Can anyone confirm if Reaper will successfully import 32-bit float files from the likes of the Zoom F6 or Sound Devices MixPre-6 recorders. According to Sound Devices, Reaper v6 will but not sure about v5. Many thanks
Can anyone confirm if Reaper will successfully import 32-bit float files from the likes of the Zoom F6 or Sound Devices MixPre-6 recorders. According to Sound Devices, Reaper v6 will but not sure about v5. Many thanks
Reaper has always supported floating point files (both 32 and 64 bit). (At least if they are standard WAV-files. You should check if the Zoom or Sound Devices produce some unusual format.)
__________________
I am no longer part of the REAPER community. Please don't contact me with any REAPER-related issues.
I've just downloaded a few 32 bit float files from the Sound Devices site and imported them into v5 and they seem to play just fine. Sound Recordists I know are starting to use the Zoom and Sound Devices recorders so I just wondered in advance if I might encounter issues in Post but actually all seems fine. Reaper does it again. Phew! Be interesting to know if any of you are starting to receive these types of files now the prices of the recorders are becoming more reasonable.
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
I'm using 32-bit float all the time now. It's perfect for showing up to live classical events and not needing to worry about setting levels. The only thing I need to focus on is good microphone placement. And, yes... Reaper definitely handles the files just fine. I'm in the habit of processing the raw files through SoundDevice's Waveagent first but I probably don't have to.
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
Nope. The SoundDevices and Zoom devices utilize multiple convertors with a dynamic range so wide that the microphones would clip before the device does. The gain knob in 32-bit float is post-fader and simply acts as a volume knob. If the printed waveform "clips" visually it is always recoverable.
Just for fun I have recorded concerts with the "gain" knob all the way down and all the way up. Same quality after adjusting the float part.
Your analog-to-digital converter is integer and it will clip at 0dB.
Quote:
Originally Posted by bachstudies
Nope. The SoundDevices and Zoom devices utilize multiple convertors with a dynamic range so wide that the microphones would clip before the device does. The gain knob in 32-bit float is post-fader and simply acts as a volume knob. If the printed waveform "clips" visually it is always recoverable.
Just for fun I have recorded concerts with the "gain" knob all the way down and all the way up. Same quality after adjusting the float part.
Your analog-to-digital converter is integer and it will clip at 0dB.
The Zoom F6 2 converters = 32 bits but I suppose they are integer. I suspect in coming years DAW interfaces might start showing up with similar and eventually become the norm.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 01-24-2020 at 11:42 AM.
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
One of the points that has been made elsewhere is that while 32-bit float is indeed excellent for worry-free live recording (as I said, I can attest!), 24-bit will work for every other case. There's plenty of headroom to work with. I aim for -12 dB peaks if I get to soundcheck beforehand but there are plenty of opera recording engineers who will go a LOT lower and are still fine and dandy.
So, will 32-bit float appear on DAW interfaces? Yep from a purely marketing point of view like 192k sample rates do (more harm than good!). Will the average DAW user at home ever need it? Nope.
From what I've read most 24-bit converters give you about 20-bits of accuracy. "Good luck" getting 32-bits of useful-meaningful information.
Quote:
with a dynamic range so wide that the microphones would clip before the device does.
I don't know anything about the Zoom design but usually an interface is "gain staged" so the preamp has plenty of headroom and the ADC gets clipped before the preamp clips. It's a bummer if your analog signal is clipped while the meters on your interface & DAW are showing plenty of headroom!
The thing you can't control is the head amp in a condenser mic. (That's why they sometimes have a pad.)
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
I think there's some misunderstanding here. We are after 24-bit plus the 8 bits consisting of the floating volume part (= 32-bit float). With 2 or 3 convertors the ability to record without setting level is certainly possible without any loss of accuracy! There's clearly some magic happening beneath the surface that only the companies themselves know about but all I can say is that the Zoom and Sound Devices work as advertised. Gain setting is, indeed, unnecessary in 32-bit float mode. You are essentially only limited by the dynamic range of your microphones.
Last edited by bachstudies; 01-24-2020 at 12:47 PM.
Correct and understandable since everyone has been taught otherwise until these newer devices came out that combine two converters - I'd have answered the same had I not known. Maybe DVDDoug can take a look at them so that he's up to speed.
__________________ Music is what feelings sound like.
Well that interesting cause as far as I know there is no such thing as 32bit float in the "real world" outside a daw. 32 bit integer yes, but not float.
But still, how can it matter if you are clipping the AD stage? Those are actual electronics and have nothing to do with the bit depth of the internal storage.
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
The end 32-float file is a composite from the 2 or 3 convertors depending on the device manufacturer. I'm not sure of the exact why they divide and conquer but essentially anything clipped is ignored and magically taken from one of the other padded converters. Think of it like the Zoom F8 safety record on secondary channels but completely automatic.
I'll repeat...it is absolutely fantastic for live concert folks who don't get chance to soundcheck AND it is primarily designed to be beneficial for video on-location sound folks who don't want to re-calibrate between capturing a whispered voice and a plane taking off.
But still, how can it matter if you are clipping the AD stage? Those are actual electronics and have nothing to do with the bit depth of the internal storage.
There is a direct correlation to the electronic parts and the bits derived. Even the quantization noise is a byproduct of having to assign values for voltages that fall between resistors, over to the closet resistor (bit). There's a heck of a lot more to it than that but in simplest terms it just analog parts turning on bits based on how high the discrete voltage at some point in time is.
IF the converters were 32 bit float, if you saved those bits to disk, they the would be one in the same. As far as preamps go, they have the DR they have. Not sure how they are handling that part.
I'm guessing a lot, let's see how far I was off.
__________________ Music is what feelings sound like.
The end 32-float file is a composite from the 2 or 3 convertors depending on the device manufacturer. I'm not sure of the exact why they divide and conquer but essentially anything clipped is ignored and magically taken from one of the other padded converters. Think of it like the Zoom F8 safety record on secondary channels but completely automatic.
I'll repeat...it is absolutely fantastic for live concert folks who don't get chance to soundcheck AND it is primarily designed to be beneficial for video on-location sound folks who don't want to re-calibrate between capturing a whispered voice and a plane taking off.
Yea it sounds like they have made a smart thing here for sure.
There is a direct correlation to the electronic parts and the bits derived. Even the quantization noise is a byproduct of having to assign values for voltages that fall between resistors, over to the closet resistor (bit). There's a heck of a lot more to it than that but in simplest terms it just analog parts turning on bits based on how high the discrete voltage at some point in time is.
IF the converters were 32 bit float, if you saved those bits to disk, they the would be one in the same. As far as preamps go, they have the DR they have. Not sure how they are handling that part.
I'm guessing a lot, let's see how far I was off.
Yes for sure, I was only talking about physical clipping in active electronics.
Thank you to all of you for commenting. I have read with interest all the replies to my original question and appreciate some of the more theoretical discussions. It is good to know that Reaper will import 32 bit float files. I shall not be concerned when presented with such files from Sound Recordists in the field knowing one can work swiftly within Reaper.
Having recorded sound in many uncontrolled situations, machines such as the MixPre and Zoom F6 can only help make our job easier, to provide clean location recordings where circumstances dictate one only gets one go at it. Example - recording a building being demolished!! No matter how experienced we are I think we can all agree there are always situations where the unexpected can happen.
I do take the point about guarding against overloading ones mic head amps. I guess the moral there is to choose the most suitable mics for the job in hand. Thank you also for the link to previous discussions on the subject.
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
The question that remains for me is whether I need to bother going through Wave Agent to sew the mono tracks into interleaved stereo (as appropriate) for Reaper. In Sequoia and Pyramix I believe I used to be able to just drag in as a poly wav but something in my arsenal of DAW software changed at some point and found myself reaching for SoundDevice's Wave Agent as an interim step and I've continued to use it ever since. I started with a Zoom F8 though and can't remember whether the files were simple mono vs poly. Oh well.
The question that remains for me is whether I need to bother going through Wave Agent to sew the mono tracks into interleaved stereo (as appropriate) for Reaper. In Sequoia and Pyramix I believe I used to be able to just drag in as a poly wav but something in my arsenal of DAW software changed at some point and found myself reaching for SoundDevice's Wave Agent as an interim step and I've continued to use it ever since. I started with a Zoom F8 though and can't remember whether the files were simple mono vs poly. Oh well.
Well, seeing as this thread has established that Reaper will import and play 32 bit float files, be brave and see if you can do everything in Reaper from the off. Why do you need to interleave the mono tracks to stereo? If you import them into Reaper you can always group them and work on them as a stereo source. Then when you are happy simply render to a stereo mix.
If your files are poly WAV's simply import into Reaper. Work with them that way if you wish or explode them within Reaper to mono files. Very easy to do.
If you do want to combine mono files to interleaved stereo, Kenny has a video somewhere that shows how.
I sense you are using a SD Mix Pre for your recordings. If you or anyone else on here, has any 32.bit float files recorded from the Zoom F6 I would love to have a sample to prove to me that Reaper is happy working with them. Then when I get an enquiry from the sound recordists threatening to send me Zoom F6 files I can hopefully say - no problem.
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
Quote:
Originally Posted by pksdp
Well, seeing as this thread has established that Reaper will import and play 32 bit float files, be brave and see if you can do everything in Reaper from the off.
Perhaps. From what I can tell, without having any poly wavs to hand on my drive, Reaper doesn't not treat them in the same way I'm used to in Pyramix. I don't like the prospect of jerry-rigging channels from a single track. Plus, I am often editing individual spot microphones to aid clean section entries etc. Now I'm remembering why Wave Agent (free!) is so handy. I also like having the main microphone pair as an interleaved stereo file before it enters the DAW to avoid user error.
EDIT: The Zoom F8 was excellent because I could choose poly wav or individual mono/stereo tracks. Plus, I could record both ways using the dual SD slots!
Perhaps. From what I can tell, without having any poly wavs to hand on my drive, Reaper doesn't not treat them in the same way I'm used to in Pyramix. I don't like the prospect of jerry-rigging channels from a single track. Plus, I am often editing individual spot microphones to aid clean section entries etc. Now I'm remembering why Wave Agent (free!) is so handy. I also like having the main microphone pair as an interleaved stereo file before it enters the DAW to avoid user error.
EDIT: The Zoom F8 was excellent because I could choose poly wav or individual mono/stereo tracks. Plus, I could record both ways using the dual SD slots!
Sometimes working in a way that gives you confidence is a good thing, sometimes it's useful to work out of your comfort zone. You may be pleasantly surprised and learn an easier way of doing things.
Just thought I'd link this for anyone interested. the dual converters + 32b float is tremendously appealing to me, as I had never heard of anything quite like this capability before. I really want the Zoom F6 for live, field recording, but, alas, have 649 reasons not to get it quite yet:
He also has another walkthough of the F6 that goes even further in depth.
I believe someone found a couple more budget interfaces where the sample rate clock had more measured jitter at 192k vs 96k. We already have the sample rate very far away from the audio data with 96k. So that was just degrading quality with increased jitter. (The problem of needing steep eq filters next to the audio band is already removed with 88.2k or 96k sample rates.)
Weather or not that shootout was controlled and true and what product that was I don't know. That was the premise of how a higher sampling rate could cause problems.
There's the part where it increases CPU use to deal with extreme HD vs HD audio but that's a 'use of resources' problem with your computer.
Unless there's some compromising workflow behind the dueling AD converters and stitching the data sets together for the 32 bit stream, this technique is legit. You're digitizing the full range of your analog head amp. Record whispers and shotgun blasts without riding the mic trim.
Careful with playback of those recordings! Good way to take a speaker or ear out!
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
Quote:
Originally Posted by future fields
Exactly, and the budget interfaces performing bad at 192khz is a budget interfaces problem.
I think yes to this but also there's an issue with 192k in general that devices can pick up all the electronic interference that plagues public venues these days. I'd much rather cut this out at the point of recording versus trying to deal with it later. 192k at any price point is just absolutely crazy when 88 or 96k is plenty to begin with. I'm actually doing more direct to 44.1k or 48k these days and finding things work out a lot better in mixing and mastering.
The Zoom F6 2 converters = 32 bits but I suppose they are integer. I suspect in coming years DAW interfaces might start showing up with similar and eventually become the norm.
That struck me too. Both Sound Devices and Zoom claim 32 bit float. I can't see why floating numbers would add anything.
Besides, we're happy with 24 bit, so 32 bit should be ample without being a floating number. Internally, for a DAW, it makes sense.
Quote:
Originally Posted by DVDdoug
From what I've read most 24-bit converters give you about 20-bits of accuracy. "Good luck" getting 32-bits of useful-meaningful information.
20 + 20 = 40. Good enough to extrapolate 32 bits, I suppose? Even the worst case, 18 bits combined still starts from 36 real bits. I haven't found any docs on the specifics of the process, tho...
Quote:
Originally Posted by DVDdoug
I don't know anything about the Zoom design but usually an interface is "gain staged" so the preamp has plenty of headroom and the ADC gets clipped before the preamp clips. It's a bummer if your analog signal is clipped while the meters on your interface & DAW are showing plenty of headroom!
The thing you can't control is the head amp in a condenser mic. (That's why they sometimes have a pad.)
The first thing that goes beserk in a condenser mic, at extreme levels, is the capsule itself, usually. The amp is mostly an impedance converter with maybe a few dB of gain. Sure, it'll distort, but that might end up passable. Sticking the membrane to the backplane, otoh, often results in tiny holes in the membrane.
Of course, you don't have to use a condenser, I guess.
Designing a regular mic amp with 130 db dynamic range is economically impossible. So they use the same trick: two preamps. If the lower amp/ADC is clipping, the digital signal comes from the higher level amp/ADC anyways. Still not as simple to do as a regular preamp, but the only way to get that close to the theoretical maximum of analog gain's S/R ratio.
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
Exactly, and the budget interfaces performing bad at 192khz is a budget interfaces problem.
It's the filtering that costs...
BTW for those worrying about jitter, every now and then, I come across a device that actually reduces jitter a lot and doesn't cost an arm and a leg. In this case, an auto SPDIF switcher on Tindie:
Both Zoom and Sound Devices strongly imply that their interfaces have 32-bit floating point signal flow end to end. If they are doing some internal conversion from fixed point then yes, this is just a marketing gimmick. But they do explicitly claim the actual recording (not just the theoretical file format) has effectively no dbFS ceiling.
So either SD uses another unknown chip (unlikely) or uses dual ADC's, as has been done for over a decade. And these still don't do float. Zoom explicitly mentions dual ADC's, as does SD's patent.
You can of course combine the two ADC values and output a floating number. Still, the floating part isn't coming from the hardware, but from a calculation.
Besides, have you read SD's explanation? It is very unclear. The entire text is about files and refers to DAWs. In fact, it shies away from even mentioning hardware.
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
I highly doubt SoundDevices or Zoom are willing to give away their hardware secrets so easily. I forget which way round it is but one company uses 2 ADCs and the other 3. Yes, we know enough that a 32-float value is calculated from the regular and padded ADCs. I'll repeat: in the real world, your microphone will clip before the device does.
There are some excellent SoundDevices videos explaining the 32-float mode. There are a couple that show the effect of zero "gain" and full "gain" (essentially just a post-fader volume control). No quality loss and zero need to actually set a gain at all.
There has been extensive debate about this elsewhere and I'm totally convinced there's nothing shady going on. As mentioned, this is hardly new tech but its implementation in portable sound recording devices is amazing. I have successfully trialed in 32-bit float mode and can vouch for the quality of recording for classical concerts. I'm pretty sure Tony Faulkner uses the MixPre10 for live classical.
Last edited by bachstudies; 01-27-2020 at 02:04 PM.
There has been extensive debate about this elsewhere and I'm totally convinced there's nothing shady going on.
Could you please show me where?
I've been searching everywhere and all I seem to find is the same mantra. Almost ad verbatim the text from Sound Devices. Even the slutz aren't awake yet.
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
It's a bit of a long read but worth it. I think I might have posted some of the SD videos there too. In any case, there are far more knowledgeable people talking about the ins and outs of it than me. But, in the end, I stand by my latest conclusions...
The preamp has nothing to do with the number of bits, one is analog and the other is digital. In side the device is a 24bit ADC those 24 bits map directly to the 24bit part of the 32 bit float. That is a 24 bit int has exactly the same information as a 32 bit float so far as the static representation of an audio signal. The 32 bits float only becomes useful when processing the signal with effects or when mixing. The actual bit depth does not change but remains 24 bits of actual audio information. Most audio interfaces from a semipro on up have 24bit (int) ADC and that is mapped to a 32 bit float for use inside the computer. Ardour uses 32 bit floats throughout for all audio, even audio from 16 ADCs is converted to 32bit float. There really is no difference between 32 bit float and 24 bit int, they are a different representation of the same thing. That is, 24 bit int converted to 32 bit float converted to 24 bit int should be bit perfect. The use of 32bit float in this device allows keeping all 24 bits of information intact at different levels. That is 24 bits in from two or three ADCs at different levels can only be expressed with 32bit float while keeping 24 bits of information. Most ADCs are single and so can express the input directly with 24 bit int. In both cases the resolution is still 24 bits.
So it's not 32 bit float, but 24 bit float?
Quote:
A device that has 32bit float outputs has a chain that looks like this:
Audio in->preamp with gain control->24bit int ADC->internal cpu->24bit->32bit float->to user’s computer. The gain control must be in the analog part of the circuit in order to make the maximum use of the bit depth without clipping. With two (or more) ADCs it would be possible to have the level control set two levels one you see and one that is hidden and exactly 20db down (or some other know value). The ADC at the higher level would be used for the output most of the time but if that adc goes over full scale the second one at 20 db down could be inserted by increasing it’s level by 20 db and switching to use that. The 32bit output would then faithfully represent the portion of the waveform that was over 0dbfs. However during that clip the bit depth would actually drop from 24 bits to 20-ish bits. It would still sound better than a digital clip. Also, it would allow recording at a higher level overall meaning a better use of the 24 bits available. However, it is questionable if regaining the last 4 bits is worth anything if the limits of the audio circuitry is only 19 to 20 bits anyway. There is not any way that using 32 bit float will help reveal preamp sound. The normal way to do that is to use a separate preamp that has the desired sound, set it’s level for that sound and then set the output level for correct ADC level in.
The person who wrote the above is called Len. Len Moskowitz?
Len seems to agree. 32 bit float is a marketing fibs...
I'll have to do with 24 bit float, I fear
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell