Topic: Tascam/Zoom/Deity 32bit recorders

[mackie80]

Hi there, finally registered after reading a lot recently about 32bit recording.

I own a H1essential and recently had a Tascam DR-10L Pro.
I had to return the Tascam as it had a strange quirk - it lost ALL settings if not turned on every 12-24hrs, even if it has full batteries. And without having time & date set, nothing works, esp not the one click recording. Dunno if it was only that particular one or if that is a general problem.

It's being said the 10L Pro does have Dual ADC, while the other Tascam handheld recorders (05XP and 07XP) only have a single ADC. I really don't understand why they do that.

I am actually tempted by the size the 10L Pro, Zoom F2 and Deity Pr-2 have, as sometimes you just need something that small.
In the Deity thread I read, it is not even clear what kind of ADC is built in it and people talk about noise after pulling recordings up.

I recorded shows with both the Tascam and the Zoom and both performed pretty well when pulled up the quieter sections (which for me is the main reason to use 32bit float as the shows I record usually have a big dynamic range from acoustic to loud). Of course, the Tascam being mono only is a drawback, but still better than not having a recording at all.

I would like to know how the Tascam DR-05XP and 07XP perform compared to the H1essential

and how the Tascam 10L Pro compares to the Deity PR-2 and the Zoom F2 - like is the Tascam probably already the best having a dual ADC nad give it another shot or should I try Deity or Zoom? I'm a bit torn right now (that settings thing was kinda nerv wrecking).

[Gutbucket]

Deity PR2 uses switching ADCs and writes 32bit output files only in mono mode.  If recording in 2 channel stereo mode it uses a single ADC per channel >24bit.  That said, it makes good recordings in 24bit stereo mode as long as you set recording levels appropriately, and is the smallest recorder I know of that provides 5V mic power.  Its noise-floor has not been a problem for me when recording typical taper concert material via mic-in or a SBD feed via line-in, as long as gain is set properly.  I've not attempted to use it for very quiet classical music recording yet though.

[mackie80]

well i know the small ones (Tascam, Deity and Zoom) are only mono in 32bit. and i am interested in how the 32bit recording quality compares.

I know the 5V feature which makes the Deity somewhat special in that size range. My mics only need 3V max, so the 2.5V any other recorder offers, is fairly enough and the 5V is not of a big interest for me.

[ICatchSounds]

well i know the small ones (Tascam, Deity and Zoom) are only mono in 32bit. and i am interested in how the 32bit recording quality compares.

Not sure what you mean by "only mono in 32bit" as they make stereo files. I can tell you as an owner of a new Tascam DR-07XP, that the 32bit level is set so low that all but the loudest tacks need boosted like crazy just to hear.  You are almost always better off recording in 24bit at the proper level that using 32bit.  The only way 32bit works well with the Tascam DR-07XP is for close voice recordings or loud distand sounds. When you bump the 32bit track up much at all it is loaded with self noise.

[Ozpeter]

Different 32 bit float devices create mono or stereo files - the smaller the device the more likely it is to only record in mono, as a crude generalisation - check the specs before purchase.  I'm sorry your experience with the Tascam recorder has not been good when using its 32 bit float format.  It only uses a single ADC so maybe that's the problem.

[Gutbucket]

[snip..]
I can tell you as an owner of a new Tascam DR-07XP, that the 32bit level is set so low that all but the loudest tacks need boosted like crazy just to hear.  You are almost always better off recording in 24bit at the proper level that using 32bit. [..snip]

A fundamental assumption when using 32-bit floating-point recording is that levels will always be adjusted (raised) afterward.  If you prefer not to do that then stick with 24-bit and manually adjust recording gain so as to peak at a reasonable level that doesn't require raising afterward except by way of the playback volume knob.

Expanding on what Peter said, a number of small low-cost 32-bit recorders are designed to use the limited resources of a single two-channel ADC and the two-channel circuit path leading up to it in two different ways.  They can either record two separate channels in 16 or 24bit in stereo, or both ADC channels can be used together to record a single 32-bit floating point mono channel.

[snip..]
When you bump the 32bit track up much at all it is loaded with self noise.

The test for you to make with your Tascam DR-07XP will be recording something in 32-bit mode and in 24-bit mode, match levels afterward on the computer, then compare the noise levels between the two.  Boosting levels will always boost all content of the recording including the noise. The important point of comparison for you will be if the noise is similar in both cases or louder in one in comparison to the other.

[ICatchSounds]

[snip..]
I can tell you as an owner of a new Tascam DR-07XP, that the 32bit level is set so low that all but the loudest tacks need boosted like crazy just to hear.  You are almost always better off recording in 24bit at the proper level that using 32bit. [..snip]

A fundamental assumption when using 32-bit floating-point recording is that levels will always be adjusted (raised) afterward.  If you prefer not to do that then stick with 24-bit and manually adjust recording gain so as to peak at a reasonable level that doesn't require raising afterward except by way of the playback volume knob.

Expanding on what Peter said, a number of small low-cost 32-bit recorders are designed to use the limited resources of a single two-channel ADC and the two-channel circuit path leading up to it in two different ways.  They can either record two separate channels in 16 or 24bit in stereo, or both ADC channels can be used together to record a single 32-bit floating point mono channel.

[snip..]
When you bump the 32bit track up much at all it is loaded with self noise.

The test for you to make with your Tascam DR-07XP will be recording something in 32-bit mode and in 24-bit mode, match levels afterward on the computer, then compare the noise levels between the two.  Boosting levels will always boost all content of the recording including the noise. The important point of comparison for you will be if the noise is similar in both cases or louder in one in comparison to the other.

"The test for you to make with your Tascam DR-07XP will be recording something in 32-bit mode and in 24-bit mode, match levels afterward on the computer, then compare the noise levels between the two.  Boosting levels will always boost all content of the recording including the noise. The important point of comparison for you will be if the noise is similar in both cases or louder in one in comparison to the other."

I have made that exact test. My point is that since this is a single ADC, Tascam sets the 32bit super low to stop clipping. If I match 32 bit and 24 bit recording the exact same sound at the exact same level, the 32bit version is always way noisier than the 24 bit file. 

[Gutbucket]

Good to hear you performed that test to make certain, as its the only way to make sure.

Welcome to the real world of engineering where results are confirmed by measurement and actual use - which unfortunately all too often gets obscured behind a curtain of marketing and popular perception confirmed by rhetorical persuasiveness rather than by a clear understanding of objective data.  The sad thing to me is that kind of thing still occurs even with these kinds of products, where you'd expect actual performance would be more important to what I imagine is a more generally astute and engineering minded customer base.

[ICatchSounds]

Good to hear you performed that test to make certain, as its the only way to make sure.

Welcome to the real world of engineering where results are confirmed by measurement and actual use - which unfortunately all too often gets obscured behind a curtain of marketing and popular perception confirmed by rhetorical persuasiveness rather than by a clear understanding of objective data.  The sad thing to me is that kind of thing still occurs even with these kinds of products, where you'd expect actual performance would be more important to what I imagine is a more generally astute and engineering minded customer base.

Yes, it is just sad that the performance of the DR-07XP is that much worse in 32bit vs 24bit. If i had have know that upfront I would have saved $60 and bought the non -  32bit version.

[Gutbucket]

Agreed! 

Hey Youtube reviewers, shine some light on this stuff and hold the manufacturer's marketing departments to task!  Need to call them out on this kind of thing.

[Ozpeter]

Agreed! 

Hey Youtube reviewers, shine some light on this stuff and hold the manufacturer's marketing departments to task!  Need to call them out on this kind of thing.

Perhaps ICatchSounds could simply upload his test results to YouTube - these days we can all be reviewers!  Though probably Mr Judd would get more views...

[Ozpeter]

Presumably as the higher the preamp gain, the lower the noise (as I understand it), then that could account for the difference in noise between the low-gain 32 bit float recording and the high gain 24 bit recording.  In an ideal real 'pure dual ADC 32 bit float' recorder, I would hope that the quality of the signal chain would be worked backwards, so that the stages before the ADC were up to the 32 bit float challenge.  In a device offering other bit rates and with gain controls actually working on the preamp, maybe that results in compromises.  Maybe devices offering multiple bit rates should be avoided. (I'm just putting the thought out there, not claiming necessarily to be right!).

[ICatchSounds]

I discovered one more issue I have with the Tascam Dr-07XP.  I like to use a set of stereo 3.5mm Clippys in my small recorders. The Tascam only puts out 2.17 volts. I made a video comparing the 3.5 voltage to some other recorders and it was the lowest by a good margin.