Topic: Tascam FR-AV2 new compact 32bit float recorder.

[grawk]

With photography, it's artistic choices that are enabled by using manual mode.  You can do something different than the system is going to do, so your pictures are demonstrably different, and that can't be fixed in post.

With 32 bit audio vs 24 bit, using just digital gain, once normalized both files are identical, unless you got clipping in 24 bit.  So you lose by using 24 bit, rather than gain.  But that said, I usually record in 24 bit because that's what the gear I use in the field does, unless I'm bringing the big recorder.

[Gutbucket]

This looks interesting!
https://reverb.com/item/85619654-earsight-the-one-for-tascam-fr-av2-stereo-field-recording-microphone

Dig that all-in-one with a removable baffle idea. 3D printing nicely applied, as seems to be the case in a number of the Immersive Soundscape products.  Should be relatively easy for DIY'ers with 3D printing skills to emulate.

[aaronji]

Or maybe support the guy who developed and implemented the idea...

[Gutbucket]

Sure.  Implemented at least.  Old idea we were DIY'ing around here at TS 20 years ago.

[Ozpeter]

... I usually record in 24 bit because that's what the gear I use in the field does, unless I'm bringing the big recorder.

I usually record in 32 bit because that's what the gear I use in the field does, unless I'm bringing the big recorder.   In my case the choice would probably lie between the Zoom H1 XLR and the dear old Tascam 60D - the former of course being radically smaller than the latter. 

Actually, these days if I had to make a paid-for classical music recording like in the good old days, when it often just needed a Sennheiser MS pair in the right spot, I would connect the mics to the the Zoom, then connect the output thereof to the Tascam (because standard practice in the old days was to use two recorders for overall backup).  The Zoom annoyingly has no MS settings, but the Tascam does, so for replay to artists I could use the Zoom but listen to it through the Tascam, with its MS monitoring mode.  You can even screw the Zoom on top of the Tascam to make a kind of one-piece device.  But I digress - again...

[grawk]

I’d use the Tascam as the preamp and the zoom for the recording in that scenario. For me the lower noise floor matters most.

I use either the dpa dvice into my phone or the deity pr2 for most situations and the Sonosax r4+ when I can bring the big rig.

[Ozpeter]

I’d use the Tascam as the preamp and the zoom for the recording in that scenario. For me the lower noise floor matters most.

The Tascam DR-60D (Mk I which I have, or the later Mk II) noise spec is -120dBu or less, while the Zoom H1 XLR is -122, so the Zoom has a slight advantage (partly I guess because it's about ten years younger!).  I do think that Zoom is a bit of an under-rated product, and maybe using the H1 designation was a mistake, as it makes people expect it to be noisy, and it isn't.  Of course the FR-AV2 noise level is -127dB, for which you pay!

[TheJez]

Sorry, but I think the airbag analogy is ridiculous. A better analogy is the Canon A-1 SLR camera that was introduced in 1978. That was the first professional 35mm camera to feature a fully automatic exposure system. There were already cameras with either aperture priority or shutter priority, but the A-1 was the milestone that evolved to the "anybody can take a good photo" goal of all photography equipment after that. Of course, composition and balance and artistic  qualities still had to come from the skill of the photographer, so when the masses could run out and get a fully auto slr by the early 1980s, suddenly there were a million amateur photographers still taking shitty photos, but now they were not over or under exposed. Even after Nikon, Minolta, and Pentax jumped on the automatic camera bandwagon, their top line professional cameras could still be switched to full manual with a good old center weighted match needle light meter.
 
Why would they make cameras with fully manual controls if automatic cameras could that work better. Well, it's that the "better" part only applied to amature weekend snapshot types, pros knew how to select shutter speeds and apertures using a hand held light meter and their industry was not screaming "help we need an automatic camera to make all the settings for us". The only reason automatic exposure systems were found in the top end cameras afterwards was purely for the purpose of selling them to pro wannabes who had no clue how that stuff worked. Same thing happened a decade later with auto focus. No pro ever said "I wish I didn't have to focus manually". 32bit float is a cool feature, but it's not to someone who rarely if ever needed to have, IMO of course.

Hmm, well, I guess when using the camera analogy, then automatic exposure seems to me more like the auto gain control or limiter. Then everybody can make recordings without clipping. Going to 32bfp seems like more like they finally invented a film or camera sensor with such a high dynamic range that you can take photos where shutter time and amount of light don't matter anymore. Even the brightest lights with extremely long exposure times won't over-expose, and the darkest scenes can be shot with very short exposure times and still capture all the details in the dark. Only when you decide to print it on paper with its limited dynamic range (i.e. convert your audio to 24 or 16 bits), you have to determine what to do with the shot (i.e. compress, normalize, whatever) to fit it in the limited dynamic range of the target. There are no real artistic considerations regarding setting the record level, its just something you have to do 'good enough for the circumstances', which can be skipped completely when recording to 32bfp. Other stuff (getting your stealth equipment in, positioning, mic orientation etc) I do like to consider a work of art 
Don't get me wrong: I would have loved to have a recorder that can record in 32bfp yet still allows to set the analog gain! It would re-introduce the risk of clipping in the analog domain (and therefor re-introduce the 'art of setting the record level right'), but also allows to optimize the gain stage for lower noise.

[Ozpeter]

My short way of arguing the merits of 32 bit float (which I'm sure we have discussed to death in these forums over the years anyway) is this - if we'd had 32 bit float from the outset, and then one of the manufacturers suddenly brought out the first ever 24 bit audio recorder together with new fangled gain controls to make sure we didn't clip stuff etc, would we all have rushed to buy it, on the grounds that it would be a better way of doing things?  Personally I don't think so, but of course some might disagree!

[aaronji]

The value of these recorders is not in the 32-bit floating point aspect, which is just a convenient storage format, but the use of multiple ADCs. Multi-ADC designs in 24-bit, such as the original MixPres, offer pretty much the same advantages (wide dynamic range).

[rastasean]

Sure.  Implemented at least.  Old idea we were DIY'ing around here at TS 20 years ago.

Exactly....just like the birth of your oddball microphone configuration DIY project.

[unidentified]

My short way of arguing the merits of 32 bit float (which I'm sure we have discussed to death in these forums over the years anyway) is this - if we'd had 32 bit float from the outset, and then one of the manufacturers suddenly brought out the first ever 24 bit audio recorder together with new fangled gain controls to make sure we didn't clip stuff etc, would we all have rushed to buy it, on the grounds that it would be a better way of doing things?  Personally I don't think so, but of course some might disagree!

Totally agree

[aaronji]

Exactly....just like the birth of your oddball microphone configuration DIY project.

Well, to be fair, the Immersive Soundscapes product, containing, as it does, mics, a 3-D printed body, and Neutrik connectors, is a pretty far cry from taping a piece of cardboard in between your handheld's on-board mics. And baffling microphones has been a thing for, what, 100 years? That didn't arise from taperssection.com...

Immersive also has an ORTF version with cardioids. And they are available for the MixPre, PMD661 (MKII), and F3. They both spec a max dBSPL of 110 dB, though.

[Gutbucket]

Exactly.  Not a new idea - neither here or there.

Long live DIY.. AND long live the folks who provide gear for others who don't have the desire or skill to do it themselves.  Fortunately there's room in the world for both.  As well as a long history of support for both in the history of amateur concert taping.

Edit- ..and those two worlds converge in interesting ways.  A key demographic here at TS has long been members who build things for other tapers.  Huge thanks to them.  Immersive Soundscapes extends that in an interesting way by making available products well within the reach of some DIYers thanks to the development of 3d printing which supports both DIY and low-quantity commercial production efforts.  Win win, and a cool convergence. I'm all in for the consolidation of good ideas and the implementation of them by both means when the result is better recordings.

[datbrad]

I see how my photography analogy fell apart due to the complexities that differ between the two technologies. The intent of my first post was merely to share my experiences with different sensitivity microphones through the FR-AV2. After reading through this thread again, there is definitely a "set and forget" camp standing strongly behind 32bit float as the answer to that need.

Here is what I know. Max input specs on recorders are based on gain set to minimum. If you use a hot mic, 10mv/pa to 15mv/pa, you will have to turn the gain down all the way at a rock show. No control of balance, and no real sense of headroom. However, if you run a low sensitivity condenser or a dynamic microphone and are able to apply -any- gain, you know that your input levels are unlikely to push past the +4 threshold. So it doesn't matter if the gain is analog, digital, or a combination, being able to add 15-20db of gain tells you the incoming signal from the mic has near zero chance of overloading the recorder mic inputs. It also doesn't matter if you are running 24bit or 32bit, the benefits of having more control are the same. One thing I have learned over the years with recorders is they all have an optimal gain level. I think the optimal level for this recorder is revealed by the +18db factory default gain setting. The thing that's great about this recorder is it's broad feature set and applicability across wide array of uses and users.