the analog and digital stages are all well matched, so 0dB=0dB across the board and users can just set their levels anywhere above -40 dB and still have 100 dB of dynamic range and 40 more dB of headroom
-
A 24 bit fixed file format can accommodate a 144dB signal which exceeds the throughput of this recorder.
I think that 144 dB is a theoretical maximum. In reality, no ADC comes close. As Dan Lavry put it, "There is no such thing as true 24-bit conversion and there won't be in my lifetime." The first iterations of the MixPre, as well as the new Scorpio and 833, for example, specify a dynamic range of 120 dB for their 32-bit ADCs. I suppose that is why the 32-bit float machines need to use multiple ADCs and combine the data with, as jerryfreak memorably termed it, a little "digital fuckery"...
-
Hmm, specs on that say mic input will do +14 dBu but with limiters on. I wonder what it is without limiters that they felt the need to pad that spec. (Ha!)
F8 with Advanced Limiter switched on (implemented post ADC) attenuates input by 10dB prior to the ADC, so presumably +4dBu (that's not enough!)
F8N allows for phantom power on line-input, effectively increasing that to +24dBu with limiters on instead of +14dBu with limiters on (which should be sufficient)
I didn't notice it until you pointed it out, but the F6 does phantom with line-in also, specifying +24 dBu max level but no mention of limiter. I would have never thought to connect a mic directly to a line input. I wonder how the additional 2000 ohms of input impedance changes things on the low end of the signal spectrum. In other words, I question whether running straight to phantom-powered line in really gains you 20 dB dynamic range (probably not). You would only do that for constantly loud stuff.
-
This thread on JWSound gets a bit into the implementation of SD MixPre II vs. Zoom F6:
https://jwsoundgroup.net/index.php?/topic/34205-zoom-f6-vs-sound-devices-mixpre-ii-series-dynamic-range/ (https://jwsoundgroup.net/index.php?/topic/34205-zoom-f6-vs-sound-devices-mixpre-ii-series-dynamic-range/)
-
One question I have is would a 32 bit float recording sound better than a properly captured 24 bit recording with all other things being equal?
My sense is that the answer is no, but that doesn't take away from the benefits of having a larger margin of error in the field.
Dolby B and later Dolby C expanded the margin of error for tapers using compact cassette in the early 1980s. But skill and experience continued to be a requirement to make the best recordings. PCM and later DAT reduced the need to run hot to reduce noise since tape saturation was no longer a factor. 20 bit, and later 24 bit A/D chips that dithered to 16 bit improved the lower level signal quality by reducing the impact of quantization noise, slightly increasing the margin of error setting levels a little more. Still, precision in level setting was a key factor separating the best recordings from the rest, exposing the less skilled and experienced unless they got lucky.
Moving from 16 bit to 24 bit recording offered the improvement of allowing for higher sampling rates, but primarily freed the taper from having to "nail" a recording in 16 bit. The skill level and amount of experience needed plummeted with 24 bit.
From everything I've read, 32 bit float recording is just another step in this evolution, making it almost stupid easy to make a good recording in the field. The notion of "set it and forget it" is realized to the point that only operator errors like forgetting to hit the record button can get in the way.
With all of these advancements, isn't it ironic that there were far more active tapers back when it was harder and required more skill to get a worthy result than are active now?
-
Hmm, specs on that say mic input will do +14 dBu but with limiters on. I wonder what it is without limiters that they felt the need to pad that spec. (Ha!)
F8 with Advanced Limiter switched on (implemented post ADC) attenuates input by 10dB prior to the ADC, so presumably +4dBu (that's not enough!)
F8N allows for phantom power on line-input, effectively increasing that to +24dBu with limiters on instead of +14dBu with limiters on (which should be sufficient)
I didn't notice it until you pointed it out, but the F6 does phantom with line-in also, specifying +24 dBu max level but no mention of limiter. I would have never thought to connect a mic directly to a line input. I wonder how the additional 2000 ohms of input impedance changes things on the low end of the signal spectrum. In other words, I question whether running straight to phantom-powered line in really gains you 20 dB dynamic range (probably not). You would only do that for constantly loud stuff.
It does not gain you 20dB unless the limiter is on, and it's not entirely clear what's happening with the limiter. As I reported in the F8 thread, you can overdrive the F8n line input (set to lowest gain) into square waves with it metering -5.5dBFS. If you get a board feed, you may clip the input without an inline pad; I had a board feed drive it into clipping and ran the limiter just last week). I frequently run MKH mics at shows with it set to line rather than mic, lowest mic gain setting not low enough. I don't run the limiter, I run gain low enough to avoid it.
-
Well I have had the night to cool down and I am glad I did not post last night. I support everybody's right to an opinion and their ability to share it, but this thread was to be all about recording with 32bit float. It is new, it is different, the approach is different, and it would be great to have a place here to discuss as we use and learn this new feature on a few existing decks and more on the way. That is what I asked for in the thread, so I was a bit (or more) annoyed to see that after a nice start, the thread was totally hijacked with the same "stuff" that to me has been stated and restated. There are plenty of long rambling threads where people have already decided what they like or do not like about 32bit recordings or the company's that so far are selling them. I have no interest in that debate and I know I am not alone. I had a MixPre6 and loved it, but when the version II was announced, I sold that deck to a very good friend and decided to dive in to the new deck because when you record on the fly in different venues with different mics and different types of music to me what could be bad about not needing to worry about your gain settings? Again, everyone has a right to their opinions, but jeez, there are pages and pages of whining about cards and talking about why 32bit is no good. I just have no interest in thumbing through pages and pages of that to learn what those who are using 32bit are doing or learning or discovering. I just want a spot for that to be the focus as I think it is a pretty big deal. I have been doing audience recordings for 48 years in every conceivable form and deck in every conceivable format and I just only rely on my ears. I love the technical stuff, but audience recording has its own set of variables and rules. The only thing that never changes is that everything changes, and this concept, 32bit float on a portable deck is pretty amazing so far in my 2 outings. Whether I run super-hot or barely a bit hotter, I feel we are onto something here and those who do embrace it or those who want to learn more about how it is working in the field deserve one thread that talks about it. So, I hope that we can let the next thread not be the thread of anything more than what was originally stated.
I orginally locked this, but all of those that want to discuss what this thread became are welcome to. 8)
-
Thanks dallman,
Apologies for my part in taking this thread in a direction contrary to your intent. Your idea to open another dedicated to hands-on user reports and experience with 32-bit float recorders is a good one. I also think it wise to continue the overview and technical discussion concerning 32bit floating-point recording in this thread, where the technical aspects have been laid out pretty up to this point. The other threads where this has been discussed thus far are more broadly dedicated to specific recorders and other aspects about them. Hopefully keeping this overview and technical discussion about 32bit-float going here will keep it from inevitably spilling over into your new thread. I don't intend to let that happen on my part, but I can easily see it happening otherwise.
~regards
-
A 24 bit fixed file format can accommodate a 144dB signal which exceeds the throughput of this recorder.
I think that 144 dB is a theoretical maximum. In reality, no ADC comes close. As Dan Lavry put it, "There is no such thing as true 24-bit conversion and there won't be in my lifetime." The first iterations of the MixPre, as well as the new Scorpio and 833, for example, specify a dynamic range of 120 dB for their 32-bit ADCs. I suppose that is why the 32-bit float machines need to use multiple ADCs and combine the data with, as jerryfreak memorably termed it, a little "digital fuckery"...
I'm referring to the file storage format written to memory, not the analog to digital conversion. The multiple-ADC switching scheme is a clever work-around of the real-world practical limits imposed by traditional single-ADC conversion which Lavry was referring to, extending DR to 142dB.
My point is that 32-bit float is useful internally for summing and DSP (just as it is in a DAW), yet is unnecessary as a storage format because 24bit fixed provides sufficient storage capacity (just as it does in a DAW). The 142dB of actual throughput will fit within the 144dB of dynamic range storage capacity of a 24bit PCM file. Storing in 24bit fixed files loses nothing useful, reduces the throughput burden of writing data to the SDcard (a currently reported problem), is more universally usable and playable, and significantly reduces file size with high channel counts.
In the current MixPre-II implementation there is some quantization noise way down there due to SD not applying dither in the conversion from 32bit-float to 24bit-fixed (Discussion about that here (https://taperssection.com/index.php?topic=191640.msg2310715#msg2310715)). Paul mentions SD has up to this point not seen the need to do so, which indicates how insignificant the problem is (even though they are now highlighting it as example of superiority for 32-bit float recording - see here (https://www.sounddevices.com/low-signal-32-bit-float/) - erroneously in my opinion, for the reasons stated in the link above). If SD were to apply dither in the conversion from 32bit-float to 24bit-fixed there would be no quantization noise, yet the noise floor would increase by a few dB. So maybe we'd get ~140dB total range or something instead of 142dB, which is still more than enough. And if someone really needed to amplify so much that the noise-floor of the recorder became evident (and it isn't completely buried deeply beneath microphone self-noise and the environmental noise-floor, which is almost certain) they'd find a dither noise-floor rather than quantization artifacts.
For those reasons, I'd like to see Sound Devices implement dithering to 24-bit in combination with the new 32bit-floating point conversion, in addition to direct 32bit floating point storage.
[edit- SD page link fixed, thanks voltronic]
-
Gutbucket - your embedded link to SD's site is broken:
https://www.sounddevices.com/low-signal-32-bit-float/ (https://www.sounddevices.com/low-signal-32-bit-float/)
Here is another relevant JWSound thread:
https://jwsoundgroup.net/index.php?/topic/34174-floating-point-v-fixed-point-wav-files/ (https://jwsoundgroup.net/index.php?/topic/34174-floating-point-v-fixed-point-wav-files/)
-
One question I have is would a 32 bit float recording sound better than a properly captured 24 bit recording with all other things being equal?
not for concert recording purposes. same for 16 bit. if you are careful with levels and make a recording that peaks near zero your room noise is still 30-40 dB above the noise limit of the digital data
-
On the F8/F8N overload point and Advanced limiter-
It does not gain you 20dB unless the limiter is on, and it's not entirely clear what's happening with the limiter.
Not sure about the 20dB figure. I understand the Advance limiter to reduce sensitivity by 10dB, providing that much additional headroom before clipping should otherwise occur, but that 10dB includes the action of the limiter and presumably increases the noise-floor by the same 10dB. Threshold is not set directly nor clearly defined, with only the not to exceed level being set by the user. I've speculated that the Advanced limiter increases ratio as the signal rises higher above the threshold, reaching the stated infinity:1 ratio just below the do not exceed setting - basically a 10dB soft knee. It doesn't sound like inf:1 when lightly engaged. It has worked to keeep the F8 out of clipping for me, but I'd rather it be just a safety measure rather than the only way I can avoid overload when using high-sensitivity mics - and that's basically how I'm currently running it.
Zoom rep states- When the limiter of the F8/F8n is activated for a channel it automatically cuts 10 dB of gain from the analog preamp. That 10 dB is automatically added back after the AD converter. This is a way we build an extra 10 dB of headroom to avoid clipping over the already high dynamic range (A/D Converter has 120 dB dynamic range, with limiter effective dynamic range is 130 dB)
As I reported in the F8 thread, you can overdrive the F8n line input (set to lowest gain) into square waves with it metering -5.5dBFS. If you get a board feed, you may clip the input without an inline pad; I had a board feed drive it into clipping and ran the limiter just last week). I frequently run MKH mics at shows with it set to line rather than mic, lowest mic gain setting not low enough. I don't run the limiter, I run gain low enough to avoid it.
Hoping I can do the same as you to avoid clipping if I switch to F8N by using line-in and the input sensitivity turned all the way down. What is the sensitivity of the MKH mics so that I might compare to better confirm this? I'd forgotten the -5.5dBFS overload point behavior you found, which doesn't inspire confidence. Are you able to keep your recordings using MKH > line-in lowest gain > no limiter peaking below -5.5dBFS?
-
I have not had an F8n problem with MKH into line input, definitely too hot for mic input in many cases with a lot of what I do. 4060 not far behind it.
Senn. MKH 800 TWIN 40mV/Pa = -28dBV
Senn. MKH 20/30/40 25mV/Pa = -32dBV
Neumann TLM 103 23mV/Pa = –32.5 dBV - they quote a max output voltage +13dBu
DPA 4060 20mV/Pa = -34dBV
Neumann KM184 15 mV/Pa - they quote a max output voltage +10dBu
Neumann KM131 12mV/Pa = -38.4dBV
-
I have not had an F8n problem with MKH into line input, definitely too hot for mic input.
That's good to hear that the phantom+line is a useable option for really hot mics. I've never had a device with that capability, so have never tried it. I'll have to borrow some hot mics to try it on my F6.
It's worth noting that the Zoom recorders have lower input impedance than the SD MixPres. The F8n is 2 kΩ mic; 2.6 kΩ line. F6 is a bit higher at 3 kΩ mic; 5 kΩ line. MixPre II is 4 kΩ regardless of setting. That may not be the whole story when it comes to how hot a level it will deal with, but it's part of it.
-
From everything I've read, 32 bit float recording is just another step in this evolution, making it almost stupid easy to make a good recording in the field. The notion of "set it and forget it" is realized to the point that only operator errors like forgetting to hit the record button can get in the way.
Perhaps I'm at risk of picking on something you didn't quite mean the way you worded it, but there's a danger all the way through this discussion that we're saying that a recording which does not suffer from clipping at one end or noise at the other is a good recording.
In recent years, I doubt whether any appraisal of the real merits of a recording by any competent recordist will have included comments about poor level setting. A good recording in the field is still, and always will be, dictated first by the choice of microphones and their placement relative to the sound source. Next, these days, comes the choice of preamp, and last in importance is the choice of recording device, and the format used to record in (bits and sample rates etc). That's not to say it is of no consequence, and of course all items in the signal chain have to be matched at a technical level and at an overall quality level, but it's dangerous to risk any statement that devalues the fundamental skill that separates a competent field recordist from a beginner, that of mic selection and placement. We have to be careful that we don't make statements about recordings that are the audio equivalent of "that's a great photo, you must have a fantastic camera".
From what I'm reading above, choice and placement of mic is even more important with these devices. This now becomes the equivalent of level setting - which has been moved from a knob to the mic, and if you've got it wrong your amazing new recorder can't help you. As always, put a hot mic close to a loud source and you'll still risk clipping - or an unhot mic too far from a low level source, and you'll still risk noise (apart from all the other considerations affecting mic placement).
However, I do agree that we're seeing an important new phase in the development of audio recorders - we started with 14 bits (effectively, as I recall it), then 16, then 24, and now 32 bit float, and then... perhaps we're at the end of the road.
-
From everything I've read, 32 bit float recording is just another step in this evolution, making it almost stupid easy to make a good recording in the field. The notion of "set it and forget it" is realized to the point that only operator errors like forgetting to hit the record button can get in the way.
Perhaps I'm at risk of picking on something you didn't quite mean the way you worded it, but there's a danger all the way through this discussion that we're saying that a recording which does not suffer from clipping at one end or noise at the other is a good recording.
In recent years, I doubt whether any appraisal of the real merits of a recording by any competent recordist will have included comments about poor level setting. A good recording in the field is still, and always will be, dictated first by the choice of microphones and their placement relative to the sound source. Next, these days, comes the choice of preamp, and last in importance is the choice of recording device, and the format used to record in (bits and sample rates etc). That's not to say it is of no consequence, and of course all items in the signal chain have to be matched at a technical level and at an overall quality level, but it's dangerous to risk any statement that devalues the fundamental skill that separates a competent field recordist from a beginner, that of mic selection and placement. We have to be careful that we don't make statements about recordings that are the audio equivalent of "that's a great photo, you must have a fantastic camera".
From what I'm reading above, choice and placement of mic is even more important with these devices. This now becomes the equivalent of level setting - which has been moved from a knob to the mic, and if you've got it wrong your amazing new recorder can't help you. As always, put a hot mic close to a loud source and you'll still risk clipping - or an unhot mic too far from a low level source, and you'll still risk noise (apart from all the other considerations affecting mic placement).
However, I do agree that we're seeing an important new phase in the development of audio recorders - we started with 14 bits (effectively, as I recall it), then 16, then 24, and now 32 bit float, and then... perhaps we're at the end of the road.
Sorry, I thought when I said in my semi-rhetorical question at the top of my post "with all other things being equal" that removed those obvious factors from the recording device equation. Of course location, source quality, and mic placement come before the recorder in order of importance. I've actually made the point about skill and experience being more important than gear cost several times over the years on this forum. And while it's true that a naturally gifted or massively practiced golfer can beat a guy with the best clubs made using an old crappy set due to more advanced skill, I'm starting to pull back on that notion somewhat with photography and audio recording with the advancements in technology.
Take photography, when I started out using a 35mm SLR in 1980, it had a manual light meter called "match needle", and was significant because before they were around, light meters were separate hand held devices that photographers had to know how to interpret in order to set f-stop and shutter speed using one. You had to know how things like depth of field were directly related to aperture for portrait work, for example, when you couldn't preview it. And while advancements in camera technology can't improve poor composition and balance, something only skill and training can do, many of the aspects of cameras that also had to be mastered no longer require the same long learning curve. Just like with analog tape saturation, film exposure to maximize the amount of silver left on the negative was very important to understand in the past, but no longer applies in the digital era. It's like knowing how to navigate a boat with paper charts and dead reckoning turned into a quaint ability no longer needed once GPS technology became well established. I guess the point of my remarks was that 32 bit float removes one more skill set from the equation of requirements for a good recording, knowing how to set proper levels to maximize the performance of the recorder.
It means that skills around factors other than location and mic placement are being eliminated, just like portrait mode on modern phone cameras eliminates the need to understand why the subject is in focus and the background is out of focus when it's used, but they do need to know when to use the setting for composition, which absolutely reduces the learning curve just like automatic exposure meters in 35mm did a couple decades earlier. Hope that clarifies my comments from before.
-
A 24 bit fixed file format can accommodate a 144dB signal which exceeds the throughput of this recorder.
I think that 144 dB is a theoretical maximum. In reality, no ADC comes close. As Dan Lavry put it, "There is no such thing as true 24-bit conversion and there won't be in my lifetime." The first iterations of the MixPre, as well as the new Scorpio and 833, for example, specify a dynamic range of 120 dB for their 32-bit ADCs. I suppose that is why the 32-bit float machines need to use multiple ADCs and combine the data with, as jerryfreak memorably termed it, a little "digital fuckery"...
I'm referring to the file storage format written to memory, not the analog to digital conversion.
Gotcha. But further to the conversion itself, I think it is kind of curious that SD doesn't offer 32-bit float on their new pro-level recorders (the Scorpio and 833). Perhaps they were already too far down the development/testing pipeline to add that feature and there will be II versions of those in the near future as well. Or maybe it is not in demand by professional customers. Or some other reason. In any event, I would be interested in hearing why this highly-touted feature isn't on those recorders...
-
It's like knowing how to navigate a boat with paper charts and dead reckoning turned into a quaint ability no longer needed once GPS technology became well established.
You would have to be an idiot as a sailor (or pilot) not to learn old-fashioned navigation skills and have the correct charts and gear around in case your GPS fails. The sea (and sky) are unforgiving masters. Definitely some worse potential outcomes than a clipped recording...
-
one thing to consider, from a marketing perspective, is that its a relatively easy-to-implement feature that can be hyped
it costs a lot of money to improve analog front ends significantly
it likely costs a lot less to use multiple (relatively) cheap ADC chips in parallel
even the $200 sony A10 uses dual ADCs for higher dynamic range
realistically while 32bit float may save a recording or two, for most people who are familiar enough with the their gear to get levels within 20 dB of where they should be, it actually offers zero improvement to the end product
-
Gotcha. But further to the conversion itself, I think it is kind of curious that SD doesn't offer 32-bit float on their new pro-level recorders (the Scorpio and 833). Perhaps they were already too far down the development/testing pipeline to add that feature and there will be II versions of those in the near future as well. Or maybe it is not in demand by professional customers. Or some other reason. In any event, I would be interested in hearing why this highly-touted feature isn't on those recorders...
I thought I'd read somewhere that 32-bit float was going to be available on the 833 with a future firmware upgrade.
-
^ Interesting. Do you recall where you saw that? On SD's FAQ, it says, "The ability to record in 32-bit float is a future possibility." Whether or not that can be done in firmware was left unsaid (i.e. are all of the necessary ADCs already in there)...
-
^ Interesting. Do you recall where you saw that? On SD's FAQ, it says, "The ability to record in 32-bit float is a future possibility." Whether or not that can be done in firmware was left unsaid (i.e. are all of the necessary ADCs already in there)...
It's annoying me because I'm sure I've read it somewhere, but can't find it now. :shrug:
-
It's like knowing how to navigate a boat with paper charts and dead reckoning turned into a quaint ability no longer needed once GPS technology became well established.
You would have to be an idiot as a sailor (or pilot) not to learn old-fashioned navigation skills and have the correct charts and gear around in case your GPS fails. The sea (and sky) are unforgiving masters. Definitely some worse potential outcomes than a clipped recording...
I wasn't talking about celestial navigation, I was talking about Dead Reckoning which is all worked out on the chart itself, and is absolutely an obsolete skill. It was dropped from the USCG Master and Pilot commercial license requirements around 2014, about a year after NOAA issued the last printed navigational charts. I don't know about where you are, but if you relied on a paper chart last updated in 2014 sailing near the Eastern Shore or at the mouth of the York or Rappahanock Rivers, you would run hard aground in no time.
-
^ I didn't say "dead reckoning" specifically; I said "old-fashioned navigation skills and have the correct charts and gear". In aviation, it is required to have the right charts and a magnetic compass and both pilotage and dead reckoning are a part of the tested body of knowledge. I am less familiar with sailing, but I am pretty sure a magnetic compass and charts are required for anything more than a few miles from the coast almost everywhere. Also, it is kind of disingenuous to say that charts aren't printed anymore, as both aeronautical and nautical charts are constantly updated and can be obtained electronically and used either on an electronic device or printed and, at least in the case of FAA charts, can be obtained on paper from "approved print providers". This is pretty off-topic, though, so if you would like to further discuss the wisdom of relying solely on GPS without some old-school redundancy, feel free to PM me...
By the way, I say this as someone who has hundreds of hours of pilot in command time and has private, commercial, instrument, multi-engine and seaplane ratings.
-
^ Interesting. Do you recall where you saw that? On SD's FAQ, it says, "The ability to record in 32-bit float is a future possibility." Whether or not that can be done in firmware was left unsaid (i.e. are all of the necessary ADCs already in there)...
It's annoying me because I'm sure I've read it somewhere, but can't find it now. :shrug:
Looking at the spec sheets for the Scorpio and the 833, both state that they have 32-bit D/A converters, but only record in 16 and 24 bit depths. 32-bit A/D converters have been out and implemented for a bit now, but it's only very recently we're seeing portable recorders doing 32-bit floating point recording. So the question is: do these top-line SD units have multiple A/D converters in place already?
-
I believe any 32-bit A/D must use multiple overlapping converters to achieve that level of performance.
-
I believe any 32-bit A/D must use multiple overlapping converters to achieve that level of performance.
Exactly my point. Having a 32-bit A/D, but only one of them, means you'll never get 32-bit floating point recording. I mean, I guess you could, but it wouldn't be worth the trouble.
These 32-bit A/D chips have been out for a while now, but it's only these very new MixPre II units and the Zoom F6 that are using them with the multiple-overlapping structure.
There was a poster on GS who pointed out that the Stagetec has been around for a long time doing gain-ranging A/D. Their Truematch technology in their current converters have been out for a few years now, and they do 32 bits with a 158 dB dynamic range. I do not know if they ever integrated this with a recording media system though - just outboard converters, consoles, and routers.
-
The Scorpio and 833 have the hardware architecture to support 32-bit float in the future. This includes >1 ADC
As pointed out by someone earlier, use of more than 1 ADC to achieve wider dynamic range is not a new idea - its been around for decades.
We patented our method of multistage ADC because it is a new unique approach which we believe greatly improves handling the transition between ADCs compared to other implementations. If you want to know more about it you can read our patent online. Warning - its complicated and much of the math goes over my head so don't expect further insights from me.
-
The Scorpio and 833 have the hardware architecture to support 32-bit float in the future. This includes >1 ADC
As pointed out by someone earlier, use of more than 1 ADC to achieve wider dynamic range is not a new idea - its been around for decades.
We patented our method of multistage ADC because it is a new unique approach which we believe greatly improves handling the transition between ADCs compared to other implementations. If you want to know more about it you can read our patent online. Warning - its complicated and much of the math goes over my head so don't expect further insights from me.
Thanks for that info, Paul. I'm sure people who bought Scorpios just before the MixPre II came out are reassured to know this.
I did try to read the patent a couple weeks ago, but it's way above my comprehension level.
-
Genuine question - in case anyone can be bother to answer what may be a silly one - what is the difference between using one of these new recorders, compared with recording in 32 bit float on a laptop (with a decent rig in front of it)? Is the breakthrough that it's all in one box? Was I actually doing this years ago in Audition?
-
The routing bus and capture are 32 but the converters are 24. My Motu 16A shows up as 32 but that’s just the bus depth.
-
Genuine question - in case anyone can be bother to answer what may be a silly one - what is the difference between using one of these new recorders, compared with recording in 32 bit float on a laptop (with a decent rig in front of it)? Is the breakthrough that it's all in one box? Was I actually doing this years ago in Audition?
The constraint is not 32bit floating point storage, but real-world bottle-necks prior to it.
One needn't adjust input trim/gain on a recorder if:
1) The total dynamic range of the signal will fit within the available dynamic range of the storage format..
AND
2) The total dynamic range of the signal will fit within the available dynamic range through the recorder's signal path, up to the point where the signal is stored in that format..
AND
3) The actual upper and lower dynamic range values of the signal fit comfortably within the upper and lower dynamic range limits of the recorder's signal path, up to the point where the signal is stored.
The laptop is handling part 1
The rest of the rig is responsible for parts 2 & 3
-
Any info how 32Bit Float will work with compressors and software limiters in real life?
If your source have a lot of really loud splashes and you bring back loudest part to visible range in one click, other parts of track gain became too low. Simple normalize tool for selected regions only or manual gain adjust with curve is ok for simple situations, but for instant random loud splaches you need some kind of automatic limiter. So will those software plugins see that hidden dynamic range in 32Bit mode? Is it possible that software limiters in 32Bit mode will produce same result as analogue limiters or even better?
UPDATE: I downloaded samples from SoundDevices website and play with very simple AU limiter plugin in Cocos Reaper. All you need do is set attack/release time. Seems works great. Artifacts-free real time recovery! :headphones:
-
When a recording on the unit is going well into the red, what happens to the monitoring? Do you hear distortion that won't be heard after normalising, or... how does it work? I guess it's down to where in the chain the monitor signal comes from.
-
When a recording on the unit is going well into the red, what happens to the monitoring? Do you hear distortion that won't be heard after normalising, or... how does it work? I guess it's down to where in the chain the monitor signal comes from.
My guess you will hear distortion that won't be heard after normalising. MixPre limiters are disabled in 32 bit mode.
Here are some processed examples (original, normalized only, non normalized with software limiter plugin):
https://www.dropbox.com/sh/rm6c9sdgimeoy98/AACXdJ3tpd5Mi5QJYyO7gKG7a?dl=0
It is also interesting what strategy will work better for sound "highlights" recovery?
a. normalize all track and then use compressor to boost silent parts withiout clipping loudest parts.
b. don't normalize track and recover hidden loudest parts with limiter.
-
DSatz, Gutbucket, and a few others raised some excellent points about tempering expectations regarding 32-bit float recording. This is valuable information we all need to pay attention to, but as has been said before: the real test is in the implementation.
To that end, I would like to hear from other 32-bit float recorder owners if they have run into anything of concern.
So far, recordings from my Zoom F6 have been flawless. These recordings have swung very wide dynamic ranges, and I do not hear any noisefloor modulation, artifacts, or anything else that shouldn't be there. If there is anything nasty going on, it is buried way down in a level that I cannot hear it, nor can my spectral analysis show it.
-
DSatz, Gutbucket, and a few others raised some excellent points about tempering expectations regarding 32-bit float recording. This is valuable information we all need to pay attention to, but as has been said before: the real test is in the implementation.
To that end, I would like to hear from other 32-bit float recorder owners if they have run into anything of concern.
So far, recordings from my Zoom F6 have been flawless. The recordings I have made have swung very wide dynamic ranges, and I do not hear any noisefloor modulation, artifacts, or anything else that shouldn't be there. If there is anything nasty going on, it is buried way down in a level that I cannot hear it, nor can my spectral analysis cannot show it.
I agree, my recordings have sounded excellent and very much like the 24bit recordings I have been making for years. I have played with running hot and I have played with recording conservatively and it does not seem to make a difference. I have seen very positive tests where things have really been pushed but I am not looking to make a point when I record, I am looking to grab as good a recording as I can, so I am not pushing into anything crazy. I have used both the F6 and the MixPre6II and I am happy with both decks performance and sound.
-
Excellent feedback, thanks!
Good to hear if I had kept my order with Zoom that all would end well. But I switched on the delay because I couldn’t be sure it would be ready for my late September and October recordings where I used the MixPre-6 II successfully.
-
got into a discussion on another board where it was asserted that 32-bit float offers similar 'precision' to 24-bit. their discussion was in regard to DSP.
heres the bulk of my post. Whycome i got what look like rounding errors if the container can represent 1500 dB of data to the sample?
----------------------------------------------------------------------------------------------------------------------------------------
im not really mathematically equipped to discuss it at that level
though it seems they are talking about precision of calculation of 32-bit float in DSP
i could kind of see this in a way. for a given mantissa value, there is a finite 'precision' those numbers can represent. and then the exponent in a way 'slides' that scale of representable numbers up and down, but it doesnt go any 'deeper' or more precise in calculations
seems like almost two different discussions. (DSP vs representing large differences in headroom)
im not super concerned about how the LSB in a DSP is handled, as much as i am about not clipping my signal by running out of data 'container capacity' for a better recording.
i did a test in soundforge. i generated a 1 sec 1khz sine at FSD and saved that file
then i used 'volume' effect in soundforge to adjust the volume by -120.00dB. saved that file
then i used volume again to adjust the volume +220.00 dB, which took it to an expected +100 dB signal on playback , saved that file
then i used volume to reduce it exactly -100.00 dB which put it at an expected 0dB peak
when i inverted this over the original 0dB generated file, it canceled it.... almost. there is some resulting random noise with peaks at -135 dB
so it can be argued that i indeed represented this waveform at ~24 bit precision at volume levels ranging over 200+ dB difference. however at the end of the day, the DSP precision of that file after multiple computations was indeed right around 24-bit.
could be multiple accumulated rounding errors at the 32-bit noise floor summing up to make some random low-level noise?
______________________________________________________________________________________________________
responses from that forum:
Right, it's about what you could call instantaneous dynamic range. 24 bit signed integers effectively have 23 bits plus sign but any integer gain adjustment could either clip or throw away bits. Each operation in 32 bit floating point essentially normalizes the result to keep 23 bits plus sign in the result. There numerical noise issues when doing a large number of re-scales or effects but these are usually minor.
and another response, emphasis mine
This interactive demo should be self-explanatory:
https://www.h-schmidt.net/FloatConverter/IEEE754.html
An implied bit when exponent bits are non-zero (1), (23) bits of mantissa and (1) sign bit.
The benefit of floating point is it has a near constant precision at different scales, that makes digital processing and gain staging much safer and consistent.
-
jerryfreak, nice experiment. Thank you for it.
I don't agree with the last part of the final statement (in boldface); it even seems propagandistic to me--an attempt to make the shifting digital noise floor seem like a virtue when it isn't one. I don't mean that it's necessarily a defect, either, because if it's low enough at all times, no one will hear it shifting. But "it successfully evades detection" is the best that can be said about it if so.
--best regards
-
DSatz, Gutbucket, and a few others raised some excellent points about tempering expectations regarding 32-bit float recording. This is valuable information we all need to pay attention to, but as has been said before: the real test is in the implementation.
To that end, I would like to hear from other 32-bit float recorder owners if they have run into anything of concern.
So far, recordings from my Zoom F6 have been flawless. These recordings have swung very wide dynamic ranges, and I do not hear any noisefloor modulation, artifacts, or anything else that shouldn't be there. If there is anything nasty going on, it is buried way down in a level that I cannot hear it, nor can my spectral analysis show it.
Bumping this thread for anyone considering a 32-bit float dual-ADC recorder...
6 months in, and I can report zero glitches to date with my Zoom F6 running in this mode. While I can't speak for the specifics of the exact implementation on other units, the input gain on the F6 is fixed in this mode. I tend to turn up the faders a bit just so I can see levels popping up during setup, +20 to +40 depending on which mics I'm using. Keep in mind that this is only affecting post-ADC fader level as the gain is fixed. This means that the resulting file may have areas significantly beyond 0 dBFS, but in this format you can just pull it down and post; there really isn't any clipping.
For me, I am never going back to 24-point fixed unless another 32-bit float dual-ADC device I am using down the road doesn't work as well as this one does. I have posted on some other threads that I am often both performer and recordist for my concerts, so I cannot monitor my recordings which have wide dynamic ranges, and peak levels can be unpredictable. Because of this, this new implementation truly is a game-changer for me, because I no longer have to rely on safety tracks and can use all of my inputs without have to think about levels.
Yes, one still needs to be careful not to overload the input stage, as has been discussed. I don't own very sensitive mics so it's a non-issue for me. If it is a concern for you, the Zoom F6 has a Line + Phantom mode which has a much higher max input level, or you can go with the Sound Devices MixPre-II series which has a much higher max level in all recording formats.
I am eagerly awaiting to see how the Tentacle Track E recorders perform once released.
-
ive come across some good recent posts from other forums with respect to bit-depth and fixed vs float, that i thought i would share here
It appears the most state of the art gear uses 32-bit fixed, not float
here is an example of a modern dual-ranging ADC, the Merging ADA8S, with some astonishing specs, using its 'dual gain' topology. it literally is utilizing close to 24bits of resolution through its entire signal path which would put it quite a bit above the specs of the 32-bit float recorders we are discussing here
https://youtu.be/IMkUcWr2-cg?t=629
heres a direct demo of a sine wave recorded peaking at -110dBU and boosted 80dB in post. the new ADA8S is the one on the left and is pretty astonishing:
https://www.merging.com/products/interfaces
some comments ive culled on fixed vs float re:this gear and its implementation
audiosciencereview:
Basically the floating point approach is just an array of ADCs (or a multichannel ADC) recording at different levels. The old Sony PCM-D100 used a similar approach -- recording two copies at the same time, but with a difference of 12dB. Of course, it only supports 24-bit integer and DSD.
While 32-bit float "only" has 25 bits of precision, don't forget it wins 32-bit integer when sample values (yes, sample values, not the overall level of the whole file) are below -42.1442dBFS (7 bits below full scale) and above 0dBFS (integers will clip). Floating point on the other hand always store 25 bits of data within the ~1500dB range. Since no combination of ADC, mic and preamp can achieve 25 bits of instantaneous dynamic range, the floating point approach is much more sensible.
from Gearslutz, someone asked Merging tech support the following question
Will the new ADA8 also be 32-Bit like Anubis? And speaking of 32-Bit conversion, Anubis specs doesn't say 32-Bit float, so it's fixed correct? And if that's the case, what made you choose 32-Bit fixed instead of float, I'm not saying one is better than the other, I'm just curious as to why you made the decision to use 32-Bit Fixed, and not Float?
response:
In terms of raw resolution, 32 bit fixed point is much better than 32 bit floating point (which only gives the equivalent of 25 bit resolution).
In other words, if converting a 32 bit fixed point signal to a 32 bit float signal and then back again to 32 bit fixed point, one would lose the 7 least significant bits from the original 32 fixed point number.
In terms of Dynamic range however (headroom, noise floor), 32 bit float is of course incomparably better (about 1500 dB) than fixed point (about 186 dB) but in most real situations 1500 dB is uselessly large.
And with respect to using a full 32 bit fixed point path in the Anubis it does make most sense, since the AD chip is able to deliver up to 32 bit fixed data and the DA chip is able to eat up to 32 bit fixed data as well.
So making sure to preserve all those 32 bit during the whole signal processing chain including through the mixing engine offers maximum transparency in the Anubis (and soon in the new ADA8 as well).
at the end of the day, we are stuffing 12-16bits of information from our microphones into a box. at the end of the day it doesnt practically matter if the box you put it in has 24 bits, 25 bits, or 32 bits of space. All that matters is that you get it in the box ;D
as mentioned above when your sample data is below -42dB, 32bit-float offers theoretical advantage. Youd be hard pressed to hear any quantization noise over the noise floor of the room, the mics, and the analog input of your recorder, however.
-------------
here is an example of a modern ADC chip, the 8-channel AK5578
you can see how multi-channels of a chip are typically combined to boost S/N
S/N: 121 dB (8-to-4 mode: 124 dB, 8-to-2 mode: 127 dB, 8-to-1 mode: 130 dB)
https://www.akm.com/content/dam/documents/products/audio/audio-adc/ak5578en/ak5578en-en-datasheet.pdf
-
ive come across some good recent posts from other forums with respect to bit-depth and fixed vs float, that i thought i would share here
It appears the most state of the art gear uses 32-bit fixed, not float
here is an example of a modern dual-ranging ADC, the Merging ADA8S, with some astonishing specs, using its 'dual gain' topology. it literally is utilizing close to 24bits of resolution through its entire signal path which would put it quite a bit above the specs of the 32-bit float recorders we are discussing here
https://youtu.be/IMkUcWr2-cg?t=629
heres a direct demo of a sine wave recorded peaking at -110dBU and boosted 80dB in post. the new ADA8S is the one on the left and is pretty astonishing:
https://www.merging.com/products/interfaces
some comments ive culled on fixed vs float re:this gear and its implementation
audiosciencereview:
Basically the floating point approach is just an array of ADCs (or a multichannel ADC) recording at different levels. The old Sony PCM-D100 used a similar approach -- recording two copies at the same time, but with a difference of 12dB. Of course, it only supports 24-bit integer and DSD.
While 32-bit float "only" has 25 bits of precision, don't forget it wins 32-bit integer when sample values (yes, sample values, not the overall level of the whole file) are below -42.1442dBFS (7 bits below full scale) and above 0dBFS (integers will clip). Floating point on the other hand always store 25 bits of data within the ~1500dB range. Since no combination of ADC, mic and preamp can achieve 25 bits of instantaneous dynamic range, the floating point approach is much more sensible.
from Gearslutz, someone asked Merging tech support the following question
Will the new ADA8 also be 32-Bit like Anubis? And speaking of 32-Bit conversion, Anubis specs doesn't say 32-Bit float, so it's fixed correct? And if that's the case, what made you choose 32-Bit fixed instead of float, I'm not saying one is better than the other, I'm just curious as to why you made the decision to use 32-Bit Fixed, and not Float?
response:
In terms of raw resolution, 32 bit fixed point is much better than 32 bit floating point (which only gives the equivalent of 25 bit resolution).
In other words, if converting a 32 bit fixed point signal to a 32 bit float signal and then back again to 32 bit fixed point, one would lose the 7 least significant bits from the original 32 fixed point number.
In terms of Dynamic range however (headroom, noise floor), 32 bit float is of course incomparably better (about 1500 dB) than fixed point (about 186 dB) but in most real situations 1500 dB is uselessly large.
And with respect to using a full 32 bit fixed point path in the Anubis it does make most sense, since the AD chip is able to deliver up to 32 bit fixed data and the DA chip is able to eat up to 32 bit fixed data as well.
So making sure to preserve all those 32 bit during the whole signal processing chain including through the mixing engine offers maximum transparency in the Anubis (and soon in the new ADA8 as well).
at the end of the day, we are stuffing 12-16bits of information from our microphones into a box. at the end of the day it doesnt practically matter if the box you put it in has 24 bits, 25 bits, or 32 bits of space. All that matters is that you get it in the box ;D
as mentioned above when your sample data is below -42dB, 32bit-float offers theoretical advantage. Youd be hard pressed to hear any quantization noise over the noise floor of the room, the mics, and the analog input of your recorder, however.
-------------
here is an example of a modern ADC chip, the 8-channel AK5578
you can see how multi-channels of a chip are typically combined to boost S/N
S/N: 121 dB (8-to-4 mode: 124 dB, 8-to-2 mode: 127 dB, 8-to-1 mode: 130 dB)
https://www.akm.com/content/dam/documents/products/audio/audio-adc/ak5578en/ak5578en-en-datasheet.pdf
The way I interpret the two bolded statements is that if I'm exporting from my DAW in 24bit, it seems that 32bit float is the best of everything and there might not be a need to ever upgrade a recorder after this.
-
well, it depends on what youre happy with. technically im sure somebody said the same thing when good HQ pre/AD combos like the V3 came out. As easily as digital technology can become dated, 20 years later you can still go make world-class recordings with a V3, in a studio, yet alone a relatively noisy live show environment.
Heck a sax>SBM setup from the 90s was putting down phenomenal dynamic range for our purposes esp with the noise-shaping of the SBM
that said, perhaps you could have highlighted this part as well:
Since no combination of ADC, mic and preamp can achieve 25 bits of instantaneous dynamic range
it seems that the ADC portion is ahead of the rest of the signal chain, for now (forever?). However, unlike previous improvements in digital technology, the 32-bit float architecture only improves the sound when compared with a less-than-ideal setup of lesser bitrate. In other words there is no sonic advantage to a 32-bit file vs a 24-bit file at proper levels. Which depending on setup (background noise, mic noise, amount of gain used, etc) could mean anything where your peaks dont hit -12dBu to where your peaks dont hit -48dBu. Again, its a container for your data, if you are carefully putting your data in the container there will be no difference. if you are missing the container with your data its a huge problem. and an easy fix in this case, with the auto-ranging ADCs which "move the container for you" as needed.
will you want to upgrade someday? sure... when a recorder that sounds as good as your mixpre is the size of your phone. that will be another 20+ years though, and will likely only offer diminishing returns, if any, over your current setup in terms of overall sonic quality for our purposes
next stop is microphone signal path, it seems that paused for a bit with the digital technology, but i would imagine thats what we see next. Every manufacturer has their own way of powering their mics, from the permanently polarized DPAs that take a trivial amount of voltage/current, to the schoeps who have cut the power requirements of their preamps by nearly an order of magnitude in a few decades
couple these advances with digital microphone preamp technology (advantage being less power required to generate a hot signal to compete with noise), as well as the ever evolving ADC and op-amps, and, basically its only going to get better as far as tech.
so i guess to circle back, you asked "there might not be a need to ever upgrade a recorder after this."
was there a need before this? :bigsmile:
-
^ Good summary!
auto-ranging ADCs "move the container for you" as needed
-
Regarding the mic signal path:
Remember AES42 digital mics, like the Neumann D line? For those that don't know, these mics have a 28-bit ADC right in the mic body. The specs on them are amazing, and the few recordings I have heard made with them are extremely transparent sounding. It doesn't seem that standard ever got wide adoption though - I hear Neumann is discontinuing that series. Sound Devices always supported AES42 back through the 7-series.
The silver bullet for audio recording in my mind would be AES42 mics with the type of 32-bit multi-ADC converters we are discussing built right in. Now you have the mic manufacturers doing all of the "container moving" onboard, and you can simplify your recording box to a bit bucket that only has to supply 10V phantom to the mics. Lower noise all around, less chance of interference on your cable runs, lower power requirements, etc.
-
yes schoeps digital version of CMCs (CMD i beleive?) just disappeared as well about 7 years ago
i’m sure they could do it better with today’s tech. those came out in like 07-08 IIRC. that’s forever ago in digital
-
Thoughts on this? I have not heard this in my recordings.
https://youtu.be/y3431uljZ2k?si=I7I7qhsKK7TUo49s
-
Thoughts on this? I have not heard this in my recordings.
https://youtu.be/y3431uljZ2k?si=I7I7qhsKK7TUo49s
Wow. There's more information in the comments, including how Sound Devices (mostly?) fixed this with a firmware update.
Now I'm going to have to dig into some of my F6 recordings and see if I can locate this issue.
What I don't understand is why the low-gain ADC is 30 dB noisier than the high-gain one. I would expect it to be the reverse.
-
It's an unpleasant issue, indeed. Thing is, I am pretty sure that Zoom F serie recorders still have dual ADC working even in 24 bit mode, so no escape unless the manufacturer finds a solution via firmware ( if any solution is possible)
-
Thoughts on this? I have not heard this in my recordings.
https://youtu.be/y3431uljZ2k?si=I7I7qhsKK7TUo49s
Wow. There's more information in the comments, including how Sound Devices (mostly?) fixed this with a firmware update.
Now I'm going to have to dig into some of my F6 recordings and see if I can locate this issue.
What I don't understand is why the low-gain ADC is 30 dB noisier than the high-gain one. I would expect it to be the reverse.
From what I understood, the recorder engages the low-gain ADC when it detects a transient - and it does so before the signal gets strong enough that it has the chance to clip. This means that there's some amount of guessing involved, and it might engage the low-gain ADC for sudden sounds that end up not getting anywhere near clipping level. And it's precisely because it has engaged the low-gain ADC for a sound that was actually fairly quiet that the device needs to make up for it by amplifying the signal digitally, raising the noise floor of the recording. It isn't that the low-gain ADC is noisier than the high-gain one in absolute terms (you're right that this would make no sense): it's that the high-gain one has cleaner analogue gain than simply boosting the digital signal produced by the low-gain one by however much is necessary so that the levels remain consistent throughout the recording.
You might have noticed he was recording tiny sounds (they sounded like water dripping?), not a gunshot or a drum kit. I imagine that these artefacts wouldn't appear in a live music recording in any noticeable way.
Edit: actually, I paid a bit more attention this time, and I don't think it's supposed to detect transients? It just routes the signal based on the voltage the microphone produces. So the situations when the issue would happen are a little different, but the same principle still applies. Basically, say the high-gain ADC is a +30dB one. Any signal strong enough to clip with that amount of gain is automatically routed to the low-gain ADC. If it's a healthy signal, you won't hear any noise. But if it's JUST loud enough to cross the clipping threshold of the high-gain ADC (say, by 1dB), then clearly the low-gain ADC is going to need to be boosted digitally so that the levels are consistent with rest of the recording. You're not going to see an issue when the sound falls into the best operational range of each ADC. You get some noise when a sound is just loud enough that you can't use cleaner gain, but still quiet enough that the new noise floor is noticeable.
-
^ That explanation makes sense. I was thinking of the signal being amplified in the analog domain before it's converted, but that's not how these devices work. I should have realized that because I frequently correct other people on this assumption. Thanks for setting me straight.
-
It's an unpleasant issue, indeed. Thing is, I am pretty sure that Zoom F serie recorders still have dual ADC working even in 24 bit mode, so no escape unless the manufacturer finds a solution via firmware ( if any solution is possible)
Yes, and that was discussed in the comments. One poster mentioned that it's the same deal for the Sony D100.
-
I noticed my Mike Stern F3 recording from last week was really noisy. Almost analog hiss sounding.
-
I noticed my Mike Stern F3 recording from last week was really noisy. Almost analog hiss sounding.
Which mics were you using? How loud was the show?
-
I noticed my Mike Stern F3 recording from last week was really noisy. Almost analog hiss sounding.
Which mics were you using? How loud was the show?
Its this one. I figured it was because of the low pro situation. Mics have never made any similar hiss.
https://taperssection.com/index.php?topic=204608.0
-
It just shouldn't be anywhere near where it would be using the low gain adc if the show wasn't very loud. And that band doesn't strike me as one that would be inordinately loud.
-
It just shouldn't be anywhere near where it would be using the low gain adc if the show wasn't very loud. And that band doesn't strike me as one that would be inordinately loud.
Unfortunately the only reason I got it was to record >:D these particular bands. Think Im going back to mk2e's>BB>A10
-
is the hiss always there? if you can isolate it, RX can remove it automatically
-
Just the quiet parts. It's not terrible, but I might try and fix it up.
-
That is strange. I love my F3s because the preamps are so darn quiet.
-
is the hiss always there? if you can isolate it, RX can remove it automatically
Watch the video. This is addressed. It's only around the louder transients.
-
Just the quiet parts. It's not terrible, but I might try and fix it up.
I've just listened to your recording! It isn't the same issue described in the video. I hear the noise, but it's everywhere during the quiet parts, not just around the transients. It's either the microphone's or the preamp's noise floor, and I don't imagine you'd get a lot less noise with a different recorder unless your F3 is faulty. There could be other weird things going on too - I know for example that I get a lot more noise out of my Church Audio preamp if I put it down on a church pew or my desk. If I keep it on my body, it's fairly quiet (I imagine this has to do with the grounding).
-
This idea is making waves on various recording forums/blogs I read:
https://www.youtube.com/watch?v=y3431uljZ2k
tl;dr, on a 32-bit recorder, the multiple ADC's kick in depending on how loud the source is. The problem is that the different ADC's have vastly different noise floors, so hit a loud transient and your recorder will choose the low-gain ADC, which adds a lot of noise right around the transient.
Now, because the noise is different/worse only around the loud transient, it's also extra hard to profile and remove with software (versus a consistent noise throughout the whole recording).
May not apply to you all. For me, I record sopranos who go into glass-breaking-mode sometimes after a gently piano intro, so this is indeed a problem.
-
so can you post an example of where it happened on one of your recordings?
also: https://taperssection.com/index.php?topic=192100.0;prev_next=next#new
-
Having spent a happy hour reading about the subject (in documents dated prior to this video), it seems to be... complicated. One writer suggested that indeed the noise involved in 32 bit float recordings is variable, not fixed, but it's variable in the area of -144dB so it is unlikely to be audible. I'm too old to hear the noise apparently demonstrated in the video, but I did rather wonder whether the source being used was not a good choice. It sounded like some kind of sound effects file with a lot of constant low frequency background noise, plus artificial reverb (but I could be wrong) - is it the reverb noise that's happening? A much better test would be to simply use a tone generator, surely?
Anyway, I would be much more concerned about distortion (clipping) than noise for real-world examples, and the whole concept of having no level controls and no meters appeals to me immensely. The ability to walk into a classical chamber recital hall and position a 32 bit float recorder/mic at the sweet spot, turn it on and leave it to its own devices uncablednd unmonitored, and go to the bar for the performance, appeals to me immensely. Would the client complain about the subtle variable noise floor? I doubt it very much. Audience coughing and phones ringing are likely to be a greater source of irritation.
But I do concede that ultimate perfection should be sought by ultimate perfectionists
-
So has anyone heard this transient noise in practice in the field?
-
The voltage ranges for the ADCs overlap, at least in the Sound Devices implementation. That means that the sound right around the "switch" is likely to be in more than one post-ADC stream. SD's algorithm, based on regression models, enables rapid shifts from one ADC to another and since there is overlapping content, this can also be done very accurately. I would guess this can be fine-tuned to eliminate or, at least, minimize this problem. Maybe different in Zoom's implementation. I don't know and this is just speculation anyway...
-
Having spent a happy hour reading about the subject (in documents dated prior to this video), it seems to be... complicated. One writer suggested that indeed the noise involved in 32 bit float recordings is variable, not fixed, but it's variable in the area of -144dB so it is unlikely to be audible. I'm too old to hear the noise apparently demonstrated in the video, but I did rather wonder whether the source being used was not a good choice. It sounded like some kind of sound effects file with a lot of constant low frequency background noise, plus artificial reverb (but I could be wrong) - is it the reverb noise that's happening? A much better test would be to simply use a tone generator, surely?
Anyway, I would be much more concerned about distortion (clipping) than noise for real-world examples, and the whole concept of having no level controls and no meters appeals to me immensely. The ability to walk into a classical chamber recital hall and position a 32 bit float recorder/mic at the sweet spot, turn it on and leave it to its own devices uncablednd unmonitored, and go to the bar for the performance, appeals to me immensely. Would the client complain about the subtle variable noise floor? I doubt it very much. Audience coughing and phones ringing are likely to be a greater source of irritation.
But I do concede that ultimate perfection should be sought by ultimate perfectionists
Imho the question is: do you want audible noise around some transients (not all of them, just the ones that are too loud for the high-gain ADC, but a bit too quiet for the low-gain one), or do you want noise everywhere? Without multiple ADCs, if you're expecting the music to have louder passages, the one option you have is to be conservative and use low gain for everything, which means the quiet parts will have a higher relative noise floor. Sure, it's easier to profile the noise in post that way, but I'd much rather have less noise during the quiet parts in the first place. I could hear the noise around the transients, but I personally don't find it distracting at all.
-
Foreshadowing back when..
This thread, dedicated to 32bit Float recording in itself but not addressing the details of how it is implemented in specific recorders won't really be of much practical interest to tapers.. other than being useful to dispel some academic misunderstandings about what it can and can't do. In other words it will be mostly academic because what really matters is how its implemented in each specific recorder in question.
That's where the rubber meets the road and where all the current confusion lies!
jerryfreak, nice experiment. Thank you for it.
I don't agree with the last part of the final statement (in boldface); it even seems propagandistic to me--an attempt to make the shifting digital noise floor seem like a virtue when it isn't one. I don't mean that it's necessarily a defect, either, because if it's low enough at all times, no one will hear it shifting. But "it successfully evades detection" is the best that can be said about it if so.
--best regards
(Bolding above is my emphasis) Of course back then we were speculating on where potential issues could arise. The current hullabaloo about noise-floor windowing artifacts during the hand-off between multiple ADC's is a specific aspect of this. DSatz essentially called out the noise shift issue quite specifically. Going back and reading his comment about successfully evading detection brings to mind a another thought on a somewhat deeper level..
There is a deeper fundamental change going on in this shift to multi-ADC designs, which is a shift away from a recording system that is fully agnostic/isotropic in respect to the recording of whatever content fits within its bandwidth, to one which gives that up in achieving some alternate aspect deemed more valuable for the intended use case, even though doing so introduces problems for less-common uses. If evading detection of the ADC switching artifacts can be successfully arranged for the recordists who value not having to set levels far more than a truly isotropic data set, such a trade off is likely to be a welcome one for them - its good enough for their purposes by definition and they gain a quite welcome new feature, even if the scheme may not be good enough for outlier uses where a more fully isotropic data set is required for more esoteric applications such as recording ultrasonic signals, dramatically pitch shifting content, or whatever.
The machine has become specially optimized for certain uses, at the cost of being less optimized for others. Of course many other features of any recorder are optimized for its intended use, but the digital recording itself previously has not been.. unless recording to a lossy format.
I'm reminded of the development of things like noise-shaped dither, and psychoacousticly tuned lossy data compression. Those are useful tools, the successful implementation of which required careful determination of how much fidelity to the source is needed, prior to sacrificing what lies beyond the limits of perception. Its an optimization for certain use cases, which makes for a tricky game in determining where those limits might be, how they might be different for different uses, and how close one is willing to get to them in seeking to leverage human hearing perception to advantage. Noise shaped dither moves the bulk of the noise to where it is perceptually less obvious. Lossy codecs minimize storage requirements in part by discarding data deemed perceptually irrelevant. The fundamental shift is from a complete data set that contains extraneous information to one that is perceptually equivalent yet not fully complete. Fundamentally this is the same philosophical calculus based upon making a decision about what maters and what doesn't. If well implemented, it's not going to be s problem for most, but may for some. Its a sacrifice of true fidelity for all use cases, for easier use for intended target use cases.
-
Following up:
- How many ADC's do different recorders have (MixPre-3/6 ii, Zoom F3/F6, etc.)?
Presumably, one fix is to just add more ADC's to these little recorders, so instead of a 30dB noise jump on passages where it chooses the lo-gain ADC, you could have 5dB increments of background noise increase. I have heard that some 32bit recorders have 3 or more ADC's, but not sure which ones those would be.
-
The recording in that video that's getting folks talking about this is a recording of a water droplet.
Recordists making Foley or SFX recordings often use 60-70 dB gain to get those types of sounds loud enough to be usable.
I'm not surprised at all that after that type of gain adjustment that you are going to be able to hear system noise. Most people recording music or even dialog are not going to hear the low level interference from the ADC.
I'll keep staging my gain the old fashioned way using 24 bit.
-
Following up:
- How many ADC's do different recorders have (MixPre-3/6 ii, Zoom F3/F6, etc.)?
i believe the zooms have 2 and the the mixpres have 3, (according to mixpre patent filings)
-
Curious, do any the Zooms (F3/F6) or SD Mixpre II's have the option to still record 24 bit instead of 32 bit float?
-
Curious, do any the Zooms (F3/F6) or SD Mixpre II's have the option to still record 24 bit instead of 32 bit float?
MP-6 II can record either 32 or 24. F6 can record 32 or 24 or both and the F3 is 32-bit only.
-
Following up:
- How many ADC's do different recorders have (MixPre-3/6 ii, Zoom F3/F6, etc.)?
i believe the zooms have 2 and the the mixpres have 3, (according to mixpre patent filings)
^ I think so. The grist is how the switching between the multiple ADCs is handled. Can it be managed in such a way that the inherent change in noise-floor at the switchover (which is unavoidable unless some sort of noise reduction is done ahead of the switchover) is made imperceptible. To complicate things further I believe many multichannel ADC and DAC chips can be configured to combine their multiple internal units in parallel to incrementally increase dynamic range without switching between them. That avoids the switching problems but only gains a couple dB for each additional unit in parallel. Rather than switching between units set to target different levels, I assume that strategy does so by way of the signal being correlated across the multiple units while noise is not. I think that's how the strategy works anyway.. this goes well beyond my knowledge comfort zone. Perhaps both of these schemes are being used.
-
Curious, do any the Zooms (F3/F6) or SD Mixpre II's have the option to still record 24 bit instead of 32 bit float?
MP-6 II can record either 32 or 24. F6 can record 32 or 24 or both and the F3 is 32-bit only.
Keep in mind that the switching ADC architecture and the recorded file format are two separate, distinct things.
The relevant follow up question is: When the 32bit recorders that can do so are set to record 24bit files, does that mean only one ADC is used? Or is the machine still switching between ADCs in the same way as when in 32bitFP mode but writing a 24bit file? I suspect the second might be the case. If so it will not be a work around for the problem.
-
I'll keep staging my gain the old fashioned way using 24 bit.
I've stuck with that too so far. It does everything I need. That said, I don't expect this particular issue would be problematic for my use and what/how I record if I were to use a recorder which does multiple ADC switching.
The tl;dr of my post yesterday about the deeper fundamental issue: An audio recorder for music is not the same as a data-logger.
-
I am pretty sure that Zoom F series, and probably also SD, are still using multi ADC in 24 or 16 bit mode. IF anything can be done by Zoom it could be either: 1) allow a single ADC functioning when set to 24 bits mode or 2) provide a "smoother transition"between ADC with a sort of noise reduction or whatever they can figure out.
I have no technical knowledge to imagine any solution, but the posted YouTube video description it seems that Sound Devices has pretty fixed the issue with a firmware upgrade, so it is in theory possible to do something?
-
Since I am not recording water droplets and can hear absolutely no trace of noise during live music transitions, I will continue using my F3s with their great pre-amps and leave gain staging in my past along with my past fears of not demagnetizing my cassette deck heads often enough (or too often) or burning out my clutch in shifting my gears too hastily. YMMV
-
I am pretty sure that Zoom F series, and probably also SD, are still using multi ADC in 24 or 16 bit mode. IF anything can be done by Zoom it could be either: 1) allow a single ADC functioning when set to 24 bits mode or 2) provide a "smoother transition"between ADC with a sort of noise reduction or whatever they can figure out.
I have no technical knowledge to imagine any solution, but the posted YouTube video description it seems that Sound Devices has pretty fixed the issue with a firmware upgrade, so it is in theory possible to do something?
This point is addressed in the video. All of these recorders that have multiple ADCs still have all of them in the signal path even if you are reporting in 24-bit fixed mode.
Your possible solution #2 is what I understand Sound Devices did in their fix. I don't think we should assume the same will be possible in Zoom recorders. Not because of the different number of ADCs, but because the DSP hardware and software is likely to be significantly different between the two manufacturers.
-
This point is addressed in the video. All of these recorders that have multiple ADCs still have all of them in the signal path even if you are reporting in 24-bit fixed mode.
Your possible solution #2 is what I understand Sound Devices did in their fix. I don't think we should assume the same will be possible in Zoom recorders. Not because of the different number of ADCs, but because the DSP hardware and software is likely to be significantly different between the two manufacturers.
I think the biggest reason is that zoom doesn't care, and neither do most of their customers. If enough people complain, the f3n or f8x pro or whatever will have a different strategy.
-
Since I am not recording water droplets and can hear absolutely no trace of noise during live music transitions, I will continue using my F3s with their great pre-amps and leave gain staging in my past along with my past fears of not demagnetizing my cassette deck heads often enough (or too often) or burning out my clutch in shifting my gears too hastily. YMMV
Sure, you can do that as long as you are always and only running your mics direct in to the recorder. If you ever use an external preamp, or run any other signal chain consisting of multiple units into your F3, the gain staging ahead of the recorder will matter.
Gain staging will always remain relevant and important. Its not tied to any specific technology. It matters just as much inside the F3, the thing is Zoom has engineered it so you needn't be aware of it. In fact, the issue we've discussing is how the internal gain-staging across multiple ADCs is being managed somewhat imperfectly!
OT- Long live the manual transmission and drivers who know how to work them.
-
OT- Long live the manual transmission and drivers who know how to work them.
Most cars in Europe have manual transmissions and diesel engines. Cars are as disposable as bic lighters in America, they're most people's primary life investment in Europe, since people are way less transitory, and homes are oftentimes handed down from generation to generation.
-
Since I am not recording water droplets and can hear absolutely no trace of noise during live music transitions, I will continue using my F3s with their great pre-amps and leave gain staging in my past along with my past fears of not demagnetizing my cassette deck heads often enough (or too often) or burning out my clutch in shifting my gears too hastily. YMMV
Sure, you can do that as long as you are always and only running your mics direct in to the recorder. If you ever use an external preamp, or run any other signal chain consisting of multiple units into your F3, the gain staging ahead of the recorder will matter.
Gain staging will always remain relevant and important. Its not tied to any specific technology. It matters just as much inside the F3, the thing is Zoom has engineered it so you needn't be aware of it. In fact, the issue we've discussing is how the internal gain-staging across multiple ADCs is being managed somewhat imperfectly!
OT- Long live the manual transmission and drivers who know how to work them.
Thanks for the blessing. I drove manual transmission cars for decades. Modern automatic transmissions are now more efficient than manuals. Thank goodness. I am so old that I remember manual choke knobs on my first cars.
-
Most cars in Europe have manual transmissions and diesel engines. Cars are as disposable as bic lighters in America, they're most people's primary life investment in Europe, since people are way less transitory, and homes are oftentimes handed down from generation to generation.
At least in the Netherlands, this isn't really true. Most cars are manual (75%?), but the shift has been large and the rate of change is increasing as well. The vast majority of passenger vehicles run on gas, though. Probably ~80% gas. Of the other 20%, diesel is a substantial, but quickly decreasing, chunk. Pretty large increases in hybrids and electric the past few years. I also think people move/travel at least as frequently as Americans. I know more people from back home that never crossed the PA state line than Dutch that haven't traveled overseas...
[EDIT TO ADD] Manuals are considerably cheaper, so that is probably a reason they continue to sell. It is more expensive to own/operate a car here.
And, on topic, I would wager the reason SD was able to address this issue more easily is because their patented method basically estimates the instantaneous rate of change (i.e. first derivative) by running regression on a small number of samples. Any curve is essentially linear over a tiny interval, so they can leverage that to accurately estimate the shifts over very short time spans.
Also, there is no such thing as a free lunch. Maybe cheap, but not free. I suspect this is the reason that SD doesn't offer 32-bit float on their pro level recorders.
-
Continuing the OT-
I'm curious if a particular driver behavior change I notice here might be the same in the Netherlands and elsewhere, which may be partly due to the shift away from manual transmissions but I think is far more driven by mobile phone usage-
Decades ago, drivers would generally tend to coast to a red traffic light, then take off quickly as soon as the signal changed. That behavior seems to now be completely inverted- Drivers tend to rush to a red light, seemingly braking as late as possible, then slowly take off after the signal changes.
Anyone else notice this? I assume its because they are impatient get back on their phone while stopped at the light, and are slow to quit it after the change.
It's somewhat annoying as I'm still coasting, timing the light changes to preserve inertia, shifting manually. The new behavior folks speed past and swerve in front during the coast, then sit there after the light change, screwing up the well preserved inertia and timing. [shakes old man fits at cloud]
I do wonder if my current manual 6-speed (2013 era, its so good - manual transmission finally perfected) may unfortunately be the last of its breed. Its been a very long time since I was adjusting valves and timing, and longer still since I was working a choke-knob and shifting "3 on the tree"! Where's that old wire-recorder?
-
At least in the Netherlands, this isn't really true. Most cars are manual (75%?), but the shift has been large and the rate of change is increasing as well. The vast majority of passenger vehicles run on gas, though. Probably ~80% gas. Of the other 20%, diesel is a substantial, but quickly decreasing, chunk. Pretty large increases in hybrids and electric the past few years. I also think people move/travel at least as frequently as Americans. I know more people from back home that never crossed the PA state line than Dutch that haven't traveled overseas...
Yeah, my Euro experience is from 2007-2009 from living in Slovenia, so I shouldn't have assumed it was still the same today. Somewhat surprised though about the shift to gasoline. Not surprised at all about the shift to electric/hybrids.
I'd be curious to know more about transitory lifestyles in the US versus Europe. I know that in Slovenia, with only 2 million people, they're much more internally focused on Slovenia first...15 years ago anyway. There weren't alot of specialty manufacturers and/or industry and people were so proud of Slovenia...in general they had absolutely no desire to live anyplace else. The majority of jobs are people supporting other people. There's not nearly as much of a profit and loss focus by multi-national mega-company's. So people aren't as subject to economic cycles as in the US, where virtually nobody goes through their lives anymore without experiencing 2 or 3 or 4 layoffs. In Slovenia, employers were much more about providing a company that supports the lifestyles of its workforce than they were about making profits for their shareholders. At least that was my perception.
I hope it hasn't changed, honestly. I loved the fact that, in Europe, entire countries shut down for 6 weeks for summer holidays because, for six weeks anyway, the focus is on the people and the family and not the company.
-
Thanks for the blessing. I drove manual transmission cars for decades. Modern automatic transmissions are now more efficient than manuals.
about as efficient perhaps, but its impossible to be more efficient than a direct clutch connection (with a capable driver of course)
-
All this car talk has me thinking, is 32 bit float the CVT of recording formats? It's the most efficient transmission made, yet real driving enthusiasts hate them.
-
I think a dual clutch automatics (which are essentially manual transmissions operated automatically) are more efficient than a CVT, due to less friction loss through a direct clutch connection. They compensate for the lack of the CVT's infinite gear ratio by adding additional gear sets - eight or nine I in some cases I think. They are more efficient than a true manual for three reasons: There are additional gear sets to choose between. The dual clutch arrangement means any two adjacent gear sets are always simultaneously engaged and dual clutches switch between them, so there is no "down time" when the engine is disconnected from the drive train during a shift. And the shifting is computer managed, cyber vigilantly paying attention to the situation without distraction to optimize efficiency to the ability of its programming.
I just like doing the left foot right hand dance. Paddles just aren't the same. And operate the CD player at the same time. ; )
-
I have not read all of the above, but as someone who has been taping since 1971, I love the 32 bit world we have entered. Why? Because I have made great recordings. I have had no issues at all. The biggest difference is that I do not have to occasionally spend a inordinate amount of time leveling the first 10 or 20 seconds of a recording because it is too hot or too low. I have a few decks and do not at all mind going 24 bit when one of those makes sense, but using the same gear I have used for years, 32 bit is my choice. For me, how the end result sounds is what matters most. That said, I respect and admire all the dialog dissecting the respective differences in how we approach our passion.
-
I have not read all of the above, but as someone who has been taping since 1971, I love the 32 bit world we have entered. Why? Because I have made great recordings. I have had no issues at all. The biggest difference is that I do not have to occasionally spend a inordinate amount of time leveling the first 10 or 20 seconds of a recording because it is too hot or too low. I have a few decks and do not at all mind going 24 bit when one of those makes sense, but using the same gear I have used for years, 32 bit is my choice. For me, how the end result sounds is what matters most. That said, I respect and admire all the dialog dissecting the respective differences in how we approach our passion.
exactly.
-
I skimmed the Sound Devices 2016 patent for 32 bit (most of which is way above me) and note that it does describe 3 and 2 converter layouts, but noting it could be any number. I did not recognize any comment about the converter handoffs and noise control, but that could be me. I note that other manufacturers have to avoid infringing on their patent, so maybe SD understands the issue better and has the secret sauce in their patent. It would be nice to know if SD uses multiple converters in 24 bit mode, I haven't heard any issues and have not done much 32 bit.
On the clutch, I just prefer manuals as more fun. Today's automatics get better gas numbers and track lap times, but on the street that's not everything for me. I like road and machine feel, although most Americans apparently prefer automatics. I also don't use the phone while driving, even hands-free (legal requirement here).
As to the anti-theft aspect, I had an Integra manual for many years with a Club steering lock that was stolen 3 times, the last time driven 100 miles away and totaled. Integras were theft targets, so the thieves certainly were not stopped by a clutch or Club.
-
I have not read all of the above, but as someone who has been taping since 1971, I love the 32 bit world we have entered. Why? Because I have made great recordings. I have had no issues at all. The biggest difference is that I do not have to occasionally spend a inordinate amount of time leveling the first 10 or 20 seconds of a recording because it is too hot or too low. I have a few decks and do not at all mind going 24 bit when one of those makes sense, but using the same gear I have used for years, 32 bit is my choice. For me, how the end result sounds is what matters most. That said, I respect and admire all the dialog dissecting the respective differences in how we approach our passion.
exactly.
Ditto, from a guy who started taping in 1972.
-
Every car I owned from 1979 to 2011 was a manual. By that point years of driving in stop & go traffic keeping my left foot pressing the clutch most of the time had wiped out all the romance of a manual.
As for taping, the last deck I bought from Doug Oade is the Marantz PMD-561 with concert mode and the way he set the gain stages make it "almost" set and forget. He set the preamp gain so the overload point of the analog circuit and the clipping level of the A/D are matched perfectly. This allowed me to find a sweet spot to set the gain at the same starting position for 90% of shows that I never have to adjust again, or just once at the very beginning. Not the true "set & forget" convenience of 32bit float, but for a guy who used to make a dozen unecessary micro-adustments at every show I'm now comfortable only checking that the deck is running every now and then, which for me is huge.
-
Every car I owned from 1979 to 2011 was a manual. By that point years of driving in stop & go traffic keeping my left foot pressing the clutch most of the time had wiped out all the romance of a manual.
As for taping, the last deck I bought from Doug Oade is the Marantz PMD-561 with concert mode and the way he set the gain stages make it "almost" set and forget. He set the preamp gain so the overload point of the analog circuit and the clipping level of the A/D are matched perfectly. This allowed me to find a sweet spot to set the gain at the same starting position for 90% of shows that I never have to adjust again, or just once at the very beginning. Not the true "set & forget" convenience of 32bit float, but for a guy who used to make a dozen unecessary micro-adustments at every show I'm now comfortable only checking that the deck is running every now and then, which for me is huge.
Concur with the point of about losing the romance of shifting in traffic, especially when stopping uphill is involved. My German-built six speed automatic is wonderful. As for taping, yeah, my Oade-modified Sony M-10s are also great in the regard that you mention. No need for an external pre-amp and the sweet spot always seems to be at about three on the gain wheel, except in the quietest situations. That said, i still prefer the total ease of 32 bit Zoom F3, along with the security of the tight XLR connections.
-
Driving a four speed manual transmission around SanFrancisco for a few years was fun. Parallel parking on the steep hills was actually made somewhat easy by gravity. Other drivers pulling up too close behind at a light on really steep inclines was worrisome though. For I while to activate the starter I had to crawl under, so I'd almost always just park on an incline, roll out and pop the clutch. My current manual has a hill-hold feature which makes it almost too easy- if the car senses it is on an incline, the brakes remain engaged for 5 sec after releasing them or until the clutch is engaged. Almost no hill-skills required.
Here's an idea, the next round of 32bit float recorders could implement a feature that normalizes the 32bit float file after recording and writes a 24bit file from that. No loss, and everything fits in a smaller, standard fixed-point format file. Sort of a 24bit hill-hold manual in comparison to a 32bit-float auto-trans? Yeah, reaching a bit there.
-
My current manual has a hill-hold feature which makes it almost too easy- if the car senses it is on an incline, the brakes remain engaged for 5 sec after releasing them or until the clutch is engaged. Almost no hill-skills required.
I know this is still off topic but my GFs current ride is a 6 speed manual Mini Cooper and it has the hill brake thing. The first time we took a road trip and I was driving I had to stop on a really steep hill so I did the hand brake with the button pushed in thing to keep me from rolling back or stalling out and she just laughed at me. I learned in a three on the tree.
-
Back on topic....Given that the drawback of ADC switching might be a rise of noise floor, the point is that for some applications it may be audible, for others not.
Whatever the final delivery might be, i.e. 24bits 48khz, is there at your opinion a difference between recording 48, 96 or even 192 khz in order to minimize the potential issues?
-
I just pulled up the last 32-bit float recording I had made to find out if I could see and hear the problem being described.
Short answer: See? Definitely. Hear? Absolutely not.
See the attached screenshots from RX. This recording was 32fp/96kHz using my F6.
1. Hand Claps - You can clearly see the windowing as the noisefloor increases around the louder transients.
2. A Cappella - The three transients just right of center are accented syllables in the choir. You'll also see stray moments before and after where the converters switch.
3. Noise Zoomed - As it says on the tin. This shows why the added noise is inaudible - it's only showing up above 35 kHz or so.
So, is this still a concern from a sound design standpoint when captures will be pitched down, as shown in the video? Sure.
Is it a concern for studio or live music recording? Well, the noise floor is constantly shifting with the auto-ranging ADCs, at least on a Zoom F6. It's not just on the stray loud transients; it's happening frequently. But the noise is in a place that no one can hear. While I don't like that this is happening at all, it's not causing any problems in the audible range.
If anyone has a Sound Devices MixPre-II or one of the newer Tascam recorders that have auto-ranging ADCs, it would be helpful to compare their behavior.
EDIT: A user on GS has tested a Stagetec Nexus Compact and it does not do this. Maybe it's a Zoom implementation issue?
https://gearspace.com/board/all-things-technical/1425826-problems-32-bit-float-sound-design-pitching-down-white-noise.html#post16963512
-
Thanks @voltronic for posting your results. Most probably Stagetec or others have a better implementation…it is interesting though to read that Sound Devices seems to have fixed the issue via firmware update. This means that the issue was there before and that something has been possibly done. Don’t know wether Zoom will bother finding an improvement, should it be technically possible. Given the huge number of F series users- among which probably a lot of people recording and manipulating sounds- I can expect that some of all these complaints will eventually reach them and maybe lead to a fix.
The better behaviour with a SD788 is maybe due to a better hardware? It would be interesting to try the Zoom with an external preamp to tell how the overall tone of the recording changes.
-
the 788 works in a fundamentally different way.
-
From the manual and other statements made by Zoom it seems that the F serie uses two 16 bit converters; Zooms says that both in 32 bit and 24 bit mode the dual ADC is engaged.
This leaves to think that in 1 bit mode only one is working? If so it would be interesting to compare the spectral view of a 16 bit recording with a 32 bit one. Not to conclude it's better to go back at 16, just out of curiosity.
-
From the manual and other statements made by Zoom it seems that the F serie uses two 16 bit converters; Zooms says that both in 32 bit and 24 bit mode the dual ADC is engaged.
This leaves to think that in 1 bit mode only one is working? If so it would be interesting to compare the spectral view of a 16 bit recording with a 32 bit one. Not to conclude it's better to go back at 16, just out of curiosity.
I'm not sure where you got this impression, but neither of the ADCs are limited to 16-bit. One is optimized for high gain; the other for low gain.
Both ADCs are engaged no matter what recording mode is selected. You can't choose to use only one, and that's a point addressed in the video.
-
I’m sure I read it somewhere but it’s probably not true. I was maybe tricked by the fact that Zooms says that both 32 bit and 24 adopt the dual adc, but no mention about 16 bit.
-
YouTube poster has taken his video "private" for some reason, so hard for newcomers to see what the fuss was about;
https://youtu.be/y3431uljZ2k?si=I7I7qhsKK7TUo49s
-
Just bought a 32 bit recorder for this first time, and had a question. If I'm recording at 32-Bit, and then downsample to 24 or 16-Bit for distribution, will that introduce clipping because of the limitations of those 2 bitrates?
-
If you are planning to use the recording as-is with no other processing, then all you need to do is perform a level normalization to set the peak level to maximum 0dbFS, and add dither if truncating to 16-bit. If you are performing any EQ or other processing, do that before you do the above. BTW, this is not downsampling, which is changing the sampling *rate* (frequency). Rather you are changing the sample bit *depth*.
-
If you are planning to use the recording as-is with no other processing, then all you need to do is perform a level normalization to set the peak level to maximum 0dbFS, and add dither if truncating to 16-bit. If you are performing any EQ or other processing, do that before you do the above. BTW, this is not downsampling, which is changing the sampling *rate* (frequency). Rather you are changing the sample bit *depth*.
Just my 2 cents: I always set max level to somewhere below 0dBFS because you run the risk of inter-sample peaks clipping. If you are using software that measures "true" peak level then it is set up to detect these.
The term for lowering the bit depth is decimation.
My favorite tool for any jobs of this type is iZotope RX, FWIW.
-
Correction to Voltronic, the word you are looking for is "truncation" (discarding excess bits), not "decimation" (which is a step in sampling/resampling)
If I'm recording at 32-Bit, and then downsample to 24 or 16-Bit for distribution, will that introduce clipping because of the limitations of those 2 bitrates?
As commongrounder mentions, what you describe is not resampling, but a truncation of bit-length or bit-depth. That is to say, a lopping off and discarding of the least-significant bits.
Doing that can introduce clipping (or noise), but won't if you do it correctly. In essence you are moving the content from a very sloppy fitting larger "container" to a better fitting smaller one. So yes, you need to make sure it fits.. AND is positioned correctly inside the new container.
As covered in many of the discussions here at TS about "32-bit float recording", not only will all of the live music recordings we make already fit comfortably within 24bits, anything that can possibly be recorded successfully by any of the 32-bit recorders folks are currently using will also fit in a 24-bit "container". The musical content itself and everything we want to hear will always fit within 24bits.
However, because 32bit recording (or more accurately, the auto-switching ADCs schemes used in "32-bit recorders") delays the need for proper level setting until after the recording has been made, that musical content is likely to be shifted way up down in level, so its level needs to be adjusted prior to listening and/or conversion to a 24 or 16bit file format. Normalizing is the "doing the level setting afterward" part. There is no need to try and push it as close as possible to the limit by normalizing the highest peak all the way to 0dBfs. It's safer to normalize to a little less. The exact value doesn't matter. -1dB or -0.5dB is fine.
The next step after normalizing is the bit-reduction truncation. Whenever bit-reduction is performed it is always best practice to add a bit of dither (random noise) at a level just a tiny bit louder than the truncation value. Doing so essentially "smooths the sharp edge of the cut" eliminating quantization error. Even though that error is very low in level and probably not audible at all in a music recording, if made loud enough to be heard, quantization error noise sounds terrible, so a tiny bit of very low level random "hiss" (dither noise) is preferable. It may not be absolutely necessary to dither when reducing bit-length from 32-bit floating point to 24-bit fixed, in the same way that its not technically necessary to do so when reducing to 16-bit, but it is still best practice. So.. always dither whenever reducing bit-length. Its easy enough to do the right thing (usually by just checking a box to apply it). Don't worry about fancy noise shaped dither. Standard "triangular dither" (TPDF) is fine.. and is preferable if further processing might be done.
-
I'll throw in that normalisation is sometimes a built in function with 32 bit float devices (eg the Zoom H1 XLR, Zoom M2 & M3). I doubt whether anyone has yet tested the matter but I wonder whether normalising within the device is any better or worse than normalising outside the device (which does usually have the advantage of the user being able to select the maximum level). I guess I could test myself to see whether an in-device normalisation actually does the job to (say) -1dB rather than 0dB or thereabouts.
-
^ Would be cool if they included an internal function that would both normalize the 32bit float file AND produce a 24bit file from that on the recorder itself, streamlining post work where desirable. Sound Devices should be able to do that easily since the MixPre IIs already have the capability of supporting internal mixdown to a new file, I believe.
-
Ok, so I use Reaper these days for mastering, if I do all my EQ in that, and then normalize to something around 0 or below before I render it from Reaper in to a 16 or 24-Bit file, I should be ok at that point is what I'm understanding.
-
:coolguy:
-
If you are planning to use the recording as-is with no other processing, then all you need to do is perform a level normalization to set the peak level to maximum 0dbFS, and add dither if truncating to 16-bit. If you are performing any EQ or other processing, do that before you do the above. BTW, this is not downsampling, which is changing the sampling *rate* (frequency). Rather you are changing the sample bit *depth*.
Just my 2 cents: I always set max level to somewhere below 0dBFS because you run the risk of inter-sample peaks clipping. If you are using software that measures "true" peak level then it is set up to detect these.
My input: Ever since I read voltronic say this some years ago, I started normalizing to -.2dBFS
-
Darn! I go away for a few hours and miss a bunch of great discussion! I always appreciate gutbucket’s well worded explanations of the technical aspects of our craft/hobby. Thank you. Indeed, I never actually normalize right up to 0dbFS because of the mentioned inter-sample clipping possibility. I have the normalize function in my Wavelab setup dialed to -0.2dbFS. I should have made that note in my post.
Dithering of 32-bit float to 24-bit fixed is an interesting subject to me. I have chosen not to dither in this situation because the audio content inside the 32-bit float container is already very close to 24-bit in depth. Going to 16-bit will truncate a significant portion of the low level content so dithering is appropriate here. I’d be interested to hear other opinions about this (admittedly obscure) subject.
-
Sorry to bombard this thread with questions...just received my new toy today and am exploring the menu options. Most are pretty self explanatory, but I am unsure on which Sample Rate to use? 44.1, 48 or 96?
-
48. There's no chance your mics are more resolving than that. If you eventually want to do something like recording bats, and slowing them down to listen to, and you get ahold of some appropriate mics, 96 will help.
-
[snip] Dithering of 32-bit float to 24-bit fixed is an interesting subject to me. I have chosen not to dither in this situation because the audio content inside the 32-bit float container is already very close to 24-bit in depth. Going to 16-bit will truncate a significant portion of the low level content so dithering is appropriate here. I’d be interested to hear other opinions about this (admittedly obscure) subject.
Thanks for the kind words. Is it really necessary? No. Not at all. I'm not making the argument that it is. Only that its easier to just always dither and not bother oneself further about when it is or isn't necessary.
It is almost totally assured there is nothing but random noise down there at the 24bit truncation point, because after normalization the meaningful audio data isn't close to that cutoff point. That's the whole point of the music fitting comfortably within 24bits - we aren't loosing any useful information by down-sizing into the 24bit container. And.. if there was meaningful audio data down there, say some sort of still audible reverb tail decaying past the truncation point (which could never actually be heard except with a super quiet recording along with an impractically large amount of amplification - its never going to happen in a live music recording), the random noise that the reverb tail was decaying into would probably serve the same purpose as dither noise in eliminating quantization-error. So yeah, in reality there is no need to add a couple dB of additional noise down there with the application of dither.
My reasoning is more about just getting into the good habit of always dithering when reducing bit length, because its never the wrong thing to do, rather than making a special exception to that just because it doesn't really matter in this particular case.
-
48. There's no chance your mics are more resolving than that. If you eventually want to do something like recording bats, and slowing them down to listen to, and you get ahold of some appropriate mics, 96 will help.
I'm with grawk on this too. Some will argue that higher sampling rates are beneficial for post processing, to which I say ok then use the oversampling function that is built-into most modern post processing effects and crank it up to 8X or whatever your computer resources can handle comfortably. There is actually a good argument for not recording a bunch of ultrasonic content (which recording at 48 automatically achieves), or actively filtering out any ultrasonic content that's already present in material that was recorded at a higher sampling rate, prior to applying post processing with oversampling.
Fun to record bats and bugs and slow that down into audibility though. Always exceptions in certain cases.
-
The M2 does have the facility to normalise and export to lower bitrates all in the recorder. It can also normalise on the fly during playback. I would lay money that the H1 XLR will work the same way. I have probably posted this before but here is a crude test I did recording 'silence' at night in the garden, normalising it in the device, and the recording low level speech for comparison. I do recommend listening to all of it and noting the explanation that shows as it plays, without which completely the wrong impression might be given, but it's less than 2 minutes long.
https://youtu.be/Fb1r11dYpUY?si=5SyqVQzpWx5ZWwr6
-
Darn! I go away for a few hours and miss a bunch of great discussion! I always appreciate gutbucket’s well worded explanations of the technical aspects of our craft/hobby. Thank you. Indeed, I never actually normalize right up to 0dbFS because of the mentioned inter-sample clipping possibility. I have the normalize function in my Wavelab setup dialed to -0.2dbFS. I should have made that note in my post.
Dithering of 32-bit float to 24-bit fixed is an interesting subject to me. I have chosen not to dither in this situation because the audio content inside the 32-bit float container is already very close to 24-bit in depth. Going to 16-bit will truncate a significant portion of the low level content so dithering is appropriate here. I’d be interested to hear other opinions about this (admittedly obscure) subject.
To my understanding, it's exactly 24 bits of data precision.
https://www.sounddevices.com/32-bit-float-files-explained/ (https://www.sounddevices.com/32-bit-float-files-explained/)
https://en.wikipedia.org/wiki/Single-precision_floating-point_format#IEEE_754_standard:_binary32 (https://en.wikipedia.org/wiki/Single-precision_floating-point_format#IEEE_754_standard:_binary32)
1 sign bit
8 exponent bits
23 data bits; as the wiki says: 23 bits are stored explicitly, but in effect, there are 24 bits of data.
-
The Zoom H1 XLR manual is now online and that confirms that it can create either 16 bit or 24 bit files from the 32 bit float originals, while normalising at the same time. Or it can normalise while retaining 32 bit float. The original files are retained of course.
-
48. There's no chance your mics are more resolving than that. If you eventually want to do something like recording bats, and slowing them down to listen to, and you get ahold of some appropriate mics, 96 will help.
Ha, can't imagine I'll ever have a need to do that, 48 will work just fine. Tried it out last night, and am confused on something. It created two files for each recording, one titled "TrLR", and another titled "TrMic". I did a little searching, and from what I could understand, the backup was created at a lower recording level to avoid clipping. I thought clipping with 32-Bit Float was considered impossible to begin with? Is having a backup file like that necessary?
-
Lr is the xlr inputs and mics are the built in mics
-
Hmm...weird, I wasn't able to use the XLR's last night, only the internal mics, but somehow there is audio on both files.
-
Second time using the Zoom H4 tonight, using XLR connected to a soundboard...I'm pretty disappointed. There is a lot of clipping, the big selling point was "The H4essential captures every nuance of your sound ensuring high-quality, clip-free audio in every take." I'm guessing user error, but I'm not sure what, I thought it was just set it and forget it.
-
It’s always possible to overload inputs. For consumer gear you regularly need attenuators for a board patch.
-
Max input level is LINE: +24 dBu, but that requires the trs input, according to the marking on the front. So if you had it connected to the line output of the board into the mic (XLR) input of the recorder, it would seriously overload well before the 32 bit float converters. They don't help with analog distortion before them in the signal chain. Indeed they will make a very good recording of it...
-
It’s always possible to overload inputs. For consumer gear you regularly need attenuators for a board patch.
Odd, I've used my Tascam DR-60 and 70 without issue when recording from a SBD, which is why I didn't think anything of it. Starting to wonder if I might be better off just sending it back and going with something like the Zoom H4N if I have to monitor levels anyway, at half the price. I like the equipment I have, I am just trying to downsize as much as I can.
Max input level is LINE: +24 dBu, but that requires the trs input, according to the marking on the front. So if you had it connected to the line output of the board into the mic (XLR) input of the recorder, it would seriously overload well before the 32 bit float converters. They don't help with analog distortion before them in the signal chain. Indeed they will make a very good recording of it...
I believe that it was a line output, it was a 1/4" plug. The recording is salvageable, just caught me off guard that it clipped.
-
If line-input via TRS is capable of handling an input level of up to +24dBu, it should handle a patch from the soundboard just fine. I Just checked the on-line manual and found the note below-
"NOTE
To make handling the levels of input signals easy, the H4essential sets input levels according to the types of plugs connected to the input jacks.
• When connecting mic-level devices, use XLR plugs.
• When connecting line-level devices, use TRS plugs."
-
Double check all the relevant menu settings. Phantom power off? (may not matter if not available on the TRS inputs, but best to turn it off anyway). Recorder set to record in 32-bit-floating point mode? Etc, etc?
Also it is entirely possible the signal was distorted prior to reaching the recorder. Any confirmation that it was clean? Maybe by checking with headphones prior to or during recording?
-
Double check all the relevant menu settings. Phantom power off? (may not matter if not available on the TRS inputs, but best to turn it off anyway). Recorder set to record in 32-bit-floating point mode? Etc, etc?
Also it is entirely possible the signal was distorted prior to reaching the recorder. Any confirmation that it was clean? Maybe by checking with headphones prior to or during recording?
Phantom power was off. I don't even see an option to change the bitrate, I think that's just the default and it can't be changed. I did not get a chance to listen to it during the recording, I was lucky to get the patch at all, you could tell this was not something they do all the time and they were just hoping it was the right output. Maybe with a more experienced person running sound I might get better results. I have a show today I am able to patch to the SBD as well, I'll give it another try. I'll leave samples of the recording, maybe someone can give more insight than I can gain from it.
RAW File: https://drive.google.com/file/d/10YB3syMj4nx8GCK8Q2ejivtQSh6-sRrG/view?usp=sharing
Normalized: https://drive.google.com/file/d/1rTZQZmA2F8WPvEk3NG-cDTeT6jqx7Omi/view?usp=sharing
Waveform:
(https://i.imgur.com/3KBsfWl.png)
-
I'd lay money that it's clipped at the source. Singer's mic so high level compared to the band? One 'problem' with today's recorders is that they make excellent recordings of the warts and all...
-
If line-input via TRS is capable of handling an input level of up to +24dBu, it should handle a patch from the soundboard just fine. I Just checked the on-line manual and found the note below-
"NOTE
To make handling the levels of input signals easy, the H4essential sets input levels according to the types of plugs connected to the input jacks.
• When connecting mic-level devices, use XLR plugs.
• When connecting line-level devices, use TRS plugs."
"THE" soundboard? What console? They have different max output levels...
If the speaker system is small and/or underpowered, the operator might be running the board up near clipping in order to achieve their desired output levels.
-
If line-input via TRS is capable of handling an input level of up to +24dBu, it should handle a patch from the soundboard just fine. I Just checked the on-line manual and found the note below-
"NOTE
To make handling the levels of input signals easy, the H4essential sets input levels according to the types of plugs connected to the input jacks.
• When connecting mic-level devices, use XLR plugs.
• When connecting line-level devices, use TRS plugs."
Ok, just to make sure I am understanding it...when I am using my mics or connecting to a soundboard/mixer that has an XLR output, use the XLR inputs on the Zoom. If I am connecting to soundboard/mixer with a 1/4" connection, I should be using the "Line In" input? So using a 3.5 MM input with a dual splitter and 1/4" adapters to the soundboard should work? No attenuator needed because it's 32 bit float?
-
From that manual quote, their advice is to use the 1/4" input on the H4essential for line level signals from a soundboard. So you would probably need two pairs of leads in your kit. One pair would be XLR to 1/4" TRS for when the soundboard was outputting from XLR. The other pair would be 1/4" to 1/4" TRS for when the soundboard was outputting from 1/4".
-
If line-input via TRS is capable of handling an input level of up to +24dBu, it should handle a patch from the soundboard just fine. I Just checked the on-line manual and found the note below-
"NOTE
To make handling the levels of input signals easy, the H4essential sets input levels according to the types of plugs connected to the input jacks.
• When connecting mic-level devices, use XLR plugs.
• When connecting line-level devices, use TRS plugs."
Ok, just to make sure I am understanding it...when I am using my mics or connecting to a soundboard/mixer that has an XLR output, use the XLR inputs on the Zoom. If I am connecting to soundboard/mixer with a 1/4" connection, I should be using the "Line In" input? So using a 3.5 MM input with a dual splitter and 1/4" adapters to the soundboard should work? No attenuator needed because it's 32 bit float?
What Morst and Ozpeter said.
Think of it this way.. You won't decide which input on the recorder to use based on the output connector on the microphone, soundboard, preamp, or whatever else you are recording. You will instead choose which input to use based on the signal level produced by what you are plugging in.
Microphones get plugged into the XLR input on the recorder, regardless of what output connector the microphone actually features, because they will all produce a lower "mic level" output. Sources that produce a higher level of output, such as a patch from the mixer, soundboard, or the output from an external preamp go into the TRS input, regardless of their output connector. That means you will need to use adapters to convert from XLR to TRS and vice versa whenever necessary, or use dedicated cables with the different appropriate connectors installed on either end.
This is all about selecting the appropriate recorder inputs based on the output level of whatever you are plugging into it, and has nothing to do with the recorder featuring a 32-bit floating point recording format or not.
-
If line-input via TRS is capable of handling an input level of up to +24dBu, it should handle a patch from the soundboard just fine. I Just checked the on-line manual and found the note below-
"NOTE
To make handling the levels of input signals easy, the H4essential sets input levels according to the types of plugs connected to the input jacks.
• When connecting mic-level devices, use XLR plugs.
• When connecting line-level devices, use TRS plugs."
Ok, just to make sure I am understanding it...when I am using my mics or connecting to a soundboard/mixer that has an XLR output, use the XLR inputs on the Zoom. If I am connecting to soundboard/mixer with a 1/4" connection, I should be using the "Line In" input? So using a 3.5 MM input with a dual splitter and 1/4" adapters to the soundboard should work? No attenuator needed because it's 32 bit float?
What Morst and Ozpeter said.
Think of it this way.. You won't decide which input on the recorder to use based on the output connector on the microphone, soundboard, preamp, or whatever else you are recording. You will instead choose which input to use based on the signal level produced by what you are plugging in.
Microphones get plugged into the XLR input on the recorder, regardless of what output connector the microphone actually features, because they will all produce a lower "mic level" output. Sources that produce a higher level of output, such as a patch from the mixer, soundboard, or the output from an external preamp go into the TRS input, regardless of their output connector. That means you will need to use adapters to convert from XLR to TRS and vice versa whenever necessary, or use dedicated cables with the different appropriate connectors installed on either end.
This is all about selecting the appropriate recorder inputs based on the output level of whatever you are plugging into it, and has nothing to do with the recorder featuring a 32-bit floating point recording format or not.
I feel like an idiot, but up until a couple days ago I didn't realize there was another input inside the XLR input. I ended up buying 2 of these cables specifically for soundboard/mixer outputs
https://www.amazon.com/gp/product/B07NZ2NVRV/
-
I suspect the best recording engineers all started as idiots... :cheers:
-
I didn't realize there was another input inside the XLR input.
FYI- That style of combo TRS/XLR jack is common on a lot of equipment, and usually both inputs are connected to the same circuit (the two inputs are wired in parallel), so that you can use cables terminated with either connector type.
This particular implementation where the TRS jack in the center feeds an entirely different circuit (line-in) than the XLR jack (mic-in) is more unusual, but is common to some Zoom recorders. I suppose they did it that way to eliminate the need to implement a mic-in/line-in switch. Its not that way on all Zoom recorders though. For example, the first generation F8 uses that scheme, while the later F8N and F8N-PRO models have the two inputs wired in parallel, along with an added menu switch to choose mic or line input.
-
The new Zoom H1 XLR also adopts the switch approach, FWIW (with separate switches for the two main inputs).
-
I suspect the best recording engineers all started as idiots... :cheers:
Yah, if you think you know everything, it's gonna be hard to learn anything new.
A first step to learning is to realize you need to know an answer!
A nice thing about learning a new subject is that you can make rapid progress.
After 40+ years of doing this, I'm lucky if I can learn one thing per gig without making a mistake to get the knowledge!! "Darn it I wanted to make a recording, and all I got was this valuable lesson!"
-
Someone has to say it ... Read the manual, but these days it seems to take more than that
In my heavy photograph days, no expensive gear was bought without first reading the "Magic Lantern Guide", sort of a third-party users,' manual.
(https://i.ebayimg.com/images/g/xncAAOSwAK5gAAS1/s-l300.webp)
I remember thinking that everything in life should have such a guide.... Cars, spouses, children, bosses, wardrobe....
And thanks, all, for the good info on combi-XLRs!
-
when i would get sbd feeds, if i got a dedicated output where the level could be controlled, i would ask for the level to be at 50%. it might be too low, but you cant really dick back and forth asking little higher, little lower. better lower than clipped. plus you never really know what the level is until the show is in progress. usually when you get the feed pre recorded music is being played. plus it might start out lower and gradually get louder.
-
I posed a question on someone's YouTube review of the Zoom H4Essential, and am even more confused. In theory, 32-Bit Float shouldn't ever clip, why would this be a concern then? If I still need to worry about clipping, wouldn't I be just as well off with a cheaper 24/16 bit recorder?
(https://i.imgur.com/9NXQ6DZ.png)
-
First, if you overload the analog input on any recorder, you’re going to get clipping regardless of the number of bits for the sample rate or anything else. The max input level is the max input level.
Second the h series is already cheap. Anything cheaper is likely to be worse.
-
So 32-Bit Float is basically snake oil.
-
No but it’s a solution to a problem you’re not having right now. And it’s not magic. The quality of the device still matters. Not just the format it stores audio in.
-
^ Would be cool if they included an internal function that would both normalize the 32bit float file AND produce a 24bit file from that on the recorder itself, streamlining post work where desirable. Sound Devices should be able to do that easily since the MixPre IIs already have the capability of supporting internal mixdown to a new file, I believe.
Gut, you're gonna have to demonstrate that there is substantial benefit from doing the processing on the unit versus a DAW. That touchscreen is minuscule, even compared with a laptop monitor. And control bandwidth on the mp-x is again minuscule compared to a kbd and mouse.
Why would I want to do those things on the recorder?
-
Onboard processing can suit some scenarios depending on many factors. Can be handy, and in certain video related scenarios when used with simple video editing systems it could be essential. It doesn't need a sophisticated display. All you are doing is making simple menu choices. Sample rate, tick the one required, normalise yes or no. It works fine on my Zoom M2 mic with its small display and processing is fast. You can do it in the car on the way home from the gig. Assuming someone else is driving of course...
-
I am possibly repeating myself, but 32 bit float A to D conversion cannot clip. The prior analog stages can. But these devices are designed with gain staging set so that under normal circumstances you would have to work hard to clip the analog front end. +4dBu is industry standard.
-
Meanwhile, I remembered while reading this thread that my very first digital recording of any kind, back in 1983, was... 14 bits. How far we have come! When I was asked to make it, I had never even heard a digital recording. I had a LOT to learn. I hired a Sony F1 / SLF1 system that used betamax tapes. I borrowed a pair of mics. Here are a couple of relevant links. The second link reproduces the Sony blurb which explained what digital recording was all about. Actually, I have just read there that the F1 could run at 16 bit for those wanting the ultimate quality. So maybe I did the right thing and used that option in fact.
https://www.palsite.com/slf1ovi.html
https://www.palsite.com/pcmf1ovi.html
Anyway, the end result was no disaster (imho) and it can still be heard on YouTube Music. At the time it got a gold award from a French magazine review. It's the first album in the playlist of some of my albums which I had a bit of fun putting together today, using the Discogs database to remind myself what I have recorded over the years. Some albums I had completely forgotten about!
https://music.youtube.com/playlist?list=PLW7X_zDpTwy3bogeppBSktVXgfkvgzNLf&si=6qedaTGP50lrDuMg
Anyone interested in what a very old digital recording sounds like (possibly mangled by YouTube) might care to listen to the first 30 seconds...
I guess this post is the very definition of off-topic. Sorry!
-
^ Would be cool if they included an internal function that would both normalize the 32bit float file AND produce a 24bit file from that on the recorder itself, streamlining post work where desirable. Sound Devices should be able to do that easily since the MixPre IIs already have the capability of supporting internal mixdown to a new file, I believe.
Gut, you're gonna have to demonstrate that there is substantial benefit from doing the processing on the unit versus a DAW. That touchscreen is minuscule, even compared with a laptop monitor. And control bandwidth on the mp-x is again minuscule compared to a kbd and mouse.
Why would I want to do those things on the recorder?
Just as an option which would allow the user to leverage the 32bit float advantages of not having to set input levels ahead of time, but produce a 24bit fixed output file directly on the recorder instead of having to transfer the file to a computer to do that. Wouldn't be forced to operate in it that way, it would simply be a nice to have option. Could always still transfer the original 32bit float file to the computer later.. along with its 24bit fixed derivative.
-
I plan to chain a (borrowed) Beachtek DXA-PRE ahead of the F3 the to see what sort of mayhem ensues.
The owner does corporate gigs, so I'm trying to find a time that he doesn't need it and I have a local gig to tape.
-
^ Would be cool if they included an internal function that would both normalize the 32bit float file AND produce a 24bit file from that on the recorder itself, streamlining post work where desirable. Sound Devices should be able to do that easily since the MixPre IIs already have the capability of supporting internal mixdown to a new file, I believe.
Gut, you're gonna have to demonstrate that there is substantial benefit from doing the processing on the unit versus a DAW. That touchscreen is minuscule, even compared with a laptop monitor. And control bandwidth on the mp-x is again minuscule compared to a kbd and mouse.
Why would I want to do those things on the recorder?
Just as an option which would allow the user to leverage the 32bit float advantages of not having to set input levels ahead of time, but produce a 24bit fixed output file directly on the recorder instead of having to transfer the file to a computer to do that. Wouldn't be forced to operate in it that way, it would simply be a nice to have option. Could always still transfer the original 32bit float file to the computer later.. along with its 24bit fixed derivative.
The upcoming Zoom H1 XLR has exactly this option, see screenshot of the manual...
-
^ which is where we started that discussion on 22 August. :cheers:
-
:coolguy:
I suspect other recorders will incorporate the same over time.
..and wonder what value the H1 XLR normalizes the highest peak to, since it doesn't say.
-
I will try to check that function on my M2 Mictrak device tomorrow to see what level that gives when normalised. I imagine other devices in their stable will work the same way.
-
Well, using the Zoom M2 I generated a test tone @ 440Hz using Adobe Audition, recorded 30 seconds of it from loudspeaker replay, and then used the built in export function to normalise the recording and export to 16 bit and 24 bit (two rapid operations). Then I opened the exported files one at a time in Audition, replayed them, and noted the maximum level shown (and retained) on the meters. This showed that the file was normalised within the device to 0dB. Originally the peak value of the 32 bit float file was -11dB approx.
However, this exercise revealed a drawback in using the built in normalisation function. Right at the start of the recording there was some kind of transient click, either from handling or from pressing the record button. This was actually louder than the recorded tone. So the tone wasn't normalised as such - the click dictated how the whole file was normalised. If I had been doing this with real material I would probably have edited off the click in the DAW, and then normalised the actual wanted part of the recording. This might not happen all the time, depending on circumstances - I wasn't very careful about how I was holding the device when I recorded the tone, and as it's almost midnight here I replayed the tone at a modest level not wishing to disturb the neighbours. But it does illustrate a possible hazard in undertaking whole-file internal normalisation. But, it would give you something reasonably playable in terms of level, prior to further level control operations.
[Edited to add, while talking about this much maligned device - I just recorded some room silence using the M2 by the time honoured method of shoving it under a folded duvet in the early hours of the morning in a silent bedroom. I then examined the file in Adobe Audition. In the spectral frequency display there was absolutely no horizontal lines at all (which would indicate a noise band of some kind) and that was with the range cranked up to 210dB. Looking at the frequency display there was a maximum level of about -91dB at around 100Hz. This could be city rumble. At 1000 Hz the level was about -117dB and at 5Khz the level was -126dB. The level continued to drop slightly so that at 10kHz it was -127dB. Above that it slightly declined further. I doubt whether this amount of noise (system noise plus mic noise) would be of any real consequence or concern even in a classical recording - and it seems to me for the price of the device, it's nothing to complain about. I paid $117 US dollar equivalent back in December.]
-
If line-input via TRS is capable of handling an input level of up to +24dBu, it should handle a patch from the soundboard just fine. I Just checked the on-line manual and found the note below-
"NOTE
To make handling the levels of input signals easy, the H4essential sets input levels according to the types of plugs connected to the input jacks.
• When connecting mic-level devices, use XLR plugs.
• When connecting line-level devices, use TRS plugs."
Ok, just to make sure I am understanding it...when I am using my mics or connecting to a soundboard/mixer that has an XLR output, use the XLR inputs on the Zoom. If I am connecting to soundboard/mixer with a 1/4" connection, I should be using the "Line In" input? So using a 3.5 MM input with a dual splitter and 1/4" adapters to the soundboard should work? No attenuator needed because it's 32 bit float?
Microphones get plugged into the XLR input on the recorder, regardless of what output connector the microphone actually features, because they will all produce a lower "mic level" output. Sources that produce a higher level of output, such as a patch from the mixer, soundboard, or the output from an external preamp go into the TRS input, regardless of their output connector. That means you will need to use adapters to convert from XLR to TRS and vice versa whenever necessary, or use dedicated cables with the different appropriate connectors installed on either end.
Had a thought today...right now I carry a set of cables for each possibility (XLR male to male, XLR female to male, and 1/4" male to 1/4" male). If I were to just buy some adapters such as this:
https://www.sweetwater.com/store/detail/PWP047Z--daddario-planet-waves-pw-p047z-xlr-male-to-1-4-inch-female-balanced-adapter
Would that give me the same results as a dedicated cable? I love the idea of bringing just a 1/4" male to male cable and just adding adapters when needed.
-
In terms of signal path, yes. You could plug your 1/4" TRS terminated cable in directly for line-in, and add the XLR adapter for mic-in.
In terms of practicality, consider things like how long the adapter connection becomes which is going to stick out from the side of the recorder - how you will manage that, how much leverage it applies, and how good and tight the additional signal path connections are.. stuff like that. A lot of times I prefer a short "pig-tail" type adapter that features a short wire between the two connectors just long enough to provide strain relief and the ability to fold under the recorder. The adapter-to-recorder connection then protrudes far less and is much better strain-relieved, and the cable-to-adapter connection can be neatly tucked away under or behind the recorder in the bag.
-
In terms of signal path, yes. You could plug your 1/4" TRS terminated cable in directly for line-in, and add the XLR adapter for mic-in.
In terms of practicality, consider things like how long the adapter connection becomes which is going to stick out from the side of the recorder - how you will manage that, how much leverage it applies, and how good and tight the additional signal path connections are.. stuff like that. A lot of times I prefer a short "pig-tail" type adapter that features a short wire between the two connectors just long enough to provide strain relief and the ability to fold under the recorder. The adapter-to-recorder connection then protrudes far less and is much better strain-relieved, and the cable-to-adapter connection can be neatly tucked away under or behind the recorder in the bag.
For the most part, my recorder is at ground level, or on a platform that doesn't cause any stress to it due to cable pull. I record a lot of local, smaller musicians who have a very basic setup, and usually I just plug in and set it on the floor behind their setup. I've bought a couple cases and adapted them to my recorders, so the recorder sits inside the case and I drill holes in the bottom of it to plug cables in, it keeps it safe from harm and together as one unit basically.
-
Perhaps I am at risk of being accused of not having read the whole thread, but...
In the thread related to the Zoom H4e, a user has mentioned that in the manual for the device (and not in the advertising!) the 3.5mm input is flagged as not being fed through a dual a/d converter. Subsequent research indicates that some He series devices do not claim to have dual converters at all. So what is happening? Is the audio simply going through a 24 bit converter with the output saved in 32 bit float format, and is there any actual benefit in doing that, apart from the sales benefit of having a 32 bit float sticker on the device? I am confused... again...
-
Perhaps I am at risk of being accused of not having read the whole thread, but...
In the thread related to the Zoom H4e, a user has mentioned that in the manual for the device (and not in the advertising!) the 3.5mm input is flagged as not being fed through a dual a/d converter. Subsequent research indicates that some He series devices do not claim to have dual converters at all. So what is happening? Is the audio simply going through a 24 bit converter with the output saved in 32 bit float format, and is there any actual benefit in doing that, apart from the sales benefit of having a 32 bit float sticker on the device? I am confused... again...
Good question...
Indeed it is very easy to convert the output of a single 24bit ADC to 32bfp format. I do see benefit for storing in 32bfp if math is done on the samples in the digital realm in the recorder (amplification, filtering, equalizing, mixing, limitting etc). Some of these operations could potentially result in samples > 0dB, which are not possible to store as 24bit linear samples without clipping. And even if all samples remain < 0 dB, storing the output of such operations as 24bit linear samples would require rounding, while storing in 32bfp would allow to store the resulting samples more accurately. If this difference is audible is another question...
-
Thanks for your response - I have to say that I trust it rather more than the response I got from "DeepSeek" (which I am amusing myself with at the moment), which I found more confusing than helpful! I think I will continue to ask about location recording matters here rather than there in future... I won't waste space with the full response, but here is the summary bit -
"Summary
There is no single ADC that directly outputs 32-bit float audio.
Dual ADCs are used to capture a wider dynamic range and avoid clipping, which is then processed into 32-bit float format.
32-bit float is a digital format created in post-processing, not a native output of an ADC.
This approach allows portable audio recorders to deliver the benefits of 32-bit float recording, such as immense headroom and dynamic range, even though the ADCs themselves do not natively output 32-bit float data."
-
I have to say that I trust it rather more than the response I got from "DeepSeek"
There is a lot of confusion and misperceptions and marketing nonsense about 32bfp recording on the internet. DeepSeek and it's western counterparts are only repeating what they've learned from the often misleading and incorrect content it finds on the web. Personally I'm a big fan of multi ADC 32bfp recording as I often have to deal with situations where it is unclear what a good record level would be upfront. So I like the fixed gain concept, and then accept the need for multi-ADC's to keep the noise down in the quiet parts.
-
Since your DAW almost certainly upsamples to 32-bit for processing, there is no benefit to converting a single 24-bit ADC output to 32-bit floating point in the recorder. Just normalize to -1 dB (or -.1 dB or whatever) in the last processing step. Any rounding errors will be so minuscule as to be completely inaudible in virtually any recording situation.
The dual (or more) ADC is the key to the whole thing, in my opinion.
-
Since your DAW almost certainly upsamples to 32-bit for processing, there is no benefit to converting a single 24-bit ADC output to 32-bit floating point in the recorder. Just normalize to -1 dB (or -.1 dB or whatever) in the last processing step. Any rounding errors will be so minuscule as to be completely inaudible in virtually any recording situation.
The dual (or more) ADC is the key to the whole thing, in my opinion.
If so, how can a device with only one ADC accurately claim to be a floating 32-bit device?
-
^ It is easy to turn a 24-bit PCM file into a 32-bit floating point format. It won't make any difference, however, other than increasing the file size. It is advertising trickery...
-
As someone who has just gotten in to the whole 32 Bit float thing, I am wondering...are there any drawbacks to using a recorder with it? I get some people don't believe it makes a difference between 24 & 32 bit and they think it's all a scam, but are there any obvious negatives to using a 32 bit recorder as compared to 24 bit? Or at worst is there just no difference?
-
As someone who has just gotten in to the whole 32 Bit float thing, I am wondering...are there any drawbacks to using a recorder with it? I get some people don't believe it makes a difference between 24 & 32 bit and they think it's all a scam, but are there any obvious negatives to using a 32 bit recorder as compared to 24 bit? Or at worst is there just no difference?
For the recorders, the only concern I've seen is a mostly theoretical concern that noise levels can get high when the recorder needs to switch to the high gain ADC, because the overall analog noise floor stays the same, it's just amplified.
-
As someone who has just gotten in to the whole 32 Bit float thing, I am wondering...are there any drawbacks to using a recorder with it? I get some people don't believe it makes a difference between 24 & 32 bit and they think it's all a scam, but are there any obvious negatives to using a 32 bit recorder as compared to 24 bit? Or at worst is there just no difference?
Most importantly to me, it means I do not have to worry about properly predicting how to set the gain level in advance. Set it up, turn it on, and then relax and enjoy the music. As good as 24 bit but less worries. Works for me. YMMV
-
^ It is easy to turn a 24-bit PCM file into a 32-bit floating point format. It won't make any difference, however, other than increasing the file size. It is advertising trickery...
So it negates the no-clipping advantage of actually recording in floating 32 bit?
-
As someone who has just gotten in to the whole 32 Bit float thing, I am wondering...are there any drawbacks to using a recorder with it? I get some people don't believe it makes a difference between 24 & 32 bit and they think it's all a scam, but are there any obvious negatives to using a 32 bit recorder as compared to 24 bit? Or at worst is there just no difference?
For the recorders, the only concern I've seen is a mostly theoretical concern that noise levels can get high when the recorder needs to switch to the high gain ADC, because the overall analog noise floor stays the same, it's just amplified.
Based on my extensive use of the F3, the word "theoretical" is the key word here. Never heard any such noise in my many recordings with the F3
-
So it negates the no-clipping advantage of actually recording in floating 32 bit?
The "magic" in the 32-bit float recorders is the multi-ADC setup. That's what allows them to not clip (unless the input clips), but, at the same time, keep the noise floor down. The 32-bit floating point side of it is really just a mathematically convenient way to implement the multi-ADC design.
-
^That.
There are 3 steps:
1) Use of additional ADCs running in parallel with different levels of input gain to extend dynamic range.
2) Combining the outputs of the multiple ADCs into one data stream.
3) Storing that output.
You can have the first without the other two. That is how secondary safety tracks have been implemented in some recorder for the last few decades. Those recorders stored both ADC outputs as separate files, one at a higher level than the other, which allowed the user to choose whichever one happened to fit the full dynamic range better than the other afterward. Those recorders essentially operate no differently than modern multi-ADC recorders in regard to this first step.
The combining of the outputs is the magic sauce, the tricky part, and is the step that is related to the low-level noise modulation artifacts reported to occur in some cases. In modern multi-ADC recorders this is done in real time while recording. But it can also be done afterward on the computer as long as the output of each ADC is stored separately by the recorder. TS member TheJez has created a program that automatically merges the two separate files created by "safety track" recorders afterward. That program and a more detailed description of what it does can be found in the SafetyTrackMerger (Windows + MacOS) thread- https://taperssection.com/index.php?topic=206443.msg2421571#msg2421571 (https://taperssection.com/index.php?topic=206443.msg2421571#msg2421571)
Whether the merge is done in real time in the recorder or afterward on the computer, the combined output of the multiple ADCs is then stored in some file format. That might be 24bit fixed-point, 32-bit floating-point, or whatever. Some recorders which do the merge in real time allow you to choose the output format, others don't.
The dynamic range of any one thing we are recording can fit in a 24bits fixed-point file. But the full range of all the different things we might plug into a recorder doesn't quite. Writing a 32-bit floating point file that has the capability of storing a ridiculously large range of level values fixes that, even if its not really much of a problem in actual use. The real-word usable dynamic range of any analog input recorder is always going to be limited by the dynamic range capabilities of its analog input, which is usually smaller than what a 24bit file is capable of storing. But a few ADCs working in parallel with different input gains can collectively exceed the dynamic range capability of that by just enough that the user needn't manually set input gain.. except they actually still do.. by manually switching between "line" and "mic" input sensitivities.
That such mic/line switching is still required is one clear indication that 32-bit float recorders do not actually have the super huge dynamic range capability which is often advertised (only the file format itself has that). They effectively have just enough more dynamic input range capability via their automated ADC switching to make them more convenient. We still need to make sure that we aren't overloading the input or not driving it strongly enough to avoid overloading or noise, but we can do that ahead of time when setting things up originally.
Beyond that, a 32-bit float file format has some advantages for some forms of processing. But modern programs and DAWs always convert the input file format to some larger internal workspace format before doing the processing anyway.
So there are some advantages (even though the advantages frequently don't actually reflect what is being advertised) such as less or no need to worry about level setting, at least after you've done the due diligence to make sure your recording setup is not going to exceed the actual input dynamic range of the recorder. And some disadvantages, such as having to store significantly larger files for the same length recording and sample rate, and those files not being able to be played in simple file players prior to doing some post processing and conversion back to fixed point wavs or lossless files.
-
^ Excellent lecture on the subject, thanks!
It seems to me that 32 bit float with dual converters particularly suits devices such as the Zoom M2 and M3 "MicTrak" recorders which only offer their own built in mics as the input source. This means the circuit designers can optimise the whole chain within the device on the basis of knowing in advance what signal levels the mic and its analog preamp will produce. The M2 claims 135dB "maximum sound pressure input" which is - I think - about as loud as any device will handle. And obviously the a/d converters will not distort if that SPL is encountered, nor I guess any preceding stages, as the mic output at that SPL input is a known figure to the designers.
Someone somewhere here did suggest that storing single a/d converter output in 32 bit float form could have the advantage that it would be useful to have that recorded format to hand when processing the audio within the device. The humble H2e and H1e devices that do not claim (as far as I can see) dual a/d converters do have the ability to remix, normalise and convert to other formats within the device (if the user is desperate) and so perhaps that's the rationale - apart from marketing - in using 32 bit float from their (apparently) single converters.
I have not seen references to the use of single a/d converters in marketed "32 bit float" devices anywhere apart from these forums. Have we discovered something?!
-
It seems to me that 32 bit float with dual converters particularly suits devices such as the Zoom M2 and M3 "MicTrak" recorders which only offer their own built in mics as the input source. This means the circuit designers can optimise the whole chain within the device on the basis of knowing in advance what signal levels the mic and its analog preamp will produce. The M2 claims 135dB "maximum sound pressure input" which is - I think - about as loud as any device will handle. And obviously the a/d converters will not distort if that SPL is encountered, nor I guess any preceding stages, as the mic output at that SPL input is a known figure to the designers.
Somewhat ironic that if the specifications of what is bing input is known to the the designers (such as is the case with the device's own built-in mics) they can easily setup internal gain staging to use multiple ADCs but then write the output from that to a 24bit fixed point file without requiring any level setting in the same way as a 32bit float file. That content will fit. It's dealing with a wider range of potential external inputs that gets somewhat trickier.
For folks doing production work on the cheap, a 32-bit float output may fit into their file management and post-production workflow a bit better, and that's an audience that is a focus for manufacturers, not music tapers! Fortunately it works just fine for us too. I balk a bit at the inefficiency of the unnecessarily increased storage size, the misleading advertising claims, and the potential loss of gain-staging skills by future recordists, but it pretty much works just fine so I shouldn't complain too much. Just takes a bit of time and discussion to unearth the actual truth of it.
Use it if you have it and like it and make great tapes.
-
I think that now 32 bit float is simply the norm in any newly released audio recorder - even wireless mics! Without it many potential purchasers will regard the device as old tech.
-
Certainly for consumer gear.
-
Since your DAW almost certainly upsamples to 32-bit for processing, there is no benefit to converting a single 24-bit ADC output to 32-bit floating point in the recorder. Just normalize to -1 dB (or -.1 dB or whatever) in the last processing step. Any rounding errors will be so minuscule as to be completely inaudible in virtually any recording situation.
25% smaller file size is a minor benefit.
As someone who has just gotten in to the whole 32 Bit float thing, I am wondering...are there any drawbacks to using a recorder with it?
25% larger file size is a minor drawback.
-
Given the size of video files these days, even 32 bit float audio files are not that much of a big deal. Or large deal. And of course the colossal size of inexpensive memory cards is another factor. I think the humble H2e can handle a 2TB card. That equates to a 32 bit float stereo recording of... considerable length.
-
Since your DAW almost certainly upsamples to 32-bit for processing, there is no benefit to converting a single 24-bit ADC output to 32-bit floating point in the recorder. Just normalize to -1 dB (or -.1 dB or whatever) in the last processing step. Any rounding errors will be so minuscule as to be completely inaudible in virtually any recording situation.
25% smaller file size is a minor benefit.
As someone who has just gotten in to the whole 32 Bit float thing, I am wondering...are there any drawbacks to using a recorder with it?
25% larger file size is a minor drawback.
I don't worry about size too much, I compress my original files using WAV Pack for archiving. Those are then transferred to an external HD, which are fairly cheap these days.
-
Tascam claim in a reply to my question to them on YouTube that their latest recorders must use dual a/d converters because that is the only way to get 32 bit float output. This is in conflict with Zoom who state that at least one of their devices does use a single converter to output 32 bit float from certain inputs of the device.
I have fired off this question to Tascam support in the USA -
"I am investigating whether recent handheld recorders by various manufacturers do or do not use dual a/d converters for 32 bit float output. This investigation began when another manufacturer stated that their 32 bit float device did not use dual converters on one of the inputs of their 32 bit float only device. That caused me to check the published specs for other devices, and there is a clear pattern that lowest cost 32 bit float devices do not claim to use dual converters. Publicity for your new 32 bit float DR-07XP device does not mention dual converters. But when asked about this on YouTube your representative states that it does, as they say 32 bit float cannot be achieved with a single converter (I disagree - and so does another manufacturer). Could you resolve this matter? On other Tascam devices, dual converters are highlighted in publicity. So if the DR-07XP does have dual converters, why not say so in the publicity and manual? Thanks for any info you can supply."
I will post any response I get here in due course...
-
Tascam doesn't seem to be mincing words with their 32bit primer: https://tascam.com//us/feature/32-bit_float#:~:text=In%20a%2032%2Dbit%20float,low%20gain%20ADC%20does%20not.
If dual 24bit ADCs is the way Tascam explains how they are able to produce a 32bit output, I would be shocked if there is a superior process they are choosing not to use. It makes zero sense from a product marketing perspective for a company as established as Tascam to broadcast that they aren't using a 32 bit standalone A/D unless such a thing doesn't exist. Just applying my own logic, I don't know this for a fact.
-
Bad link for me, an extra backslash. This seems to work:
https://tascam.com/us/feature/32-bit_float#:~:text=In%20a%2032%2Dbit%20float,low%20gain%20ADC%20does%20not
I am curious whether the little Zoom flaw from the deleted YT video last year is now a non-issue for all manufacturers, or if it was ever a real problem.
I also wonder if the Sound Devices patent affects other manufacturers implementations of 32 bit float, or if SD is going to start lawsuits.
-
^ They are lying to you.
This is demonstrably not true: "After you've recorded audio in 32-bit float format, any quiet sections of your recording can be increased in volume in your DAW or NLE software without adding any noticeable noise. And loud sections that would have clipped a 24-bit or 16-bit recorder can be reduced in level, without any digital distortion or clipping."
The analog input dynamic range capability of the recorder, which ranges from it's noise floor up to clipping, is determined by it's analog input capabilities, not the digital file format. Anyone can record a signal which is low enough in level to encounter the recorders analog input noise floor, which will then become apparent when the level is raised in the DAW. And as we have discussed here at TS, an overly hot input which exceeds the capabilities of the analog input will distort and clip. Either of those things can be easily accomplished. It is true that within those limits no gain adjustment is needed, and that's really great. But the limits of the real-world analog input dynamic range capability of the recorder is orders of magnitude smaller than what they are inferring from the 32bit float format and are really not that different than what we already had by correctly setting input gain manually. Not having to make that gain adjustment when recording (as long as the signal fits the analog input stage limits) is the real advantage. Wish they could be honest about it.
Here's an easy indicator of marketing bullshitism. Any time you come across an illustration like the one below in marketing materials, stop reading:
-
Really, it doesn't matter to us how many ADCs or what the magic sauce inside is. What does matter to us are the limits of the real-world analog input dynamic range capability of the recorder. I really wish they would honestly talk about that. I see a lot of smoke and mirrors talking around this essential point by manufacturers and find it shameful.
-
Really, it doesn't matter many ADCs or what the magic sauce inside is. What does matter to us are the limits of the real-world analog input dynamic range capability of the recorder.
You're right that there's a lot of marketing-nonsense spread by the companies. However, I do believe that multi ADC (if implemented properly) can and will broaden the analog dynamic input range capability. A proper implementation will not consist of just two or more ADC's, but also two or more analog input paths, each designed to handle their specific part of the total device's analog dynamic range.
It would be great if the manufactorers would be more clear how each of their devices implement the 32bfp technology, so we can separate the 'true 32bfp devices' from those who are merely 32bfp because that's the format they use to store the recordings.
-
Really, it doesn't matter many ADCs or what the magic sauce inside is. What does matter to us are the limits of the real-world analog input dynamic range capability of the recorder.
You're right that there's a lot of marketing-nonsense spread by the companies. However, I do believe that multi ADC (if implemented properly) can and will broaden the analog dynamic input range capability. A proper implementation will not consist of just two or more ADC's, but also two or more analog input paths, each designed to handle their specific part of the total device's analog dynamic range.
It would be great if the manufactorers would be more clear how each of their devices implement the 32bfp technology, so we can separate the 'true 32bfp devices' from those who are merely 32bfp because that's the format they use to store the recordings.
Responding to the bold statement: You can't increase the analog dynamic range with multi-ADC implementation. As Gutbucket explained, the analog input stage is the limiting factor no matter how many taps off of it go to however many ADCs. The devices with greater dynamic ranges are the ones that have outstanding analog stages at the front end. The analog circuitry is also one of the main places where the cheaper brands save money, and they distract you by telling you about the enormous digital dynamic range even though the analog input stage will never get you anywhere close to utilizing that enormously tall and deep digital bucket.
Case in point: The long-discontinued Sound Devices 788T has an input dynamic range of 123 dB, whereas the Zoom F8N PRO is 113 dB (both A-weighted). The 788T has a single ADC feeding a 24-bit fixed container whereas the F8N PRO has two auto-ranging ADCs feeding a much larger (numerically) 32-bit float container. Yet the old 788 gives you 10 dB better dynamic range, and that's because of its higher-quality input stage.
In recent years the spec sheets have gotten quite inconsistent with different manufacturers, making things difficult to compare.
- Zoom does not specify A/D dynamic range for any of their recent 'Essentials' recorders, but we know through other specs they likely reused the old, noisy H-series preamps so the dynamic range is going to be rather limited.
- Sound Devices specs the A/D dynamic range for the 888 at 120 dB, and for the MixPre-10 II at 142 dB minimum. I find it hard to believe that the much cheaper MP10II has dynamic range that much better than the 888, even with multiple ADCs.
- Tascam specs the Portacapture X8 A/D DR at only 113 dB, but the FR-AV2 at 132 dB or higher. But if you look farther down, the FR-AV2 has far higher input noise.
-
When these devices first started coming out, I was under the mistaken impression that each signal path had its own preamp - because in my experience, the difference between a shitty preamp and cleaner one was very noticeable. Adding +40 dB with the internal preamp of a Zoom H1 sounds a lot more noisy than dialling down the internal gain and using a better external preamp with the same amount of gain. But if you only need something like +15 dB for the signal to reach a healthy level, then the performance difference between internal and external preamps will be much closer in absolute terms. You won't hear any preamp noise either way.
So I thought these devices worked like this: one signal path would have high preamp gain (say, +40dB) fed into an ADC that then offset this by -40dB when writing it to a 32-bit float container, and a different path with little to no preamp gain, fed into an ADC with little to no offset. Combine both, and you have the advantages of high clean preamp gain for the quiet parts without worrying about overloading the device with your peaks.
But as far as I know, this is not how these devices work. A Zoom F3 appears to have the preamp gain fixed, which initially didn't make sense to me, since I always heard people say it was a conservative amount of gain. But this architecture only makes sense to me if the gain is permanently set to high (say, +40 dB), with one of the AD converters being capable of withstanding a really hot signal. Does this sound right? This would allow the device to actually take advantage of the clean preamp (which a conservative amount of gain wouldn't permit!).
What I don't fully understand is: if this is the case, is a preamp necessary at all here? If you can tweak the sensitivity of the ADCs, can't you just make the one intended for quiet sounds really sensitive, skipping the preamp stage altogether? Or is that not possible because you'd bump against some noise floor that is inherently higher than that of a quality preamp?
-
All I can say is that my F3's preamp sounds pretty darn quiet, even when I greatly boost the volume of very quiet passages in post
-
All I can say is that my F3's preamp sounds pretty darn quiet, even when I greatly boost the volume of very quiet passages in post
Same here! I am just interesting in understanding exactly how it accomplishes this.
-
This is the only true 32 bit fixed integer audio interface I can find. There are 8 more bits but it still clips signals above 0 dBFS just like 24 bit.
https://www.merging.com/products/interfaces/merging+anubis
To me the fact that cheap recorders and wireless mic receivers are all moving to 32 bit float tells me this is feature for convenience and that's it.
-
my sonosax is also 32bit integer, and has a DNR of 135db. The only analog gain stage available on this is a +20db selectable gain. It can write safety tracks as well.
-
Rairun- You've got it right. There are two or more parallel analog circuit paths (call them preamps if you like) into separate ADCs, each analog path having a different fixed gain, and the output of those ADCs are combined into a single output. That's essentially it. In modern properly designed "32bit recorders" the combining happens in real time while making the recording. But the same can be done afterward if the different ADC outputs are saved separately. That's what TheJez's cool SafetyTrackMerger (Windows + MacOS) (https://taperssection.com/index.php?topic=206443.0) tool does for older "pre 32bit float era" recorders which have a feature that records a "safety track" at a lower gain simultaneously with the "primary track". They use essentially the same method of parallel analog circuit paths with different gains feeding multiple ADC channels. Biggest difference while recording is that you'd still manually set the gain for the primary channel and usually the offset gain on the safety track as well. With modern 32bit recorders that gain setting is determined by the manufacturer in the design phase and locked down. SafetyTrackMerger then does the combining part afterward on the computer.
TheJez is spot on when he says "multi ADC (if implemented properly) can and will broaden the analog dynamic input range capability". No question about that. Its made possible by to the parallel circuit paths.
Voltronic is too when he confirms what I was saying about the dynamic bottle-neck being the analog input stage, where improvements in circuit performance are more costly. The SD788 he mentions (now old!) and especially Grawk's Sonoxax are an examples farther along that axis of input stage quality.
If twice as many input stage paths or more running in parallel are needed that gets more costly. But combining a number of less costly paths, each of which in isolation doesn't provide premium performance, may be able to achieve premium performance in aggregate. And that's a reasonable way to increase performance while keeping costs manageable. Use multiple premium input circuits and you can potentially increase performance dramatically in a costs-be-damned type of design.
But regardless of how the sausage is made, the proof is in the eating. How this stuff works internally is interesting, but as I taper I don't really care as long as the device delivers the performance that is promised. The problem as I see it is that "promise". The performance of these inexpensive 32bit recorders does not deliver the inferred (and sometimes outright) promises made in the marketing. And measurements of what the actual real world performance is tends to be buried, withheld or obscured.
Long live independent testers with the capability to run the tests that can shed light on what's actually being delivered.
/Data not promises.
-
Right! I'm particularly interested in the architecture of the Zoom H1essential because it's suspected that they are using the same preamp as the Zoom H1 and H1n. I'm very familiar with the Zoom H1, which has a gain scale of 0 to 100, where <16 is digital attenuation, 16 is no gain, 17 to 36 is digital gain in steps of +0.4 dB, 37 is fixed (and very clean!) analogue gain of +13.5 dB, and >37 is digital gain in steps of +0.4 dB (on top of the +13.5 dB analogue gain). In other words: Zoom H1 and H1n users have a binary choice between +0 and +13.5 dB analogue gain. There is no reason to use anything other than levels 16 and 37.
At level 37, a very loud rock show would clip with the Zoom H1 internal mics, so you need to set it to 16. In my experience, a pair of CA-11s + battery box into a Zoom H1 would never clip at level 37, just because the CA-11s are less sensitive. So I can totally see the Zoom H1essential using less sensitive mics, fixing gain at +13.5 dB (which is clean, if that's all the gain you need) and writing this to a 32-bit float file. Or they could simply use the same mics and skip the preamp. When used with internal mics, either option would sound noisier than the Zoom H1 when recording quiet sounds, but it wouldn't clip with louder ones. When used with a less sensitive external mic, the former would sound as good as the Zoom H1, and the latter would sound noisier, with no risk of clipping either way. If used with a more sensitive external mic, it'd either clip more easily or have more noise during quiet passages.
Personally, if that's the case, I'd much prefer the Zoom H1 over the Zoom H1e. All the H1e would be doing is pretending you don't have to set levels anymore, by simply setting it to a non-ideal level for you!
This is assuming, of course, that the Zoom H1essential DOES NOT have two ADCs. It's very possible that it does, despite the lack of info in the specs. Assuming the same preamp as the H1/H1n and dual ADC, the device would give you the same performance as the Zoom H1 for quiet sounds (good if you only need +13.5 dB; if you need more, you're screwed), but with more headroom.
-
OK, the game is up. I have just received the following email from Tascam USA customer support -
"Hello Peter,
Our engineering department has sent the following information...
DR-07XP/DR-05XP are SINGLE ADC.
Thank you for your interest in TASCAM products."
So this confirms my suspicion that low cost 32 bit float recorders which do not claim dual A/D converters do not have them. But they are happy to allow consumers to assume - like the Tascam marketing department person did when replying to me on YouTube- that dual converters are always used in 32 bit float devices. These single A/D devices are probably just expressing their 24 bit output as 32 bit float, to gain advantages in the final result which are very hard to identify.
I have several 32 bit float devices, and the only one which does not claim dual converters is the H2essential. It's fine for what it is, at the price they sell it for. The lack of recording level control simply means that they have preset the level to cover the worst possible loudness scenario and the results are ok imho. But given that all discussion of 32 bit float recording methods on the internet etc has described its benefits as being due to the use of dual converters - or more - it seems to be that companies now releasing single converter devices should include this fact in their published specifications which customers can see, and consider, before purchase.
-
I have several 32 bit float devices, and the only one which does not claim dual converters is the H2essential. It's fine for what it is, at the price they sell it for. The lack of recording level control simply means that they have preset the level to cover the worst possible loudness scenario and the results are ok imho. But given that all discussion of 32 bit float recording methods on the internet etc has described its benefits as being due to the use of dual converters - or more - it seems to be that companies now releasing single converter devices should include this fact in their published specifications which customers can see, and consider, before purchase.
Yeah. At that point, they aren't removing the need for gain control, they're just removing the OPTION to control gain in different circumstances, noise floor be damned.
-
Even 'Deepseek' is deeply (!) confused on this subject - partly because it will be deriving its response from human stuff on the internet.
"You're absolutely correct that a single analog-to-digital converter (ADC) operating at 24-bit resolution cannot inherently capture more than 24 bits of dynamic range, and expressing its output in 32-bit float format doesn't magically increase the ADC's resolution. However, there are still practical benefits to using 32-bit float format even when the ADC itself is limited to 24 bits. Let me clarify the nuances:
---
1. **Preservation of Headroom and Avoiding Clipping**
- While the ADC itself is limited to 24 bits, the 32-bit float format allows for **extended headroom** in the digital domain. This means that if the ADC's output exceeds 0 dBFS (full scale), the excess can be represented in the 32-bit float format without clipping.
- In a fixed-point 24-bit system, any signal exceeding 0 dBFS would clip irreversibly. In 32-bit float, the extra "room" above 0 dBFS allows you to recover and adjust levels in post-production, even if the ADC's output was technically clipped.
[I think that is wrong in every respect]
---
2. **Simplified Gain Staging**
- With 32-bit float, you don't need to worry as much about setting the input gain precisely. If the signal is too hot or too quiet, you can adjust it in post-production without degrading the audio quality.
- In a 24-bit fixed-point system, you must carefully set the gain to avoid clipping or excessive noise, which can be challenging in dynamic recording environments.
[Wrong again, if the digitally clipped output from an overloaded 24 bit converter is stored in 32 bit float]
---
3. **Post-Processing Flexibility**
- Even though the ADC's resolution is 24 bits, the 32-bit float format provides a much larger "container" for mathematical operations during post-processing. This reduces the risk of rounding errors or quantization noise when applying gain changes, EQ, compression, or other effects.
- In a 24-bit fixed-point system, repeated processing can introduce cumulative errors, whereas 32-bit float minimizes these issues.
[OK, I'd agree with that]
---
4. **Future-Proofing and Workflow Consistency**
- Using 32-bit float as the recording format ensures consistency with modern workflows, where 32-bit float is increasingly the standard for digital audio workstations (DAWs) and post-production.
- Even though the ADC's resolution is 24 bits, recording in 32-bit float avoids the need to convert formats later, which can introduce unnecessary complexity or artifacts.
[That I'd broadly agree with too]
---
5. **Dynamic Range Utilization**
- While the ADC's dynamic range is limited to 24 bits, the 32-bit float format ensures that the full range of the ADC's output is preserved. This is particularly useful for capturing very quiet signals, as the 32-bit float format can represent extremely small values without losing precision.
- In a 24-bit fixed-point system, very low-level signals might be lost in the noise floor, but 32-bit float ensures that these signals are retained for potential recovery in post-production.
[Wrong if the 24 bit output from a single 24 bit converter is being stored in 32 bit float]
---
6. **Practical Example**
- Imagine recording a live concert where the dynamics are unpredictable. With a 24-bit fixed-point system, you might set the gain conservatively to avoid clipping, but this could result in quiet sections being too close to the noise floor.
- With 32-bit float, you can set the gain more aggressively, knowing that any clipping can be recovered in post-production, and quiet sections will still be preserved with full precision.
[Simply not true without multiple converters making more use of the 32 bit float than a single converter]
---
### Conclusion
While it's true that a 24-bit ADC cannot capture more than 24 bits of dynamic range, expressing its output in 32-bit float format provides significant practical advantages in terms of headroom [wrong], post-processing flexibility [maybe], and workflow efficiency [true if post-processing in 32 bit float, but not really a big deal]. The 32-bit float format acts as a "safety net" for the ADC's output, ensuring that the full potential of the 24-bit resolution is utilized without the risk of clipping or loss of detail [Not true]. This makes it a valuable choice for professional audio recording, even when using a single ADC. [I'd say that a single converter system writing 32 bit float output might make the designers think carefully about end-to-end gain staging but that's probably the only good thing]
-
There a point at which my brain apparently isn't translating all this information very well, but I reckon that's no one's problem but mine.
This may or may not relate to what some of you folks have been explaining, but on the user end, it's been frustrating over the past few years, that some audio applications automatically default to 32-bit, and there doesn’t seem to be any way to discern what the original format was.
Since some people have been encoding at variable formats such as 24/44.1, 24.48, 24/96, I can’t be sure any more when I get files what they were in the first place, and half the time, the people who sent me the files don’t really have any idea either, since they either got them from someone else, or were just following instructions they didn’t really understand.
I'm attaching a few screen snaps of files which I wouldn't know how to decipher if I hadn't been provided with the information first.
-
4. **Future-Proofing and Workflow Consistency**
- Using 32-bit float as the recording format ensures consistency with modern workflows, where 32-bit float is increasingly the standard for digital audio workstations (DAWs) and post-production.
- Even though the ADC's resolution is 24 bits, recording in 32-bit float avoids the need to convert formats later, which can introduce unnecessary complexity or artifacts.
[That I'd broadly agree with too]
Apple Logic Pro has been 64 bit internal since 2013...
-
Groundhog420, try this, which I have used for years without my PC being hijacked -
https://mediaarea.net/en/MediaInfo
Or maybe you are using an Apple device - well, it may be useful to someone anyway.
-
4. **Future-Proofing and Workflow Consistency**
- Using 32-bit float as the recording format ensures consistency with modern workflows, where 32-bit float is increasingly the standard for digital audio workstations (DAWs) and post-production.
- Even though the ADC's resolution is 24 bits, recording in 32-bit float avoids the need to convert formats later, which can introduce unnecessary complexity or artifacts.
[That I'd broadly agree with too]
I don't agree with that.
I don't buy the workflow argument. I can import whatever file format into the DAW without any extra steps.
No way fixed point WAV format is going away despite adoption of 32-bit float. It will remain supported.
Converting from 24 bit to 32bit float is not complex at all and introduces NO artifacts. Just plain wrong on that account.
Ugh with AI's regurgitating erroneous human output. One of the biggest problems with AI as I see it is an underlying misjudgement of presumed authority - the assumption by users that what it spits out is actually true! I see that leading to dangerous and misleading simplifications in society at large as it becomes more widely adopted.
-
Ugh with AI's regurgitating erroneous human output. One of the biggest problems with AI as I see it is an underlying misjudgement of presumed authority - the assumption by users that what it spits out is actually true! I see that leading to dangerous and misleading simplifications in society at large as it becomes more widely adopted.
I agree whole heartedly. Bob
-
Ugh with AI's regurgitating erroneous human output. One of the biggest problems with AI as I see it is an underlying misjudgement of presumed authority - the assumption by users that what it spits out is actually true! I see that leading to dangerous and misleading simplifications in society at large as it becomes more widely adopted.
I agree whole heartedly. Bob
100%.
And there's a rule of thumb which lays it out...
https://thebullshitmachines.com/lesson-2-the-nature-of-bullshit/index.html (https://thebullshitmachines.com/lesson-2-the-nature-of-bullshit/index.html)
As we will see, LLMs are powerful tools. But they also make it easy for people to mislead us by accident, or on purpose.
The Bullshit Asymmetry Principle — also known as Brandolini's Law after the computer programmer who proposed it — is one of the most important principles in bullshit studies.
Brandolini's Law: The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.
I think we have an intuitive understanding that this is the case... It's why the Galt Gallop works so well (just keep lying and they'll never catch up to you)
-
:coolguy:
Thanks for that link!
-
Well - ok, I take your point(s). But the nub of my post was to indicate the level of confusion on this subject - even in the world of AI.
Meanwhile Tascam have posted a remarkable response to my questions on YouTube about whether their new recorders, which use 32 bit float but do not claim dual converters, do or do not have single or dual converters. Their "Product Specialist", who claims to know more than anyone in the company, states categorically that a single ADC cannot write its output in 32 bit float format, and therefore the devices must use dual converters. He (or she) also doubts my statement that their engineering department states equally categorically that the devices have single converters. The tone of the response is borderline offensive. I have pasted into my response the text of the email I received quoting their engineering department and suggested that someone further up the company should explain the company's employees diametrically opposite statements and apparent lack of product and technical knowledge.
The confusion in the company and the manner of their communications makes me very reluctant to buy one of their products again. Not that I planned to. I will post the entire exchange in the thread here related to these new models.
-
The confusion in the company and the manner of their communications makes me very reluctant to buy one of their products again. Not that I planned to.
🗑
#TrashCan
-
Tascam and zoom both have similar strategies. They make some consumer focused gear which is mediocre and some pro focused gear which is significantly better. If you can afford to step up do it. If not, you can still make good recordings, it just requires more of you as the recordist.
-
From what I've learned, the generation of recorders that came out around 2017 use dual converters but instead of writing the lower level safety track, a circuit after the ADCs compares samples in real time and writes the ones with the best dynamic range, discarding the rest. I speculate 32bit float was just the next step in evolution by using the dual ADCs to write the larger range of 32bit float instead of having to keep within the scale of 24 bit. It also might be cheaper, which could explain all the budget devices that use 32bit float now.
-
Just putting on record here for posterity in this 32 bit float discussion, that Tascam seem now to have sorted out the confusion within their own company as to (1) whether 32 bit float files can be created from a single ADC at all and (2) whether their latest launched devices do indeed have only one ADC, creating 32 bit float files that have no better dynamic range than 24 bit files. They have deleted some of their previous comments on the YouTube channel and replaced them with a new comment which reads as follows -
"32-bit float point recording and Dual ADC technologies do not necessarily go hand in hand. Dual ADCs give a far superior dynamic range, sure. But with a single ADC, 32-bit float recording is still possible and the clip gain can still be adjusted as such without any degradation. The only real difference between one or two ADCs is the disparity in dynamic range."
So now it is clear as I and others suspected, that recent low cost "32 bit float" recorders are using a single ADC and the dynamic range enhancement gained from two ADC's is not being achieved. Given the confusion within Tascam itself, consumers are going to be confused too. Hmmm.
-
Yak....It's becoming more attractive to go back to using cassette tapes. I wonder if my old D5s still work...... ;)
-
Yak....It's becoming more attractive to go back to using cassette tapes. I wonder if my old D5s still work...... ;)
A lot of rock and roll barely has 50 dB dynamic range, cassettes are awesome for that! (says morst while listening to a 1990 tape on the Nak 582 > F3 > WAV @32f96)
-
I have quickly made a YouTube video explaining this whole thing. (32 bit float audio from single ADCs not being mentioned by manufacturers).
Ironically I used the sound from the video camera, set to stereo, placed close to a noisy PC - hardly an advert for my skills with audio. But I'm short of time. That's my excuse, officer.
https://youtu.be/pAyaWH-WX-M?si=RlfKDhlKs6eY35aZ
My channel is not monetised so I am not earning from that link.
-
Thanks @Ozpeter and others for making us aware about the issue regarding single/multi ADC devices. With the risc of complicating things even more: I guess a device having two (or more) ADC's doesn't necessarily make it a very good recorder! Of course the quality of the analog pre-amps matters a lot, but also something that didn't got much attention at all here: The quality of the 'combining algorithm'! There are a zillion ways to do it, all with their pro's and con's. Even 'good' algorithms may work well most of the time but might introduce artifacts in certain specific situations. I'm not saying that they do, just saying that they might. And even if they do, it might well be unaudible. We simply don't know. What we do know is the patent by SD that describes a possible way to combine multiple ADC's into a single 32bfp recording. It doesn't even necessarily mean that SD actually uses the algorithms they describe in their patent. It probably does mean that other manufacturers will be using different ways to combine, simply to avoid patent infringement. (Or maybe they do use the SD ideas and won't tell, or maybe they are paying license fees to SD, we simply don't know...)
Bottom line:
- We now know for sure that a recorder that has 32bfp doesn't automatically also has multi-ADC.
- I guess that even if a recorder has multi-ADC, it doesn't mean it's a good recorder.
I'm almost wishing the manufacturers would have sticked to 24-bit main/safetytrack recording, so the users would still have control of the analog gain (to optimize S/N ratio, with extra safety margin due to the 2nd ADC) and the option to combine during post using the algorithm of their choice with the settings of their choice (to fix the loud clipping parts and/or improve S/N in quiet parts, whatever is needed).
-
Yup, the converters are kind of at the end of the chain and what comes before is still (or more) vital as ever. And yes, how those multiple converters combine their output can make a difference - some people claim to hear artifacts (but I don't).
You know, it seems to me that that there should be a law against describing any recorder as 32 bit float. It should be described as single converter or dual converter or even quad converter. The output format is dictated by the number of converters. It is, of itself, not of great interest. When the Tascam guy swore that you can't write 32 bit float audio from a single converter, actually he got the whole thing the wrong way round. You can't write 24 bit audio from dual converters (well, not without throwing out data, which would render the dual converters pointless). Maybe we should stop talking about 32 bit float recording here, and only talk about multi-converter recording. Companies required to describe their recorders as dual or single converter devices wouldn't be able to pull the wool over consumets' eyes, like we're seeing recently. Single or dual (multiple) converters make a difference. Output formats don't. They are simply a natural consequence of what comes earlier in the chain. And that's how they should be described.
I feel another video coming on... :o
-
You know, it seems to me that that there should be a law against describing any recorder as 32 bit float.
Haha, I get what you're saying. But I don't have a general problem as such with recorders being described as '32 bit float'. I do think there is a benefit with 32bfp storage when recorders can do post-ADC stuff like equalizing or basically anything that could change the amplitude of the samples. Personally I just want a recorder that just records, but if post-ADC features (either in realtime or not) are present on a recorder (and people actually using it), then I'd be happy with 32bfp storage, regardless if it's a single or multi ADC device. It's just that a device being 32bfp doesn't say anything about the recording quality. Nor does single or multi-ADC. Unfortunately the manufactorers are abusing the 32bfp buzz to convince us into buying new equipment. Just like we had to buy 3D and curved TV's. (Remember those?) Hopefully manufactorers will become/remain clear and honest about what exactly they are trying to sell to us, so we can judge for ourselves if we need their stuff or not. And we need to be alert to not blindly fall for their marketing tricks!
Tonight I'll be recording my first loud gig with my new Tascam FR-AV2 and enjoying the dual-ADC 'no need to worry about the record level' feature. Looking forward to it! :headphones:
-
Multi ADCs do allow for a wider dynamic range to be recorded. But indeed, they don't help with anything before the ADC, except perhaps the designers will couple a dual ADC system with upstream components which can make that wider dynamic range worthwhile. Well, possibly...
-
Multi ADCs do allow for a wider dynamic range to be recorded. But indeed, they don't help with anything before the ADC, except perhaps the designers will couple a dual ADC system with upstream components which can make that wider dynamic range worthwhile. Well, possibly...
Yeah, or they can implement two crappy cheap analog signal paths feeding two crappy ADC's, combining them with a crappy algorithm and still wave the dual-ADC 32bfp flag to make it sell :(. But indeed, on the bright side: The multi-ADC technology, combined with 32bfp does open up options to further improve recording quality at a reasonable pricing, and fortunately some devices are actually doing this!
-
You can't write 24 bit audio from dual converters (well, not without throwing out data, which would render the dual converters pointless).
You can in fact write 24-bit audio from dual converters! In my experience, with the playback system I use and the loudest listening volume I find comfortable at 0 dBFS, you can increase levels by 50-60 dB before quantization errors become just barely audible (you need to increase levels by ~118.5 dB to normalise quantization errors to 0 dBFS). Meaning: if you record a sound at the threshold of perception (defined by me here as -60 dBFS), you can normalise it to 0 dBFS, and the quantization errors will be insignificant - they will be raised to a level that would be barely perceptible in absolute silence (i.e. the errors are imperceptible when the quietest sound actually recorded has been normalised to 0 dBFS). This is to say 24-bit files have an incredible amount of dynamic range and give you a lot of leeway for editing!
In other words: manufacturers could easily use a dual ADC setup and then assign specific level ranges to write the results to a 24-bit file. Assign +4 dBu to 0 dBFS, write the data of the signal path with less gain to the -20 to 0 dBFS range, and the data from the path with more gain to -20 dBFS and under. You'd run into the microphone and preamp noise floors before bumping into quantization errors.
Don't get me wrong, at that point, I do find it better to write to 32-bit float! But you can absolutely see the benefits of dual ADC with 24-bit files, because this design is meant to address noise in the analogue signal chain, not digital noise.
-
Interesting! But is anything commercially available?
-
Interesting! But is anything commercially available?
I don't have one, but I believe the MixPre10 II engages all 3 ADCs when set to 24-bit. AFAIK you are free to set the gain (so you can clip digitally), but the multi-ADC design does its thing to reduce the analogue noise floor of the quiet passages.
-
Someone correct me if I'm wrong, but I think Sound Devices had multi-ADC designs even before their products had a 32-bit float option at all.
-
No reason a multi-ADC recorder can't be designed to save in whatever format is desired, including 16bit mp3 or whatever. Zoom F8N_Pro can be set to record 24bit fixed, and Zoom has confirmed that the recorder is using the switching dual ADC signal path when doing so. I believe the SD MixPre's can as well.
As for quantization errors, that problem was solved decades ago. Application of dither eliminates quantization errors. Just apply it regardless of bit depth. It's an easy fix. Manufacturers apply dither when saving to 16bit, yet for some reason don't bother when saving to 24bit. I supposed they decided back when that the recorded dynamic range of 24 bit was large enough that any quantization noise at such low levels didn't matter when folks were setting gain reasonably. Or maybe they determined that there was already enough random noise in the signal noise at those levels to effectively self-dither. If manufacturers still consider not applying dither at 24bits to be a legitimate design choice, while now simultaneously claiming that 32-bit float "solves the quantization problem", I smell rotten fish.
Again, whenever you see marketing materials use those stair-step drawings, best to stop reading right there because you are being lied to.
On the 'combining algorithms' TheJez mentions-
They are clever, they are cool, they are useful. But philosophically what bugs me deep down about all this is that the quality of a digital recorder used to be defined by absolute linearity within it's bandwidth - being measurably bit perfect in/out within that range. Use that available bandwidth anyway you like and be assured that whatever you put in you get back out again. That includes sound-design folks recording ultrasonics and pitching them down into human hearing range, which is apparently where multi-ADC combining artifacts of some recorder were noticed. This deeper problem is not about us not noticing or caring about the artifacts / non-linearities, it's about the shift from truly predictable linear transparent systems to ones which only need be perceptually transparent. I appreciate the useful extension of dynamic range.. when it is done right and effective. But when adopting the tech required to do that sacrifices measurable, linear, predictability I don't like where it can be too easily taken. It all too quickly becomes a marketing smoke and mirror game rather than one of measurable linearity. Reminds me of the mp3 perceptual codec thing all over again in sacrificing true linearity for perceptual psychoacoustics.
-
Someone correct me if I'm wrong, but I think Sound Devices had multi-ADC designs even before their products had a 32-bit float option at all.
The first generation MixPres have multiple ADCs, but no 32-bit float. I (accidentally) tested this out once and was able to add more than 30 dB of gain with no audible noise.
The high-end SDs (833/888/Scorpio) also used multiple ADCs without 32-bit float since their release (2019). I think they only implemented 32-bit float via a firmware update last year (May 2024).
-
Multi ADCs do allow for a wider dynamic range to be recorded. But indeed, they don't help with anything before the ADC, except perhaps the designers will couple a dual ADC system with upstream components which can make that wider dynamic range worthwhile. Well, possibly...
Yeah, or they can implement two crappy cheap analog signal paths feeding two crappy ADC's, combining them with a crappy algorithm and still wave the dual-ADC 32bfp flag to make it sell :(. But indeed, on the bright side: The multi-ADC technology, combined with 32bfp does open up options to further improve recording quality at a reasonable pricing, and fortunately some devices are actually doing this!
I don't recall if the Zoom F3 was the first commercial device to combine dual ADCs, floating 32 bit tech, and super quiet pre-amps for a few hundred dollars, but the ones I own do a great job with no audible artifacts, even when I amplify quiet passages beyond all reasonable limits in post (perhaps to hear a musican whisper the song's title to bandmates).
-
I've now been told by B&H that the very popular DJI Mic 2 wireless mic that records in 32 bit float is, as I suspected, a single ADC device. Given that I can't find any 32 bit float recording wireless mic that does claim dual converters, I suspect that all of them are pointlessly writing 24 bit data into a 32 bit float container.
I do feel that this whole "fake" 32 bit float thing is something of an industry scandal. And I've been posting about it all over the internet :alert:. I do hope someone here will tell me if in fact there is a worthwhile point in using a single converter to write 32 bit float files, before I embarrass myself further!
[Edited to add that DJI themselves have now confirmed what B&H stated. They say "DJI Mic 2 32-bit float recording adopts a brand new audio encoding and recording method, which expands the recording range and effectively solves the problem of audio overexposure." Really? From a single 24 bit ADC? How?]
-
I've now been told by B&H that the very popular DJI Mic 2 wireless mic that records in 32 bit float is, as I suspected, a single ADC device. Given that I can't find any 32 bit float recording wireless mic that does claim dual converters, I suspect that all of them are pointlessly writing 24 bit data into a 32 bit float container.
I do feel that this whole "fake" 32 bit float thing is something of an industry scandal. And I've been posting about it all over the internet :alert:. I do hope someone here will tell me if in fact there is a worthwhile point in using a single converter to write 32 bit float files, before I embarrass myself further!
[Edited to add that DJI themselves have now confirmed what B&H stated. They say "DJI Mic 2 32-bit float recording adopts a brand new audio encoding and recording method, which expands the recording range and effectively solves the problem of audio overexposure." Really? From a single 24 bit ADC? How?]
I can see how using 32-bit float would be simpler there. They do have to make sure that anything under the microphone's maximum SPL doesn't clip digitally as it goes through the 24-bit ADC, BUT after the analogue signal has been converted, they also have to allow the user to set the volume (to feed, for example, a camera) - and sure, they could simply name this volume the "output volume", but maybe they felt this would be confusing and opted for it to be both. So by saving the data to 32-bit float, it allows the user to set whatever volume they want (even if they run the risk of overloading the camera they are feeding), while also avoiding clipping the mic's internal recording digitally.
-
Devices which write 32 bit float files using only one ADC (confirmed by manufacturer) -
Zoom H1 essential
Zoom H2 essential
Zoom H4 essential only uses dual ADC on inputs 1 & 2
DJI Mic 2
Tascam DR-07XP
Tascam DR-05XP
Response awaited from Rode regarding the Wireless Mic Pro
Obviously there are loads more devices out there which might write to 32 bit float from single converters, and I may get specific info on some more, but as a rule of thumb, Zoom said that as far as their product range is concerned, if there is no claim in their publicity for the device that it has dual converters, then it doesn't. I suggest extending that way of assessing devices to all other companies.
I have started a discussion on Reddit about this - simply trying to spread the word, to undo the assumption that 32 bit float means extended dynamic range from multiple ADCs - 15,000 views in 5 hours, plenty of comments - I did say in the first line that the whole subject came up from someone's post here!
-
The Reddit thread has had 22,000 reads in 11 hours so at least the subject is being kicked around widely. One response has suggested that 32 bit float from a single ADC can be implemented with clipless dynamic range being enhanced by an entirely different method than dual ADC methods.
"It is possible however to achieve the same result with a single ADC. A dual path preamp can be triggered to change its gain on the analog side if clipping occurs and the DSP could then digitally trim up the digital signal. It's a bit harder to do as you need precise timing adjustment and the analog circuitry needs to be quite precise (maybe even calibrated at manufacturing) but overall it would have the exactly the same performance for far less costs. This is how some digital desk preamps are actually made. Analog gain is stepped and digitally compensated to give the illusion of continuously variable gain."
I think I can follow the meaning of that, just. It appears to be as if the analog signal is compressed to avoid clipping on the analog side, but then uncompressed once converted to 32 bit float data. But is that a good idea in terms of the integrity of the audio? And if it is a good idea, is it just as good as using dual ADCs? And if it's just as good as using dual ADCs, why isn't it being used in the top flight devices? And why are the companies not making the thing known, like they were happy to do with endless diagrams in publicity for dual ADC devices?
I suspect I'm getting obsessed with all this... :-X Sorry!
-
I suspect I'm getting obsessed with all this... :-X Sorry!
You might be obsessed with investigating this and writing about it, but I’m obsessed with reading about it so don’t be sorry.
-
Does anyone know if the Tascam FR-AV2 uses two ADCs for the external in (as well as for the XLR in), or is it crippled like the X8 external input?
-
The TASCAM DR-10L PRO uses dual ADC for 32-bit Float recording on its single 3.5mm (1/8") TRS jack.Tascam might have implemented this also on the FR-AV2 3.5mm external input.
-
The problem with devices which seem not to have dual ADCs on all their inputs, only on some, seems to relate to the inputs being recordable at the same time to multiple tracks. With devices that can use only one input or the other at the same time, then they only need to provide one pair of ADCs to deal with it - the input will be passed to the same dual ADCs from whichever input it in use. So, if the Tascam FR-AV2 can only record to one stereo track at once, I can see no reason why its dual ADCs should not be employed on whichever input is used. But - I don't actually know. (The Zoom H1 XLR has dual converters and there's nothing in the manual to suggest that the 3.5mm input doesn't go through them - and it is rather cheaper than the FR-AV2!).
-
I don't know what you all want, but the bottleneck of all is the dynamic range of the ADC. Also double ADC's don't double the dynamic range because they all have the same ground. If not that would be an interesting schematic for studying. The maximum possible dynamic range is the range between ground and the maximum voltage, this is not necessary the same as the total voltage of the batteries, multiplying is possible, in that case your power usage is going up.
-
I don't know what you all want, but the bottleneck of all is the dynamic range of the ADC. Also double ADC's don't double the dynamic range because they all have the same ground. If not that would be an interesting schematic for studying. The maximum possible dynamic range is the range between ground and the maximum voltage, this is not necessary the same as the total voltage of the batteries, multiplying is possible, in that case your power usage is going up.
tut
I never read anyone claim that dual 24 bit ADCs doubles dynamic range. The idea is to "stretch" the dynamic range of the written file by 8 bits. I think we all know dynamic range is limited to the analog signal, with the best being about 120 db SNR.
-
The Reddit thread has had 22,000 reads in 11 hours so at least the subject is being kicked around widely. One response has suggested that 32 bit float from a single ADC can be implemented with clipless dynamic range being enhanced by an entirely different method than dual ADC methods.
"It is possible however to achieve the same result with a single ADC. A dual path preamp can be triggered to change its gain on the analog side if clipping occurs and the DSP could then digitally trim up the digital signal. It's a bit harder to do as you need precise timing adjustment and the analog circuitry needs to be quite precise (maybe even calibrated at manufacturing) but overall it would have the exactly the same performance for far less costs. This is how some digital desk preamps are actually made. Analog gain is stepped and digitally compensated to give the illusion of continuously variable gain."
I think I can follow the meaning of that, just. It appears to be as if the analog signal is compressed to avoid clipping on the analog side, but then uncompressed once converted to 32 bit float data. But is that a good idea in terms of the integrity of the audio? And if it is a good idea, is it just as good as using dual ADCs? And if it's just as good as using dual ADCs, why isn't it being used in the top flight devices? And why are the companies not making the thing known, like they were happy to do with endless diagrams in publicity for dual ADC devices?
What you describe is companding (https://en.wikipedia.org/wiki/Companding) - essentially a form of lossy compression used to increase dynamic range throughput and has been around for a long time. Compress on the way in then expand on the way out by the same amount. Dolby and DBX noise reduction schemes are examples of compansion tuned specifically to improve dynamic range of analog tape, with the more specific goal of noise reduction. Conceptually, substitute the single ADC for the magnetic tape portion of the signal chain, with the compression being done in the analog circuit path ahead of the ADC, and expansion done digitally afterward. The trick is getting the two to match identically such that the expansion exactly counters the effect of the compression, and maintaining the required design tolerance in doing so over time. It's a legitimate work around, but is not free of trade-offs. No free lunch.
-
^^ what he said
Dual-ADC method isn't that difficult and the circuitry costs not that expensive.
At this point in the evolution of the technologies, it seems the way to go.
Will that be bested?
Probably
But modifying analog stages just seems like a step backwards, and more likely to introduce artifacts which cannot be undone.
-
Solution: a compact field recorder that accepts AES42 (digital microphone) signals. No more 32FP! :)))
-
Solution: a compact field recorder that accepts AES42 (digital microphone) signals. No more 32FP! :) ))
Aaton Cantar mini has 4... too big?
https://shop.cinneshop.com/product/aaton-cantar-mini/ (https://shop.cinneshop.com/product/aaton-cantar-mini/)
Oops company went into receivership a year and three days ago...
https://jwsoundgroup.net/index.php?/topic/62483-aaton-bankruptcy/ (https://jwsoundgroup.net/index.php?/topic/62483-aaton-bankruptcy/)
-
Aaton Cantar mini has 4... too big?
https://shop.cinneshop.com/product/aaton-cantar-mini/ (https://shop.cinneshop.com/product/aaton-cantar-mini/)
Oops company went into receivership a year and three days ago...
https://jwsoundgroup.net/index.php?/topic/62483-aaton-bankruptcy/ (https://jwsoundgroup.net/index.php?/topic/62483-aaton-bankruptcy/)
Ah naturally. I still want one though. :)
I think the Sound Devices 788t accepts AES3 and AES42 signals....
-
Does anyone know if the Tascam FR-AV2 uses two ADCs for the external in (as well as for the XLR in), or is it crippled like the X8 external input?
I asked Tascam and got this reply:
“The EXT input has the same circuit path (post mic-pre) as the main inputs.”
-
Does anyone know if the Tascam FR-AV2 uses two ADCs for the external in (as well as for the XLR in), or is it crippled like the X8 external input?
I asked Tascam and got this reply:
“The EXT input has the same circuit path (post mic-pre) as the main inputs.”
Post mic-pre?
I was disappointed to discover that the 3.5mm input on my Sound Devices MixPre 3II doesn’t use the same Kashmir pre-amps as the XLR inputs. That, together with the low 2.5v PIP means that I never use it for lavalier type mics.
There’s a lot of love for the Tascam FR-AV2 as a stealth recorder because it removes the need for setting levels and a battery box. I’ve not used mine for that yet, but do hope it’s using the same pre-amp as the XLR inputs.
-
Does anyone know if the Tascam FR-AV2 uses two ADCs for the external in (as well as for the XLR in), or is it crippled like the X8 external input?
I asked Tascam and got this reply:
“The EXT input has the same circuit path (post mic-pre) as the main inputs.”
Many thanks!!
-
I have just uploaded some tests on YouTube which attempt to show how well three recorders perform when recording something very loud (rock band) and something very quiet (fly) in the same clip. I did the test with low cost recorders as presumably that's where the problems would be most evident. And currently, as I wanted to compare dual ADC with single ADC, only lower cost devices have single ADC circuits.
I chose the Zoom H2n to represent the 24 bit option, the H2e to represent the single ADC 32 bit float option, and the M2 to represent the dual ADC 32 bit float option. Obviously there were chosen largely because I have them in the cupboard!
To record the loudest possible sound which doesn't quite clip the H2n, its level control has to be set to zero. The other two recorders have no gain control. Can all three handle the same loud music without distortion? What does the silence sound like when the music stops and only the fly can be heard? (Actually there's no fly, I just recorded a few words in a quiet room, and normalised the results).
And then what happens if I only want to record the imaginary fly in its own recording, band-free? For that the gain on the H2n has to be set to maximum (10). but the other two have no gain control so can they capture that tiny sound as well as the 24 bit device? Or in this practical experiment, the sound of quiet speaking?
Finally what difference is there between the results from the dual ADC M2 and the single ADC H2essential?
The tests were carefully done but there are a lot of variables involved, so maybe the playing field isn't completely level. For instance in the quiet room tests, the M2 waveform looks noisy but actually its mics record more LF level from the very faint suburban traffic noise outside. But still, I think the outcome is interesting.
https://youtu.be/5bqYQ87e64Y
I don't make money from my channel so I'm not profiting from this post.
-
I have just uploaded some tests on YouTube which attempt to show how well three recorders perform when recording something very loud (rock band) and something very quiet (fly) in the same clip. I did the test with low cost recorders as presumably that's where the problems would be most evident. And currently, as I wanted to compare dual ADC with single ADC, only lower cost devices have single ADC circuits.
I chose the Zoom H2n to represent the 24 bit option, the H2e to represent the single ADC 32 bit float option, and the M2 to represent the dual ADC 32 bit float option. Obviously there were chosen largely because I have them in the cupboard!
To record the loudest possible sound which doesn't quite clip the H2n, its level control has to be set to zero. The other two recorders have no gain control. Can all three handle the same loud music without distortion? What does the silence sound like when the music stops and only the fly can be heard? (Actually there's no fly, I just recorded a few words in a quiet room, and normalised the results).
And then what happens if I only want to record the imaginary fly in its own recording, band-free? For that the gain on the H2n has to be set to maximum (10). but the other two have no gain control so can they capture that tiny sound as well as the 24 bit device? Or in this practical experiment, the sound of quiet speaking?
Finally what difference is there between the results from the dual ADC M2 and the single ADC H2essential?
The tests were carefully done but there are a lot of variables involved, so maybe the playing field isn't completely level. For instance in the quiet room tests, the M2 waveform looks noisy but actually its mics record more LF level from the very faint suburban traffic noise outside. But still, I think the outcome is interesting.
https://youtu.be/5bqYQ87e64Y
I don't make money from my channel so I'm not profiting from this post.
Thanks for this! It's an interesting comparison of the devices as a whole, but it's hard to tell whether the noise during the quiet parts is due to the preamps/ADCs or the microphones themselves (I presume they don't share the same internal mics?).
One interesting thing about the Zoom H2n's 'clicks' is that many years ago, while running a very similar experiment between a Zoom H1 and a Roland R-05 (both fed by the same external mic), the Zoom H1 also displayed the same pulsating noise! It was interesting because the Roland R-05's noise floor was actually louder, but it was a constant, uniform hiss. The Zoom H1 had quieter hiss between each click (by something like -8 dB compared to the R-05), but the clicks were louder than the R-05's hiss by at least 5 dB.
-
I'm glad it was of interest. The clicks would only be heard when massively amplifying audio recorded with the gain on zero (well, I guess I haven't checked at 1 and 2...) which normally would not be something you would do. Perhaps I should have done less normalizing and let people check the quiet tests at the level they were recorded at, but then I'm not sure that would have been a level playing field either. I shall think of what I could try next...
Also it's hard to find a location which is really quiet enough for the quiet tests - which perhaps illustrates that maybe when it comes to real-world recording, we are getting too fussed about system noise (within limits). As for the difference between dual and single ADC devices, well, overall the M2 came out best imho overall, but then it's probably a better quality device end to end (mic, preamp, and the twin converters, and fully integrated because you can't connect any external source). So again, the playing field isn't level. Hmm.
-
"sync sound" has reported on the reddit thread that I started that he queried Zoom about the H2e and its single converter, and got this reply -
"Me: "Since it's a single 24 bit converter, does that mean that the total dynamic range of the recorded file can't exceed the dynamic range of that converter?"
Zoom: "Yes this is correct, and the built-in microphone itself will clip before the dynamic range would be reached."
VERY interesting. On the face of it, and in the light of my tests, what they could be doing is taking the 24 bit output from the converter, translating it to 32 bit float, and adding about 15dB to every sample. So it looks clever when you see it in your DAW like a brick wall and you have to bring it down 15dB to remove the digital distortion, but it's just playing around with the samples to make it appear to be 'real' 32 bit float, when actually it remains real 24 bits. Fortunately that messing with the samples will not affect the final quality - it's just maths - but it still leaves a funny taste in the mouth. Or the ears.
Once again, I'd say don't bother reading the 32 bit float claims. Focus on the number of converters claimed - until someone from one of the companies convinces us that there is an advantage in 32 bit float, single ADC. And of course focus on the end-to-end sound - a crap preamp or noisy mic will be very accurately captured by either real or pretend 32 bit float.
-
Zoom: "Yes this is correct, and the built-in microphone itself will clip before the dynamic range would be reached."
Maybe this is why their cheaper 32-bit recorders with built in mics have single ADCs whilst the, also cheap, H1 XLR had dual ADCs?
They figure that the mics will clip anyway so nobody will notice that we’re cheating?
I’m still quite impressed with the overall performance of the H1 XLR for the price sold.
-
24-bit has always exceeded the dynamic range of most (all?) microphones and analogue stages. That's never been the issue. The issue is that the AD converter has its own inherent ANALOGUE noise floor (not talking about quantisation errors here) - it needs to be fed a signal that is strong enough to surpass that noise floor by a wide margin, or else this shows up in the recording as hiss.
Since Zoom decided not to include gain control on the device, it makes perfect sense for them to match the strongest signal the ADC can take to the strongest signal the microphone can output without distorting. The 24-bit digital container has more than enough dynamic range to contain every last bit of useful signal it converts, i.e. from the loudest signal the ADC can take to the ADC's own analogue noise floor. Quantisation errors would be quieter than the ADC's noise floor.
This does make the device very fool-proof. The user is never going to be able to set gain high enough to clip digitally when the mic is producing a non-distorted signal. Likewise, the user is never going to set gain so low that the ADC's analogue noise floor becomes lower than the 24-bit digital noise floor (i.e. quantisation/dithering noise), which WAS possible to do in devices that allowed you to set digital attenuation instead of giving the option of using an actual pad before the analogue stage.
But again, while 24 bits is more than enough to store all that data, it needs to be stressed that this data includes the ADC's own analogue noise floor! That analogue limit has ALWAYS been the issue. So what Zoom are saying here is a bit disingenuous - by focusing on clipping levels (which no one doubted they could deal with in the first place), it is easier to ignore what happens on the other end of the range. And while their statement is also correct about the digital limitations of the quieter end of the 24-bit container (i.e. they are not an issue), it conveniently ignores why users used to accidentally clip in the first place: they added analogue gain to quiet microphone signals so that they would be as far away from the ADC's analogue noise floor as possible, and then a sudden louder noise would clip.
If you can't clip, then that also means you can't boost quiet signals far above the ADC's analogue noise floor - that is, unless you make clever use of a parallel analogue stage/ADC OR a variable analogue gain/compensation technique with a single ADC, which they are almost certainly not doing because they keep focusing on the limits of the digital dynamic range instead.
-
It must be said though that the internal microphone's self-noise, under the setup I described above, might well still be louder than the ADC's analogue self-noise. In that case, Zoom are absolutely right: there is NO need for a second ADC. At all. I don't know if this is the case, but it's a possibility (the mics would need to be fairly noisy). If so, fair play to them - they have designed the perfect recorder for that specific microphone. All bets are off if you're using better external microphones, though.
-
Yes that's essentially it.
Forgive me for making a few nit-picky corrections in differentiating the analog noise-floor from the digital noise-floor.
[snip..] The issue is that the AD converter has its own inherent ANALOGUE noise floor (not talking about quantisation errors here) - it needs to be fed a signal that is strong enough to surpass that noise floor by a wide margin, or else this shows up in the recording as hiss.
The analog noise floor of the recorder comes from from the analog circuitry ahead of the ADC (analog 'hiss'). The digital noise-floor comes from the ADC itself (quantization noise). We want both to be low enough in level, but we want the quantization noise to be lower than the analog noise, because "smooth low level hiss" is far more perceptually benign than "nasty" quantization noise. So we want the analog noise to be low, but not overly low. Dither is the intentional introduction of noise at the analog stage just ahead of the ADC which is just slightly higher in level than the onset of quantization noise.
Quantisation errors would be quieter than the ADC's noise floor.
^That is the purpose of dither. Just enough added noise to effectively randomize the value of the least significant bit through the ADC, which effectively eliminates quantization noise. The question then becomes if the noise-floor of the analog circuit path ahead of the ADC can effectively act as dither itself, or if additional dither noise need be added.
This does make the device very fool-proof. The user is never going to be able to set gain high enough to clip digitally when the mic is producing a non-distorted signal. Likewise, the user is never going to set gain so low that the ADC's analogue noise floor becomes lower than the 24-bit digital noise floor (i.e. quantisation/dithering noise), which WAS possible to do in devices that allowed you to set digital attenuation instead of giving the option of using an actual pad before the analogue stage.
My claim is that any digital recorder that allows you to record in such a way that quantization noise noise becomes apparent in low level signals is improperly designed. Either dither noise or the input stage noise-floor acting as dither (both of which are forms of analog noise) should define the low level dynamic limit of any analog input digital recorder. That noise should never be lower than the least significant bit of the digitized output.
-
The analog noise floor of the recorder comes from from the analog circuitry ahead of the ADC (analog 'hiss'). The digital noise-floor comes from the ADC itself (quantization noise). We want both to be low enough in level, but we want the quantization noise to be lower than the analog noise, because "smooth low level hiss" is far more perceptually benign than "nasty" quantization noise. So we want the analog noise to be low, but not overly low. Dither is the intentional introduction of noise at the analog stage just ahead of the ADC which is just slightly higher in level than the onset of quantization noise.
I think we are just slightly disagreeing on what counts as the analogue circuitry before the ADC and the ADC itself. As well as the quantisation noise, the ADC's circuitry produces its own analogue self-noise due to thermal noise, having its own input stage, etc. This should always be higher and more relevant than the quantisation noise (if it isn't, there is something wrong with the design). I like thinking of it as the ADC's noise (rather than, say, part of the preamp stage) because you can't skip it - it doesn't matter if you use Line-In or Mic-In, nor if you use internal or external gain. For our purposes, we have to make sure the resulting signal from the microphone + preamp stages reaches the ADC at a level that makes this noise insignificant.
When I mentioned that you can run into quantisation/dithering noise, I was thinking of situations where the recorder allows you to set digital attenuation after the entire analogue chain and digital conversion. The more digital attenuation you add, the more of the signal you "submerge" into dithering/quantisation noise waters. But a recorder should not be designed in a way where digital noise is higher than the ADC's own analogue self-noise.
Does that make sense?
-
It does and I think we are essentially saying the same thing. I'm just specifying some points of definition I think will be conceptually helpful to others.
Any analog circuit will have a noise-floor that cannot be lower than that of the thermal noise, including an ADC (the "A" in ADC). But since dither is analog noise and quantization noise is a digitization artifact I think it best to not lump the two together. Conceptually, when I speak of ADC quantization error I'm talking about the decimation and digital part of the conversion and I think it helps to consider the analog circuit noise and dither noise as being separate from that. Dither is compensation for quantization noise, necessarily made before quantization artifacts can occur in the digitization process, applied in an entirely different "analog realm" prior to the "digital realm" conversion steps, even if both happen within the same physical ADC chip.
I do disagree with the wording here-
[snip..] The more digital attenuation you add, the more of the signal you "submerge" into dithering/quantisation noise waters.
Once digitized, the noisefloor becomes fixed relative to the rest of the signal. We can alter level digitally and the noise-floor shifts along with the signal. If we amplify a lot so that low level signals become perceivable, the noise floor is amplified along with the signal and can also become perceivable. The boat rises along with the tide, both are being raised or lowered by the same amount. The signal "submerging" into noise (into analog thermal or dither noise in a properly designed ADC rather than into quantization noise) all happens upon or prior to the digitization process.
-
I do disagree with the wording here-
[snip..] The more digital attenuation you add, the more of the signal you "submerge" into dithering/quantisation noise waters.
Once digitized, the noisefloor becomes fixed relative to the rest of the signal. We can alter level digitally and the noise-floor shifts along with the signal. If we amplify a lot so that low level signals become perceivable, the noise floor is amplified along with the signal and can also become perceivable. The boat rises along with the tide, both are being raised or lowered by the same amount. The signal "submerging" into noise (into analog thermal or dither noise in a properly designed ADC rather than into quantization noise) all happens upon or prior to the digitization process.
The noise floor won't shift along with the signal, though! I mean, it will do so if you are talking about the noise that has been digitised, as long as it still fits in the bits available. If the noise floor of the recording is -90 dBFS and you lower it by -10 dB, then yes, the new noise floor will be -100 dBFS because this still fits in the 24-bit file. But as you reach the lower end of the 24-bit format's dynamic range, the audio data that is just immediately above it will drown in dither or quantisation noise (if the device or software doesn't apply dither when saving the file).
Another way to think of this: you've recorded a show with a high dynamic range, and you peaked at -1 dBFS. The quietest songs are around -45 dBFS. The recording's noise floor (venue noise, preamp noise, whatever analogue noise was created) sits at around -60dBFS, which is audible if you crank up the volume for the quiet parts. Now, if you attenuate this by -80 dB, save the file to 24-bit, open it again and amplify it by 80dB, you will have increased the noise floor in relation to the quiet music because the dither or quantisation noise that was applied when you saved the file will have drowned the original noise floor! This is why audio editors default to working at 32-bit float - because then you can attenuate it and amplify it by however much you want without introducing more noise/artifacts. Then, when you've got your ideal levels set, you can export to 24 or 16 bit.
-
Just to add to my last point:
If you open an audio file in Audacity and use the silence command on it, you can amplify it infinitely without hearing any noise because Audacity works in 32-bit float. Now, if you export this silent file to 16-bit and 24-bit files and then reopen the files in Audacity you will find:
- You can normalise the dithering noise added by Audacity to 0 dBFS by adding +71.224 dB to the 16-bit file.
- You can normalise the dithering noise added by Audacity to 0 dBFS by adding +118.474 dB to the 24-bit file.
This is by using the best quality 'Shaped' dither under Preferences > Quality.
My point was that any attenuation you do digitally on a 16 or 24-bit file, if it is actually saved to the file instead of a temporary 32-bit float one, is not a lossless operation. If the device (or software) applies attenuation after the conversion, you are losing data. The only lossless operations that you can save to 16 or 24-bit files are simple ones like copying, moving, splicing, deleting, etc. As soon as you manipulate the actual sound waves with EQ, compression, amplification, and so on, you're not getting the exact same data back by doing the inverse operation.
-
Wow. Lots to absorb there. Could any of this account for why 32 bit float recorders seem to peak at +15dB so that it has to be attenuated in post? (By that I mean that when you record into them at almost analog clipping level, the file is recorded to +15dB instead of 0dB as one might logically expect?)
-
Totally agreed on the last two posts.
The noise floor won't shift along with the signal, though! I mean, it will do so if you are talking about the noise that has been digitised, as long as it still fits in the bits available.
^ That's it. I was referring to the initial digitization through the ADC only, not any additional subsequent processing.
Yes, if performing some subsequent operation other than simple ones like copying, moving, splicing, deleting, etc. then both the operation space in which the calculations are performed and the saved output format must accommodate the potential increase in wordsize. One might get away with digitally amplifying or attenuating after the ADC in a recorder which uses a larger internal operation space and still save all meaningful audio data within 24bits by truncating empty zeros at the top or random bits at the bottom, but there's nothing that keeps the user from overdoing it and exceeding the limits "meaningful audio data". Technically its not bit-perfect even if it still contains all meaningful audio data.
Simple take away is that if a recorder is going to allow the user to digitally amplify/attenuate, or perform any other non-simple operation after digitization, it should use a larger calculation space AND save to 32-bit-floating-point. Otherwise, best not to provide those functions on the recorder.
But if not allowing for those operations, the dynamic bottle neck of any of these recorders remains their analog circuitry and not the 24bit file format. All the clever tricks being used to increase dynamic range such as analog gain ranging ahead of a single ADC or multiple analog paths through multiple ADCs or whatever could still be applied to usefully increase real world dynamic range while writing a standard 24bit file.
Thanks for the engaged discussion!
I just wish that whatever shenanigans are going on under the hood had been applied to improving 24bits recorders without muddying the waters with the whole 32-bit float output file thing. 24bits is fully capable of accommodating that with all the niceties that folks now tend to associate with "32-bit" such as no level setting in these kinds of small all in one recorders. I don't really care what those shenanigans are, as long as the output is as identical to the input as possible, which is the metric that counts.
-
Wow. Lots to absorb there. Could any of this account for why 32 bit float recorders seem to peak at +15dB so that it has to be attenuated in post? (By that I mean that when you record into them at almost analog clipping level, the file is recorded to +15dB instead of 0dB as one might logically expect?)
Only reason I can think of is that the signal level captured in the raw recorded file is no longer being effectively normalized upon capture by the user setting an appropriate input gain. Because of that initial playback levels prior to making any level adjustment afterward will tend to be much lower than they used to be. Increasing level in the output file by 15dB may just be some compensate for that. 15dB being somewhat arbitrarily decided upon as being about right, but the specific amount not really mattering. Keeps folks unfamiliar with the new 32-bit-float ways from thinking "man this is really low in level unless I amplify it", and also makes direct playback without making a level adjustment more reasonable.
But that's just my guess.
-
And one more thing - I've been saying that here for years - in the M2 for instance (dual converter device) there are options to normalise to a copy of the file, to normalise during replay, and to export to 16 or 24 bit. Am I right in thinking that these features will be work more accurately and losslessly if the stored recording is 32 bit float rather than 24 bit? There are also broadly similar things possible with the single ADC H2essential - mixing its tracks, normalising, exporting. Then again, a DAW can do that stuff with a 24 bit file but in a 32 bit float processing environment. But maybe with the files in 32 bit float at the outset, it makes the internal processing (using low grade processors probably) more efficient?
-
Am I right in thinking that these features will be work more accurately and losslessly if the stored recording is 32 bit float rather than 24 bit?
The mantissa is gonna be 24 data bits either way, but having a wider dynamic range on the container could pick up actual signal rather than blankness for a bit or three while it's shifting.
Just wondering if anyone can create a short file to be used as a dual-ADC obstacle course?
Ideally it would be music, but if speech, birdcalls, electronic sounds, or thunderclaps are more telling, then let's use the most difficult source to track, in order to reveal issues and/or differences between implementations!?
#ADCobstacleCourse
-
Curtis Judd has just released a video titled “32-bit Float Recorder Myths You Should Know About”.
I suspect that is in response to some of the discussion here and elsewhere.
-
Curtis Judd has just released a video titled “32-bit Float Recorder Myths You Should Know About”.
I suspect that is in response to some of the discussion here and elsewhere.
I saw that - I think it is intended as another contribution to the general "merits of 32 bit float" debate - but indeed he did briefly mention the dual / single ADC subject.
The key issue there, for me anyway, is that there is a bucket load of info about how dual ADC 32 bit float is implemented. (Well, apart from the exact way the output of the two devices are merged which may vary between devices, and there are some who can hear and/display resulting artifacts.)
But single ADC 32 bit float remains a mystery, by which I mean, sure it can be simply 24 bits written to 32 bit float resulting in no real audible difference, or it may be something potentially better than that if well implemented, like a kind of analog compression / digital expansion thing which seems quite plausible. But we don't know for sure, or even at all. And given that it's being used for marketing, and given that it's messing with our audio data, someone (presumably in a vendor company) should at least give a summary of what is going on, and why it's good.
-
But single ADC 32 bit float remains a mystery, by which I mean, sure it can be simply 24 bits written to 32 bit float resulting in no real audible difference, or it may be something potentially better than that if well implemented, like a kind of analog compression / digital expansion thing which seems quite plausible. But we don't know for sure, or even at all. And given that it's being used for marketing, and given that it's messing with our audio data, someone (presumably in a vendor company) should at least give a summary of what is going on, and why it's good.
Yeah, it would be good to how they implement this!
I've long understood the theoretical principles of this, but something that I hadn't fully figured out until recently was the actual bottleneck of the gear I used. When I used to tape with CA-11s>STC-9000>Zoom H1, I disliked the amount of noise I got during quiet passages when I had set the gain for other much louder parts. My interest in 32-bit float devices - and my eventual purchase of the Zoom F3 - was due to that.
But it was only after running some tests with the Zoom H1 and Zoom F3 side by side that I understood not only their practical limitations, but the microphones' as well. I made a post about it (https://taperssection.com/index.php?topic=206442.msg2421559) that was a bit too long, but essentially, I found that as long I was adding around 20dB of clean gain to the CA-11's output (either with the STC-9000 alone or a combination between the STC-9000 and the Zoom H1's internal gain - the Zoom H1 only provides 13dB of fairly clean gain and gets worse from there), the Zoom H1 and the Zoom F3 performed exactly the same. With the CA-11s, EIN maxed out at around -118.5 dB, no matter how much more gain you threw at it, because at that point the microphone's own noise floor became the bottleneck.
In short, if you are using a pair of CA-11s with a Zoom H1, you get no benefit from using any more gain than +20 dB. So if that gives you enough headroom, you will get no performance benefit from pairing the CA-11s with the Zoom F3 instead. I'd say that for about 70% of the shows I attend, I get more than enough headroom with the CA-11s +20 dB into the Zoom H1. At the loudest show I've ever recorded (Mogwai), I used +12 dB and peaked at -4 dBFS. This means that if I had used the Zoom F3 then, the noise floor of my recording would have been around 8 dB quieter than it turned out to be.
Note that +20 dB gain paired with the CA-11s is not a very strong signal! If you're recording a singer-songwriter, you might be peaking at -30 dBFS or even less, i.e. a lot of headroom and an end result as good as if you were using the best multi-ranging ADC device around.
Now, of course, this calculation changes entirely when you're using different mics - both sensitivity and self-noise specs would affect the ideal amount of gain you'd need to get the least amount of noise as possible. What I suspect these budget devices are doing is simply giving the user fewer options to optimise for the internal mics (maybe compromising a little bit on the quiet/clipping ends to strike an acceptable balance). For those of us using external mics, though, I can't see how that's a good thing.
-
I woke up this morning devising new tests of all this stuff that I could carry out using devices I have. Will I bother? We'll see...
When you mention how they may be optimizing these devices for internal mics, well, that's partly my (often repeated, sorry) view of the Zoom M2 (and M3) - as they have no inputs, Zoom must know that they have no way of blaming anyone else for bad outcomes, assuming they are correctly placed. So they must be optimising the setup of the preamp and dual ADCs (claimed in marketing to be derived from F series devices) with a mic that isn't a load of crap, otherwise if it was, there's no point in the device at all. And in my simple tests online, it's not an embarrassment.
-
tl;dr- If its a closed system that can only record through built-in mics, it really doesn't need a dual ADC.. even if its output is 32-bit files.
If it's a closed system, meaning no external input other than the built-in mics, it should be relatively easy for manufacturer's to get away with providing no level controls at all, including gross mic/line and fine gain adjustments, because the manufacturer can just adjust things however is needed to match the output parameters of the internal microphones they are using to the inout parameters of the ADC they are using. ..AND since the dynamic range of any of the built-in mics they are likely to use is going to be less than 100dB, they would have no problem at all saving that output as a 24bit file. A 32-bit-floating-point file option would be fine as well if that helps them sell more recorders, but wouldn't provide any higher quality. Same for dual ADCs.
Dual ADCs or other strategies along with writing 32-bit files is only technically helpful if the recorder has external inputs that need to accommodate gear with output parameters that are unknown to the manufacturer. Even though the dynamic range of any one of those "unknown input sources" could be accommodated without using such strategies if it's output parameters were known to the manufacturer, accommodating a sufficiently wide range of "unknown inputs" requires either manual input gain setting by the user or increased dynamic range via dual ADCs or some other strategy. But even with multiple ADCs, gross input gain in the form of mic/line input sensitivity still needs to be set manually by the user.
-
I see what you mean but the M2 (and I think the M3 but its MS setup makes things slightly more complex) quotes a max SPL of 135dB in the specs - which must include the mics, and all the way through to the 32 bit float output. As memory serves me, 24 bit would handle up to 144dB (?) so on the face of it, 24 bits would cover that ear damaging input. Just.
But as I think I've mentioned here - or elsewhere, oops - the M2 does some pretty efficient (faster than realtime) normalizing and converting onboard, and can even play back normalized on phones without prior conversion, so maybe 32 bit float helps a bit with that processing, in theory if not in practice.
(But would the ADC handle 135dB input at its analog stage without fear of clipping that front end of the ADC? Thus needing one with lowered input to accept high levels, while the other does the quieter stuff maybe with even raised gain right before the conversion stage of the chip? I have a horrible feeling that in my dotage I am sending this discussion round in circles...)
-
But as I think I've mentioned here - or elsewhere, oops - the M2 does some pretty efficient (faster than realtime) normalizing and converting onboard, and can even play back normalized on phones without prior conversion, so maybe 32 bit float helps a bit with that processing, in theory if not in practice.
I guess I'm missing the point here... Normalizing is one of the easiest operations one can do on audio. It doesn't matter nowadays if the processor is doing integer or floating point arrithmatic, both will be very fast, and maybe even outsourced to an onboard DSP. By far the biggest bottleneck is reading/writing the input/output samples to the slow SD card, not the calculations. Back in the days, microprocessors typically only had integer arrithmatic on board, meaning that floating point calculations were emulated by integer software algorithms, which made them slow. Fortunately those days are over.
-
Here we go again - in a YouTube video, my post on Reddit raising the matter of single ADCs in a 32 bit float recorder has been quoted at length, and then discussed at length, with a video explanation specially provided by Zoom's usual chap whose name escapes me.
https://youtu.be/soX4F5JkQ6w?si=UPM_hXWj7QST3ewp
I still think his explanation doesn't really say why writing 32 bit float from a single 24 bit float ADC is a good thing, although he does say that the ADC used is of a fundamentally higher quality than past devices.
Chiefly I think the good thing is that the myth that 32 bit float devices all have dual ADCs is now busted and public, and I am prepared from my own limited experience with the H2e to say that its lack of gain controls doesn't seem to matter, but I would still advise anyone buying a device marketed as 32 bit float to check the number of ADCs, as I can't believe anyone would argue that one ADC is as good as two. If you can afford two, or more, go for it.
-
Thanks for updating us on this stuff. I do not own one of these decks yet, but this type of information will be informative when I do buy a new deck.
-
I still think his explanation doesn't really say why writing 32 bit float from a single 24 bit float ADC is a good thing, although he does say that the ADC used is of a fundamentally higher quality than past devices.
He admits that it is merely a handy thing for technical reasons, chiefly post production.
If one were to take a 24-bit recording and load it into a 32 bit float workstation, would that be the same thing though?
-
I still think his explanation doesn't really say why writing 32 bit float from a single 24 bit float ADC is a good thing, although he does say that the ADC used is of a fundamentally higher quality than past devices.
He admits that it is merely a handy thing for technical reasons, chiefly post production.
If one were to take a 24-bit recording and load it into a 32 bit float workstation, would that be the same thing though?
Yes, that's almost exactly the same thing, except that with a conventional 24bit recorder, the recording might be clipping if the analog gain was set too loud during recording.
It was kind of funny to see his example clipping guitar recording that was fortunately repairable because it was stored as 32bfp. What he didn't mention was that the only way to get a clipping recording with that recorder is by applying digital gain, as the output of the 24bit ADC will never exceed 0dB per definition!
So indeed, if the recorder allows digital gain, then it is possible to set the gain 'too loud' and make it clip. Only when stored as 32bfp, this clipping is repairable, which is nice.
The alternative would be to NOT allow digital gain at all. Just store the output of the ADC as 24bit. Then the user can never set the gain too loud and there is never any clipping to be fixed. It does mean that any level/balance correction must be done in post, while in the current situation the level/balance correction can be done during recording. I do understand that they opted for digital gain + storing as 32bfp, it seems like a sensible choice for a recorder with fixed analog gain, a single ADC and only built-in mics just because of this level/balance correction benefit. I'm sure some folks will appreciate that feature, although I personally couldn't care less.
-
I think he's using the H1essential which has no gain control at all, neither digital nor analog. They seem to set the output from the ADC to exceed zero in the 32 bit float file even if the limit of the max SPL of the mic hasn't been reached. Perhaps the idea is that typical sound levels will stay below zero, and thus may not need fixing in editing - louder sounds which would have gone into the red in older devices without significant reduction in analog gain now go over 0dB in the 32 bit float files, needing to be fixed later.
-
I think he's using the H1essential which has no gain control at all, neither digital nor analog. They seem to set the output from the ADC to exceed zero in the 32 bit float file even if the limit of the max SPL of the mic hasn't been reached.
Wow, amazing. If that is indeed the case, then they not only have a fixed analog gain, but also a fixed digital gain! That's the only way to get samples above 0dB from a single ADC. (As said, the maximum output of a single fixed point ADC is 0dbFS per definition.) In that case they might as well have simply stored the 24bit output of the ADC. Then there is really no point at all to make this device 32bfp, apart from marketing (or e.g. production benefits by sharing the same design, chips and software components for multiple recorders)
-
^ That.
As a 'closed system' which includes the entire signal chain from microphones to saved digital file, the designer can presumably set analog gain during the design stage such that the level out of the microphone input stage optimally matches the input sensitivity of the ADC. Once that has been done, there is no real imperative to provide user controlled 'gain.. other than as a way of achieving better normalization for quiet recordings, yet that can be managed afterwards. If the design of the recorder uses a single ADC per channel rather than being one which effectively switches between multiple ADCs as a way of increasing dynamic range, performance shouldn't differ regardless of the output being saved as 24bit fixed or 32bit float. The data content of both should be identical. Only difference is the size of the resulting files and how the output file is handled after being transferred for editing.
-
I won't pretend to understand the technical part of all this (in fact it makes my head spin), but all I know is when I record using my Tascam DR-40X at conservative levels for louder shows, I still get some clipping. When using the Zoom H4E, I never have. So in my personal experience, 32 Bit Float has made the recording process much more fun, not having to worry about setting levels just so, or monitoring them throughout the show.
-
I won't pretend to understand the technical part of all this (in fact it makes my head spin), but all I know is when I record using my Tascam DR-40X at conservative levels for louder shows, I still get some clipping. When using the Zoom H4E, I never have. So in my personal experience, 32 Bit Float has made the recording process much more fun, not having to worry about setting levels just so, or monitoring them throughout the show.
It is possible to overload the input stage on any recorder with a line in.
Once overloaded, running with low "knob" levels will prevent digital "over level," but the signal is already flattened on peaks due to design limitations of the input stage.
(There is a max level that any input can accept, over which, distortion will occur.)
If you are overloading the built in mics on the DR-40X, then no amount of turning the levels down will stop the "brickwall" distortion.
Move farther away from loud sources if this is the case.
-
I won't pretend to understand the technical part of all this (in fact it makes my head spin), but all I know is when I record using my Tascam DR-40X at conservative levels for louder shows, I still get some clipping. When using the Zoom H4E, I never have. So in my personal experience, 32 Bit Float has made the recording process much more fun, not having to worry about setting levels just so, or monitoring them throughout the show.
It is possible to overload the input stage on any recorder with a line in.
Once overloaded, running with low "knob" levels will prevent digital "over level," but the signal is already flattened on peaks due to design limitations of the input stage.
(There is a max level that any input can accept, over which, distortion will occur.)
If you are overloading the built in mics on the DR-40X, then no amount of turning the levels down will stop the "brickwall" distortion.
Move farther away from loud sources if this is the case.
Like I said, it makes my head spin :lol: I'm not using the built in's on either recorder, I use externals plugged in to the XLR ports
-
Like I said, it makes my head spin :lol: I'm not using the built in's on either recorder, I use externals plugged in to the XLR ports
OK, well, then this is really a topic for a different thread...
Something in your signal chain is maxing out before the analog>digital conversion.
I'm sure you'll be happier when you get that sorted.
-
Curtis Judd just put out a 2-video series: '32-bit float recorders & differences in audio quality'.
The first video is a nice background on the tech, and the second gets into more specifics as to which specific products implement this tech well, and how those that don't do it well cut corners.
https://www.youtube.com/watch?v=mxWH8V5BaYk
https://www.youtube.com/watch?v=f_9TM9QZPdo
-
I won't pretend to understand the technical part of all this (in fact it makes my head spin), but all I know is when I record using my Tascam DR-40X at conservative levels for louder shows, I still get some clipping. When using the Zoom H4E, I never have. So in my personal experience, 32 Bit Float has made the recording process much more fun, not having to worry about setting levels just so, or monitoring them throughout the show.
It is possible to overload the input stage on any recorder with a line in.
Once overloaded, running with low "knob" levels will prevent digital "over level," but the signal is already flattened on peaks due to design limitations of the input stage.
(There is a max level that any input can accept, over which, distortion will occur.)
If you are overloading the built in mics on the DR-40X, then no amount of turning the levels down will stop the "brickwall" distortion.
Move farther away from loud sources if this is the case.
Like I said, it makes my head spin :lol: I'm not using the built in's on either recorder, I use externals plugged in to the XLR ports
Tascam DR-40X
Maximum Input Level -3dBV (MIC) / +20dBu (LINE)
Zoom H4E
Maximum Input Level MIC: +4 dBu (1.78 dBV) / LINE: +24 dBu (21.78 dBV)
The Zoom has a max input level almost 5dB higher than the Tascam. If you switch to LINE and it will still supply 48V (assuming 'externals' means condenser mics), you won't overload. I ran a Sony Minidisc forever with DPA 4060's feeding it set to line level, and it was fine. If the mics have pads, consider using them. I don't know if the Tascam has a built in mic pad option; that would also solve it. If you are using an external preamp with fixed gain, it may be too much for the Tascam and not the Zoom. All of these things are gain staging considerations that will affect analog overload before signal ever gets to the AD converter stage.
I noted way back at the beginning of one of these discussions that the Zoom F8N/whatever has much lower input headroom than the Sound Devices 'whatever'. The Zoom F8n (non 32 bit) at least has a pad that can be switched in for mic use. Doesn't mean one is better or worse than the other, they are design decisions that balance different expectations/needs, so one has to adapt to those parameters.
The bad assumption with these devices is that the headroom window is adequate for all uses. Hot condenser mics at loud events (or up close to loud things) typically output line level signals rather than mic level signals. Remember part of the market is people in the field recording bird calls, so they don't assume a LOW amount of gain. The Rolls Royce device here would have adjustable analog gain with metering, and 32 bit float capture with metering, then you could really control the headroom relationships. I'm not aware of anyone making that device. Yeah, it's somewhat redundant; ok, so give us a separate analog overload indicator along with the converter metering.
-
Seems that what's driving all this is the simplification of gain staging by the user in the recorder. As mentioned, the need for proper gain staging doesn't actually go away, its just made less critical for a significant part of the typical range. That is until it exceeds that safety bubble range and once again becomes critical, like it always was.
I'm all for reducing potential failures, but I wish manufacturers would do so without obfuscating the details of gain-staging and instead make the monitoring of levels and the awareness of where the signal fits into the maximum dynamic range of the recorder more transparent to the user. Don't need to dumb things down in order to make things easier when the system could instead be made more transparent and empowering..
The Rolls Royce device here would have adjustable analog gain with metering, and 32 bit float capture with metering, then you could really control the headroom relationships. I'm not aware of anyone making that device. Yeah, it's somewhat redundant; ok, so give us a separate analog overload indicator along with the converter metering.
I'd like to see an overload indicator on the analog input stage, followed by a peak/VU meter with a scale that shows the entire effective dynamic range through the converter, from the recorder's noise-floor up to 0dBfs. That's supposedly a range of about 142 dB through a current SD Mixpre recorder. The scale might be made non-linear to fit if necessary. Primary role of the meter when recording is to indicate: Modulation (meter "dancing movement" indicating active signal); Peak level; VU level; and in this case noise floor as well. Showing the entire range allows for seeing modulation at all input levels, and allows for determining where the current noise floor and peak level actually fall within the available input range. User can also easily visually determine the approximate total dynamic range of whatever signal is being recorded.
Extra credit for including a metering mode that actively changes scale with content, starting fully "zoomed in" and automatically "zooming out" as necessary to keep whatever the current peak-hold value happens to be at the top of the meter scale and the current minimum noise-floor value at the bottom of the meter scale. The meters will then show the full range of active modulation with maximum possible visual resolution. Pushing the "peak reset" button resets the display to some comfortable maximum zoomed in range or whatever the current modulation range happens to be if if exceeds the max zoom default. Display scale numbering across the top of the meter updates with each automated "zoom out" step as the uppermost and lowermost values indicated on the display change with each excursion past the previous peak and minimum value.
-
I suppose that what they do is to cater for the vast majority of users' needs, and tough luck on the rest... but, having said that, has anyone actually encountered an issue with recording on a proper 32 bit float device, in terms of wishing that they could have increased the analog gain, or reduced it? Genuine question, I'm not taking sides here (wouldn't dare...:) )
-
No, I have not encountered any such issues. I'm perfectly happy with the 32-bit system in my three Zoom F3s. And I've been recording shows for more than 50 years, starting in the cassette era.
-
I usually use a Zoom F3 paired with CA-11 cards, and after running some tests, I found that the microphones are the bottleneck in terms of self-noise. Whether I plug them straight into the F3 (with a battery box + adapter) or add a pre-amp to the chain with a lot more clean gain, the level difference between a test signal and the gear's self-noise remains the same.
I also ran the same test with a pair of Clippys, which are way more sensitive and have less self-noise than the CA-11s. With Clippys>Battery box>Adapter>F3, the difference between the test signal and the gear's self noise was about 4dB greater than the result I got with the CA-11s>Battery box>Adapter>F3 rig - that is, changing the mics increased the signal to noise ratio by 4dB. With the Clippys>Pre-amp>Adapter>F3 rig, the difference was 5dB. This means that adding loads of gain between the Clippys and the Zoom F3 only increased the signal to noise ratio by 1dB.
My conclusion is that unless your microphones have EXTREMELY low self-noise, it's pointless to worry about the Zoom F3's gain staging.
-
I was checking spec sheets for a couple of Zoom 32bit float recorders and notice repeated use of the following wording (this taken from the F3):
Equivalent input noise: −127 dBu or less (IHF-A) when waveform magnification is ×1024 with 150 Ω input (My empasis)
Apologies if this is a basic question or have been covered before, but why would the input noise be affected by the waveform magnification?
I seem to recall early discussions and confusion about the F3, and came out with the impression that the waveform magnification setting on the display was immaterial and would not affect the quality of an 32bit recording when normalised later.
What is it I am not understanding here?
-
I was checking spec sheets for a couple of Zoom 32bit float recorders and notice repeated use of the following wording (this taken from the F3):
Equivalent input noise: −127 dBu or less (IHF-A) when waveform magnification is ×1024 with 150 Ω input (My empasis)
Apologies if this is a basic question or have been covered before, but why would the input noise be affected by the waveform magnification?
I seem to recall early discussions and confusion about the F3, and came out with the impression that the waveform magnification setting on the display was immaterial and would not affect the quality of an 32bit recording when normalised later.
What is it I am not understanding here?
You're not misunderstanding. You actually understand the unit better than Zoom does, apparently. The F3 manual has some glaring inconsistencies and downright false or misleading statements. One of the biggest being the "Waveform Magnification" business, where they state at one point that changing that setting does not affect the recorded level. But, as anyone who has used the F3 can tell you, it absolutely does affect the recorded level. The thing is, that level is only being adjusted post-ADC, so it's just a digital fader rather than an input trim or analog gain setting. The analog preamp gain feeding the ADC on the F3 is a fixed, non-adjustable level. This is the case for all Zoom recorders when recording in 32-bit float format (the F6 and F8 give you a 24-bit fixed option in which case the analog preamp gain is adjustable).
The statement from the specs you quoted strongly implies that analog gain is adjustable, because you typically measure EIN with the preamp gain at max level. Since we know preamp gain is fixed and we know what "waveform magnification" actually does, the portion of the statement you bolded should not be there.
-
That's a great clarification - but I wonder whether when they said it doesn't affect the recorded level, perhaps they were treating the massive dynamic range of the 32 bit float data as something not related to recorded level? In other words, changing the 'magnification' is not really changing the level - same as when you look at an insect through a microscope, it's not actually getting a whole lot bigger? In other words, the 'recorded level' is set unchangeably by the preamp, and the post ADC gain is simply changing how it is numerically expressed?
Show me a hair and I'll split it...
-
But, as anyone who has used the F3 can tell you, it absolutely does affect the recorded level.
This was an excellent question by Niels, and a great answer by voltronic.
But (and please correct me if I'm wrong) isn't it so that on the F3, the waveform magnification setting at the moment of starting the recording determines the level of the recorded audio, but changing the waveform magnification while recording does not change the level of the recorded audio?
When selecting a 32bfp recorder a while ago this is how I understood the F3. I actually liked this behavior as it makes so much sense in a 32bfp recording device with fixed analog stage. The Tascam FR-AV2 which I eventually bought does change the recorded level when changing the digital gain during recording. So in practice, I never touch the gain setting while recording, as it doesn't add anything to the quality of the recording and can only result in extra work in post. On the F3, at least, one can modify the magnification to get a nice waveform on the display without changing the recorded level!
-
Thanks @voltronic for taking the time to explain.
I have a feeling of deja-vu.
It has probably been covered repeatedly in some of the sometimes very long ongoing threads, but it has only recently become relevant for me when I got one of those 32bit recorders.
-
The entire range between −127 dBu at the bottom and 0dBuFS at the top fits easily within 144dB!
-
The entire range between −127 dBu at the bottom and 0dBuFS at the top fits easily within 144dB!
True... but if you allow the user to digitally amplify (or magnify or whatever it is called) the sound, you might end up above the 0dB at the top... Then the enormous dynamic range provided by the 32bfp storage format becomes beneficial and takes a source of error away.
-
If it ain't broke..
There is good reason for implementing a multiple ADC switching architecture. Make the dynamic range of the recorder larger than that of the microphone. I welcome that. But there is no good reason for forcing 32bit storage in place of 24bit in combination with it. 24bit is already larger than the increased dynamic range the recorder provides by way of the new multiple-ADC architecture. In contrast to multiple ADC switching, I don't welcome the unnecessary and unwanted complication of forcing the 32-bit file format on users.
It's not a tall ask to set gain initially for whatever microphone is being used, based on the specs of the mic. One never need touch the gain setting again until switching to a microphone with significantly different specs. That supports the more technical users. Casual users wouldn't need to do either. Good design choices empowers users rather than unnecessarily constraining them.
/rant
-
I'm quite convinced the 32 bit float thing is purely an advertising gimmick, when the device has a single ADC. I recently asked DJI whether their wireless mics (three versions now) which trumpet recording 32 bit float audio internally have dual ADCs. They confirmed what I expected, they don't. Off the top of my head I don't recall seeing any wireless mics advertised with 32 bit float internal recording by any manufacturer claiming dual ADCs in advertising. (But maybe some just don't bother to claim of course) And there seem to be lots of them these days.
The technical point of 32 bit float from single ADCs seems to be simply that the digital format is the same as many other devices now produce, so if you are working in post-production with 32 bit float data from multiple recording devices, it's all the same format, and you don't have to spend (possibly small amounts of) time reformatting 24 bit data as 32 bit float.
-
From a gain-staging perspective, processing at 32f makes things hella easier.
I mix to 32f now and then save those files before applying processing (gain/limit) where I output to 24 bit.
Running with 32f as a default probably saves the manufacturers a lot of customer service calls.
I can imagine a workflow where audio gets passed from one app to another and it sure is better if it's not clipping on any one of those.
-
Storage is cheap, and 32 bit has benefits. If it’s available I use it. Same for 96khz sample rate.
-
Help me to understand what I'm missing. I'm fully onboard with processing, transferring between applications, and whatever else at 32bit ..but I've long had the ability to do that with 24 bit raw recorded files which I can simply import into the 32bit float workspace of a digital editor with no additional steps - no reformatting of the raw 24 bit input files necessary. Same goes for 16 bit source files or the combination of the two. Once in the editor, I can do whatever processing is required in 32bit float and set the output to whatever format I desire. How does 32 bit input files ease workflow over 24 bit input files? Is what I'm doing somehow unique to Samplitude? I don't think so, but maybe. Additionally, without opening them in the DAW I can preview and modify the raw 24 bit files easily - playing them, splitting / interleaving tracks etc.
Same is not true for 32bit raw files. Not as easy to do many of the simple tasks. I'm not trying to be a contrarian here, I just don't understand the ease of workflow argument. Help me to see the advantage after recording.
I'm quite convinced the 32 bit float thing is purely an advertising gimmick, when the device has a single ADC. I recently asked DJI whether their wireless mics (three versions now) which trumpet recording 32 bit float audio internally have dual ADCs. They confirmed what I expected, they don't.
To my understanding (as outlined above) this achieves nothing useful to the user.
-
Well... good questions. Such benefits as there might be are subtle rather than dramatic / essential. I am currently archiving DAT recordings dating from 1987 onwards. Classical music live stuff. 60+ done, I have yet to encounter any clipping (perhaps excluding applause but the clipped transients in that are not really obvious). And I haven't been thinking, wow this would have been so much better if the audio was 24 bit, let alone 32 bit float. But I did have to be careful with levels at the time and I did have to rely on a lot of experience. Especially with operatic sopranos... capsule splitters, some of them.
These days I don't do serious recording but if I did, I'd lean towards 32 bit float for insurance as it were, but not expecting any obvious audible difference.
-
^ Right. Those advantages are what a multiple-switching-ADC architecture achieves on the recording end. I understand and fully approve of that part!
-
Without checking back all the pages of this thread, I may be repeating something already posted, but in the course of investigating something or other, I just came across this research document where a 32 bit float audio interface was tested in remarkable depth. It might be of general interest to those in this thread. Or incredibly boring...
https://www.researchgate.net/publication/374782111_An_experimental_investigation_of_32-bit_float_ADC_systems_-Zoom_UAC-232_Test_Report_using_Multi-Instrument
-
Thanks Peter. Saw that back when and re-read it again over the weekend. The tests confirm Zoom's switching ADC architecture achieves the goal of extending dynamic range in a way which seems to be fully sufficient for our concert recording needs. And it somewhat explains how the same may not be true for some esoteric sound design uses due to the level of its noise floor modulating along with the ADC switching, which does not cause any problem for concert recording where the noise floor is fully masked by either the music itself or by other sources of noise in the signal chain at the levels around which the switching occurs. Presumably the results carry over to the Zoom recorders using the same or similar tech as the interface. Would love to see such an analysis done on the recorders TS members are using.
What isn't confidence inspiring is the parroting in the introduction of a marketing-driven untruth we've discussed previously, made without any backing attribution (as usual). That's unfortunately become par for the course in marketing materials but is inexcusable in a supposedly rigorous technical paper reporting test results. Would love to know if peer-reviewers of the paper (if any) called it out. This establishing statement made in the introductory paragraph struck me as admirably clear and precise: "Ideally, the noise level would only cause the least significant bit to toggle while the maximum level would fully utilize all bits." Reading that I thought, "Yes! That is the goal, clearly and concisely stated." [smiles] ..and based the limits of the device as explored in these tests, that easily fits within 24 bits" [raised eyebrow]. Only to be immediately let down upon reaching the second paragraph by the claim stating- "Audio that is too loud will be clipped at the top of their waveforms while audio that is too quiet will lack fidelity or even be buried in the noise floor (see figure below).'' (< bolding for emphasis is mine) Clipped, yes. Buried in noise-floor yes. Lacking fidelity no. In making that claim, the onus is on the author to explain why digital sampling theorem suddenly no longer applies.
Making it more egregious is the inclusion of the misleading marketing illustration that follows, showing a stair-step output response at "weak signal" levels. Whenever you see that you should immediately be wary, because in making that claim (seemingly always without attribution) the author has almost always sabotaged their own credibility.
Scale up the analog output from any digital recording which was properly frequency limited to fit within the passband imposed by the sampling-rate (achieved via proper design of the ADC), and which, in the author's own words, "Ideally, the noise level would only cause the least significant bit to toggle.." and and show me stair steps. There are none. The presence of noise sufficient to cause the LSB to toggle randomly eliminates low-level quantization error. Injecting dither noise if needed to achieve that has long been correct practice for an old problem solved long ago. The presence of additional noise further removes all possibility low-level quantization error occuring.
Otherwise, great to see these kinds of in depth test results and, ignoring that difficult to overlook bit of bs in the introduction, the testing seems good to me. Experts are welcome to correct me if I'm missing something in those results which might prove problematic for concert recording.
More on why the stairstep drawings are misleading and claims of "lack of fidelity" are bs here (great video made 12 years ago, link to which has been posted a number of times around TS)-
https://youtu.be/cIQ9IXSUzuM?si=-YpzlQRgKEOQIOIe&t=216
Reaction / further-explanation to the above above (posted a 2 years ago, the truth is unchanged)-
https://youtu.be/cD7YFUYLpDc?si=aMevn5fMs8keOkmq
-
^ All good thought provoking stuff!
This morning I had a new thought. For once.
Companies making single ADC devices writing 32 bit float data from the 24 bit output of the ADC could be adding a multiplier to the data to make it look like real 32 bit float. Say the multiplier was 1.2 - so what would have been written as 0.99 became 1.188 - and if 1.0 equated to full scale, the audio would look digitally clipped when editing. But then of course you could reduce the level by a factor of 1.2 (or more) and the clipping would be gone. And the user would think, wow, real 32 bit float helping to avoid clipping. But actually it would be a con. Hmmm. I wonder if that is happening?
Note that my example has been expressed very crudely - I hope my point is clear.
-
The multiplier is always a power of 2, so you would never see a multiplier of 1.2.
https://en.wikipedia.org/wiki/Single-precision_floating-point_format (https://en.wikipedia.org/wiki/Single-precision_floating-point_format)
an IEEE 754 32-bit base-2 floating-point variable has a maximum value of (2 − 2−23) × 2127 ≈ 3.4028235 × 1038. All integers with seven or fewer decimal digits, and any 2n for a whole number −149 ≤ n ≤ 127, can be converted exactly into an IEEE 754 single-precision floating-point value.
-
The multiplier is always a power of 2, so you would never see a multiplier of 1.2.
https://en.wikipedia.org/wiki/Single-precision_floating-point_format (https://en.wikipedia.org/wiki/Single-precision_floating-point_format)
an IEEE 754 32-bit base-2 floating-point variable has a maximum value of (2 − 2−23) × 2127 ≈ 3.4028235 × 1038. All integers with seven or fewer decimal digits, and any 2n for a whole number −149 ≤ n ≤ 127, can be converted exactly into an IEEE 754 single-precision floating-point value.
Confession - that's over my head! But maybe that doesn't change what I am saying. In a DAW you can raise the gain of a normalized undistorted 32 bit float file, so that it now peaks above the maximum, and on replay it sounds clipped. And of course you can later reduce that gain back to where it was and it is no longer clipped on replay. What if the last process in the single ADC on the recording device increases the signal level unnecessarily as it outputs to the 32 bit float file, in just the same way as you can boost it in a DAW, so that when you open the file in your DAW it looks and sounds clipped, but you can normalise it back to an unclipped file, and think that the 32 bit float format has done something useful?
I've just being doing some very simple tests with a single ADC 32 bit float recorder, recording from its internal mics. Zoom claim that the device can record the maximum specified SPL level of the mic without clipping. When I record a test tone from a loudspeaker at a neighbour-annoying level with the device held right up to the speaker, and open it on my DAW, it's over the top and clipped. When I normalise it, the waveform is perfect (zoomed in to sample level) and it sounds fine. But why is the audio being written to file above full scale when it doesn't need to be, as the max SPL of the mic wasn't reached? Surely they could configure the device so that when the mic hits its maximum SPL, the 32 bit float file just reaches full scale? As it is, a user might think that without 32 bit float, the audio would be unrecoverable, whereas if they didn't bump up the digits as they are written, there would be no problem?
Maybe what you are saying is that 32 bit float is bumped up by a factor of 2 as part of the standard, not as part of a con by manufacturers.
Anyway, I am actually impressed that the cheap single ADC recorder I tested (Zoom H2 essential) can actually record a very loud test tone without distortion - after normalising. Even if the manufacturer is playing tricks. Or not.
-
I don't think IEEE is trying to mess with you, it's just that the multiplier bits in 32-float give a very wide dynamic range because it's powers of 2.
note that the bold in my quotes is my emphasis in my post above.
32-bit, base two means it's 32 on/off bits aka binary, aka base-2
https://en.wikipedia.org/wiki/Institute_of_Electrical_and_Electronics_Engineers (https://en.wikipedia.org/wiki/Institute_of_Electrical_and_Electronics_Engineers)
Seems like you have not reached the limits of the loudness your mics can handle yet, but making over-level-looking WAVs is sufficient for this demo.
-
The multiplier is always a power of 2, so you would never see a multiplier of 1.2.
https://en.wikipedia.org/wiki/Single-precision_floating-point_format (https://en.wikipedia.org/wiki/Single-precision_floating-point_format)
an IEEE 754 32-bit base-2 floating-point variable has a maximum value of (2 − 2−23) × 2127 ≈ 3.4028235 × 1038. All integers with seven or fewer decimal digits, and any 2n for a whole number −149 ≤ n ≤ 127, can be converted exactly into an IEEE 754 single-precision floating-point value.
I think the important word in the quote about IEEE 754 32bfp is 'exactly'. It means we may no be able to multiply the samples by exactly 1.20000000000, but instead it will be multiplied with 1.199999998796 or whatever. I also guess you may be misinterpreting the 8-bits exponent inside the 32 bits, which is used to multiply the lower 23 bits with 2^exp. This is just a way to store a floating point value, and has nothing to do with the ability to multiply two whatever floating point values. So we can multiply samples with a value (very close to) 1.2.
What floating point format would it be if a value cannot be multiplied by (something very close to) whatever other floating point value.
-
I've just being doing some very simple tests with a single ADC 32 bit float recorder, recording from its internal mics. Zoom claim that the device can record the maximum specified SPL level of the mic without clipping. When I record a test tone from a loudspeaker at a neighbour-annoying level with the device held right up to the speaker, and open it on my DAW, it's over the top and clipped.
Typically 32bfp devices have some form of digital gain control. Basically this is just some logic that simply multiplies the output of the single or dual ADC samples. This is done in the 32-bit realm, so it won't introduce clipping when the multiplied value goes over 0dbFS. On the F3, this is the 'wave magnifier' (1x, 2x, 4x, 8x etc) or whatever it is called. On my Tascam FR-AV2, it is a slider than can be set from 0dB to +40dB or something. I don't own a Zoom H2 Essential, so I just scanned its manual and it seems there is indeed no way to control the digital gain on this device. As you can get levels > 0dBFS when recording, it seems Zoom indeed picked a fixed digital gain > 1x . Is it bad intention to justify the 32bfp format? I guess we will never know, but personally I doubt that. I guess they just picked a value that gives reasonable output levels on what they deem 'normal usage'.
@Ozpeter: Reading the manual a bit further, I do notice that it is possible to export the files to 32bfp, without normalization, yet with using the 'mixer settings'. In the mixer, you can set the digital gain from -80dB to +40dB. Could it be you used that way to get your files into your DAW, unintentionally applying a digital gain > 0dB?
-
No, although I did experiment a bit with that feature this afternoon. You have to go through some hoops to process the already recorded files, and I did check the 'mixer' when recording wasn't doing anything (set to zero). I do admit that if the device was configured so that the loudest audio without mic or preamp clipping was full scale in the 32 bit float file, then quieter stuff would be rather quiet for immediate listening. Maybe that's why they digitally amplify it, as you imply.
Not directly on the immediate topic, but related, I did a little test where I recorded the loud tone from the speakers, close to analog clipping (which the H2e warns you about on its display), and also recorded some 'silence' in a quiet room under a rolled up duvet. The result seemed to me quite good when analysed. When the combined files were normalised, the 440Hz tone peaked at zero of course, and at that frequency the subsequently recorded silence was at about -100dB on a frequency analysis display. At 10kHz, the 'silence' was down to about -120dB, and slightly quieter at 20kHz. The chances of anyone being disturbed by system noise with those figures would only be likely if the recording involved a small bird 100m away, in a very quiet country location, with a lot of gain added later. Or so I think.
-
I guess they just picked a value that gives reasonable output levels on what they deem 'normal usage'.
That's my presumption.
If a hot input signal clips the input buffer or the "high level" ADC, such clipping occurs in the analog realm prior to, or in the course of digitization, and is irreversible.
If clipped via a digital increase in gain after digitization to 32bit-floating point, it's reversible.
Many are likely to confuse those two things and assume it's the 32bit float part of things that's responsible for avoiding the clipping while recording, when its actually the multiple-ADC switching scheme.. which will clip upon reaching it's realworld upper level input limit.
-
Not directly on the immediate topic, but related, I did a little test where I recorded the loud tone from the speakers, close to analog clipping (which the H2e warns you about on its display), and also recorded some 'silence' in a quiet room under a rolled up duvet. The result seemed to me quite good when analysed. When the combined files were normalised...
Assuming they match their analog and digitals component well, then one would think the upper limit of the analog stage would be just below the upper limit of the ADC (which would then output just-below-0dBFS samples). As you recorded very close to the analog stage upper limit, the resulting 'loudness' of the recorded file (so without normalization) should give a reasonable estimation of the fixed digital amplification they apply. Do you remember how much gain reduction was applied when normalizing?
When the combined files were normalised, the 440Hz tone peaked at zero of course, and at that frequency the subsequently recorded silence was at about -100dB on a frequency analysis display. At 10kHz, the 'silence' was down to about -120dB, and slightly quieter at 20kHz. The chances of anyone being disturbed by system noise with those figures would only be likely if the recording involved a small bird 100m away, in a very quiet country location, with a lot of gain added later. Or so I think.
Agree... Likely situations where one would attach a less sensitive mic (e.g. for high dynamic recordings or high SPL situations) would be more challenging. Having no analog gain control and just a single ADC would likely drop down the SNR.
-
Unfortunately I am not at home for several days now, but I will respond to above question re levels as soon as I can.