Topic: External Preamp into 32 bit (Zoom F3/F6, SD MixPre-6II, etc.)

[Ozpeter]

For personal recordings of course there may be a personal preference for the sound of a particular preamp.  But for commercial recordings, or others mainly for an external audience, perhaps it could be argued that neutrality and the benefits of the most minimal signal path should be the objective.  That is partly why I don't record for others any more.  I know my preference would be for a brighter sound which my elderly ears would actually consider flat, but which would grate on the ears of others.

[fanofjam]

I enjoy that people are finally accepting the concept of tweaking your recordings in post to improve the sound or to dial in a 'flavor' you like.  I've always done that and for so many years so many ts'ers seemed to think that was blasphemous or contradictory to capturing the 'original sound at the venue'.

[goodcooker]

I enjoy that people are finally accepting the concept of tweaking your recordings in post to improve the sound or to dial in a 'flavor' you like.  I've always done that and for so many years so many ts'ers seemed to think that was blasphemous or contradictory to capturing the 'original sound at the venue'.

I've always tried to make my recordings more "listenable" after the fact. Venues I frequent are often boomy with sketchy sound systems.

I want the recording to be the best it can be so I've gotten a lot more aggressive with editing as the years go by.

[fanofjam]

I enjoy that people are finally accepting the concept of tweaking your recordings in post to improve the sound or to dial in a 'flavor' you like.  I've always done that and for so many years so many ts'ers seemed to think that was blasphemous or contradictory to capturing the 'original sound at the venue'.

I've always tried to make my recordings more "listenable" after the fact. Venues I frequent are often boomy with sketchy sound systems.

I want the recording to be the best it can be so I've gotten a lot more aggressive with editing as the years go by.

Same...or if the sound system is fine and I'm not DFC and bass can overpower the balance.  I'm always adding a mid to middle-high boost to my recording to enhance presence which almost always gives my recordings a better balance.  That said, I sometimes have to be careful to leave the top frequencies, leave the frequencies like the high-hat alone or even back off of them a bit.  Don't want the recording sounding spitty.

Goodcooker...bet you're no different than me though where you spend, like an hour, tweeking and playing around with your original and you're happy and master everything.  Then a day or two later, you relisten and hate the master and say to yourself, WFT was I thinking?  LOL.  Well maybe not 'hate' the master but it just doesn't sound nearly as good as you thought it did at the time.  Happens to me all the time.

[Gutbucket]

[snip]..spend, like an hour, tweeking and playing around with your original and you're happy and master everything.  Then a day or two later, you relisten and hate the master and say to yourself, WFT was I thinking?  LOL.  Well maybe not 'hate' the master but it just doesn't sound nearly as good as you thought it did at the time.  Happens to me all the time.

Happens to me too, probably everyone.

To me, this reflects the fundamental moral issue at the root of the choice of doing more or not when processing the recordings we make, far more so than the arguments I often hear made for "preserving the sound as it actually was in the venue", as if any one singular unprocessed recording is really some kind of accurate document of how it sounded live.  If that actually were the case, all unprocessed recordings of the same performance would sound nearly identical, which we all know is not the case.  In any case, making it sound "more right" is closer to the spirit of the performance and enjoyment of the listeners.

The keys to minimizing bad processing choices and making sure we are making things better and not doing harm are these, I think:
1) Reliable monitoring that is flat enough, extended enough, and revealing enough to be reliably dependable in sufficiently informing us about whatever processing decisions we are making, as well as how the recording is likely to translate to other playback systems.
2) Being aware that no mater how good our monitoring, our subjective listening perception changes dramatically depending on what we are focusing on, how long we've been listening, how tired we are, etc.
3) Double checking a day or two later, ideally through an additional system, and ideally by another pair of trusted ears.

Its amazing how a day or two later, listening with fresh ears, I often immediately hear some fundamental EQ problem I'd made or missed, usually within the first few moments or notes before my brain has a chance to acclimate to the sound of the recording.

Like the sentiment, "The king is dead, long live the king".. Trust your ears, but don't always trust what you think your ears are telling you!

[unclehoolio]

I have a Zoom F6 and Oade 248 and have wrestled with this. I like the 248 sound but I think you can nearly reproduce it with FX plugins and EQ in post. And the F series preamps seem to do a very good job making a clean recording. So I've decided I want to make the cleanest original recording and preserve the ability to manipulate it as technology and my preferences evolve. I also now use an external battery with the 248 so it’s a lot of extra gear and I usually need to decide between bringing the 248 or a second pair of mics, though I could replace the 248’s internal battery if I had the time / desire.

lemme know if you decide to sell that m248 please.

ha, that was my first thought as well!

Also, I'm probably missing something important here, but hasn't the 'issue' of applying a pad (-20db or whatever it may be) to the preamp stage to accommodate line level input (and thus the related issue of 'including' the deck's preamp's flavor/color) been around since DAT recorders, if not before?  Was line in on a DA-P1 / hhb PortaDat / SV255 / D10 proII or even AD1000, etc. a different signal path than mic in?

[SMsound]

OK, so if you have a 32bit MixPre-3 or MixPre-6, then is there any reason to use an external preamp for the extra 'aux-in' 1/8" TRS channels 3/4 on the MixPre-3 or channels 5/6 on the MixPre-6? (The aux-in channels have no builtin preamp)  Should you just use a battery box? This link:

https://www.sounddevices.com/low-signal-32-bit-float/

Makes it look like having the extra 20dB of gain or whatever from using an external preamp on those extra channels is basically useless in 32bit, unlike the old 24bit MixPre 3/6 models where many of us run an external pre on the aux-in channels.

Or am I reading this wrong... It also looks like maybe the builtin Kashmir pre's on the new MixPre-3/6 don't do much in 32bit mode (no need for them).

[voltronic]

If you are using standard full-size mics, I would definitely use the best external preamps you have available to feed that aux TRS input. The way a high quality preamp loads a mic can benefit its sound beyond that of a simple battery box. The fact that you may not need to set gain very high on that external pre when the recorder is in in 32FP mode does not preclude this. I would caution you, however, to be careful of the output level of the external pre since the 1/8" TRS inputs on these recorders expect "consumer level" rather than +4 dB "professional" level. For this reason, I use the 1/8" "TAPE OUT" on my Shure FP-24 (SD MixPre) when feeding inputs such as this, rather than the XLR outputs which are much hotter level.

Setting the recorder to 32FP mode does not mean its mic preamps now aren't doing much. They still are a major contributor to the sound quality you get out of your mics. The difference when using this mode is that the gain level becomes much less important because of the autoranging ADCs (on the Zoom F series recorders the gain actually becomes fixed in this mode). The preamps are still very important.

As has been discussed many times before in threads related to these types of recorders, it is still possible to overload their input stages, regardless of 24-bit fixed or 32-bit float modes, or mic vs. line setting. All that special sauce that allows you to have increased dynamic range is happening after the input stage, meaning that the dynamic range is still partially limited by the levels that input stage can handle. In practice, you would most likely need to be recording something very loud with very sensitive/hot output mics to overload the inputs on these recorders (especially the Sound Devices units) but it is still possible. Anytime you needed to use input pads in the past, you probably still need to use them.

[SMsound]

If you are using standard full-size mics, I would definitely use the best external preamps you have available to feed that aux TRS input. The way a high quality preamp loads a mic can benefit its sound beyond that of a simple battery box. The fact that you may not need to set gain very high on that external pre when the recorder is in in 32FP mode does not preclude this. ...

Interesting, but I don't follow -- why is a preamp important/helpful for these extra MixPre Aux channels if preamp gain isn't needed in 32bit mode (as far as I can tell) and 48v phantom power can be generated by a battery box? The SD link I showed above appears to suggest that you can make great recordings in 32bit  at "inaudibly" cold levels. I feel like I'm missing something here -- on these extra Aux-In channels in 32bit mode, why use an external preamp and not just a phantom power box (putting aside from the fact that we like the extra flavor of SD's transformer pre's, and supposing we just want a clean recording on the 2 Aux tracks)?

[voltronic]

If you are using standard full-size mics, I would definitely use the best external preamps you have available to feed that aux TRS input. The way a high quality preamp loads a mic can benefit its sound beyond that of a simple battery box. The fact that you may not need to set gain very high on that external pre when the recorder is in in 32FP mode does not preclude this. ...

Interesting, but I don't follow -- why is a preamp important/helpful for these extra MixPre Aux channels if preamp gain isn't needed in 32bit mode (as far as I can tell) and 48v phantom power can be generated by a battery box? The SD link I showed above appears to suggest that you can make great recordings in 32bit  at "inaudibly" cold levels. I feel like I'm missing something here -- on these extra Aux-In channels in 32bit mode, why use an external preamp and not just a phantom power box (putting aside from the fact that we like the extra flavor of SD's transformer pre's, and supposing we just want a clean recording on the 2 Aux tracks)?

Because it's not just about gain, nor is it about "coloring" the sound (unless that's really what you want). A high quality preamp interacts with the mics in a way that will give you the best sound possible even at low gain levels.

This post explains it much better than I could, but I encourage you to go through that entire thread as it's a great discussion on the topic.
https://gearspace.com/board/all-things-technical/1419328-do-you-still-need-great-preamp-before-entering-your-audio-interface-2023-a-2.html#post16834374

[marcb]

Thanks for sharing that thread. I noticed the point about “steady current” being an important feature. I run my F6 with an external battery through the USB as the primary power source and fresh AAs in the battery sled as a backup (my understanding is if the USB stops supplying power the unit switches to the battery sled seamlessly). Should I be concerned this approach could result in inconsistent current and compromise performance?

[voltronic]

Thanks for sharing that thread. I noticed the point about “steady current” being an important feature. I run my F6 with an external battery through the USB as the primary power source and fresh AAs in the battery sled as a backup (my understanding is if the USB stops supplying power the unit switches to the battery sled seamlessly). Should I be concerned this approach could result in inconsistent current and compromise performance?

I don't think there's any cause for concern there. I would just monitor your external battery level and have a spare to swap in, because the internal AAs aren't going to last long running 4-6 mics. Or, get a couple large-capacity L-mount batteries (6800 mAh or larger) that you can hot-swap. One of these runs 6 mics on the F6 for a VERY long time.

[SMsound]

If you are using standard full-size mics, I would definitely use the best external preamps you have available to feed that aux TRS input. The way a high quality preamp loads a mic can benefit its sound beyond that of a simple battery box. The fact that you may not need to set gain very high on that external pre when the recorder is in in 32FP mode does not preclude this. ...

Interesting, but I don't follow -- why is a preamp important/helpful for these extra MixPre Aux channels if preamp gain isn't needed in 32bit mode (as far as I can tell) and 48v phantom power can be generated by a battery box? The SD link I showed above appears to suggest that you can make great recordings in 32bit  at "inaudibly" cold levels. I feel like I'm missing something here -- on these extra Aux-In channels in 32bit mode, why use an external preamp and not just a phantom power box (putting aside from the fact that we like the extra flavor of SD's transformer pre's, and supposing we just want a clean recording on the 2 Aux tracks)?

Because it's not just about gain, nor is it about "coloring" the sound (unless that's really what you want). A high quality preamp interacts with the mics in a way that will give you the best sound possible even at low gain levels.

This post explains it much better than I could, but I encourage you to go through that entire thread as it's a great discussion on the topic.
https://gearspace.com/board/all-things-technical/1419328-do-you-still-need-great-preamp-before-entering-your-audio-interface-2023-a-2.html#post16834374

The thread you linked is interesting and timely. However, Kevin seems to be saying that,  aside from clean (or colored) gain (which is apparently unneeded in 32 bit mode), the other thing a good preamp does is provide very consistent 48v phantom power. Wouldn't a simple high-quality phantom power box then be able to accomplish the same quality of recording in 32 bit mode? I may be mistaken, but I believe that a consistent 48v battery box is actually not too terribly hard to achieve compared to clean gainstages or even clean  48v power converted from A/C, like in the Pueblo/Millenia/etc. he's talking about.

On the other hand, is the point moot anyways due to the fact that the Aux-In on the mixpre's has a higher noise floor that may compete with a weak mic signal? On the original mixpre-6 (24 bit), I believe SD built their Aux-In with a basic preamp with the same electronic components as Zoom H5/H6 on the, rather than the noiseless Kashmir Preamps on their main channels. So, even if that Aux-In gain is set to zero, I'm not sure if it going through those components adds some noise that makes it into the recording (I have no idea -- just a thought).

[breakonthru]

Thanks for sharing that thread. I noticed the point about “steady current” being an important feature. I run my F6 with an external battery through the USB as the primary power source and fresh AAs in the battery sled as a backup (my understanding is if the USB stops supplying power the unit switches to the battery sled seamlessly). Should I be concerned this approach could result in inconsistent current and compromise performance?

I don't think there's any cause for concern there. I would just monitor your external battery level and have a spare to swap in, because the internal AAs aren't going to last long running 4-6 mics. Or, get a couple large-capacity L-mount batteries (6800 mAh or larger) that you can hot-swap. One of these runs 6 mics on the F6 for a VERY long time.

depends on the current draw of the mics. ive seen 4 eneloops go 3 hours with 6 mics. while thats tight for a show id have no worries running one set that way. usually keep a usb power brick in the bag for phones and misc power and will use that if the L series is low. AAs strictly backup but get used on those days you bone it and leave your F6 on and the L series is dead when you hit the venue

[Gutbucket]

OK, so if you have a 32bit MixPre-3 or MixPre-6, then is there any reason to use an external preamp for the extra 'aux-in' 1/8" TRS channels 3/4 on the MixPre-3 or channels 5/6 on the MixPre-6? (The aux-in channels have no builtin preamp)  Should you just use a battery box? This link:

https://www.sounddevices.com/low-signal-32-bit-float/

Makes it look like having the extra 20dB of gain or whatever from using an external preamp on those extra channels is basically useless in 32bit, unlike the old 24bit MixPre 3/6 models where many of us run an external pre on the aux-in channels.

Or am I reading this wrong... It also looks like maybe the builtin Kashmir pre's on the new MixPre-3/6 don't do much in 32bit mode (no need for them).

Unfortunately that example doesn't tell us anything about the 1/8" AUX-in.  Rather, it is a comparison between recording in 24-bit verses 32-bit float mode (a misleading one IMO*), with both examples using the same signal path through Mic Input 1. 


*Why misleading?  Because Sound Devices has told us that dither is not applied when writing 24 bit files with the Mix-PreII, but does not say anything about that in the examples above.  We were informed of this by the SD rep posting here upon release of the 32-bit float feature when I asked him specifically about it.  When I pressed further asking why not, he did not say.  As best practice, dither should be always applied anytime bit depth is reduced.  Its application eliminates the quantization noise easily heard and seen in the 24-bit example above.  SD does properly apply dither when writing 16-bit files. If they had also included a 16bit example above, we would expect the noise floor to be significantly higher (9 bits higher than in the 24bit example), yet the piano note would decay cleanly into that noise floor without any quantization noise.  If dither had been applied to the 24bit file the same thing would occur, except with with a noise floor 8 bits lower than in the 16bit example.  In comparison to the provided 24bit example without dither applied, the noise floor would increase by one bit and the horrible sounding quantization noise would be eliminated. 

Apparently SD made the decision that it was unnecessary to apply dither when deriving 24bit files from the 32bit float ADC.. presumably because any situation where this might be an issue never actually arises in real life use!  Ok, no problem.  Design & engineering decisions are all about fine tuning the fit between real world requirements and a real world implementation.

The problem is that SD can't have it both ways at once.  It either matters or it doesn't.  Their design & engineering department apparently decided it doesn't matter - enough so to not bother implementing known best practice, while their marketing department finds it compelling to claim that it does - providing the example above as evidence. 

What the above example actually provides is evidence that SD doesn't apply dither when writing 24bit files and and example of what that sounds like.  If I was more cynical I might suspect this design choice was made intentionally to help market 32-bit float by way of this kind of non-real-use comparison, but the primary take away of all this for me is a strong confirmation of the maxim- Never let the truth get in the way of good marketing.

I trust their engineering team, and not their marketing team so much.