Jump to content

Preferences: 24 Bit Recording


darkecho

Recommended Posts

i would think its only for the bit depth converters work at,

and 32floating is the bit depth internal engine works with, i dont think it makes any sense exporting in 32bit audio since most of the programs have a different way of using that extra 8bits, which would result in incompatibility.

i think the 8bits is only used for the internal headroom the audio engine works with and its not written on the files itself? or is it?

i cant recall but i think cubase/nuendo dont write 32bit files, rather 24bit.

 

besides, your hardware does not support 24bit, so the audio written with 32bit floating point is actually still 24bit audio. i cant think of any audio interface that supports more than 24bit.

so if you take into consideration ONE file which did NOT clip while recorded (if it does then it is corrupted any way), adding extra 8bits wont give you anything except bulk data

Link to comment
Share on other sites

Why can't I save files at 32bit? This setting is for the bit depth of recorded and frozen audio, correct?

I don't know about frozen audio, but recorded audio is definitely saved at 24-bit if that option is checked. A quick look at your audio bin will confirm this. (Click Media, then Bin).

 

Search this forum for further explanation on why recording in 32-bits is not necessary. 24-bits provides all the headroom humans will ever need. And if your search doesn't satisfy you, pick up a copy of Nika Aldrich's "Digital Audio Explained", and turn right to the last chapter on digital audio myths. The preceding chapters serve to support the conclusions drawn in that chapter.

 

BTW Cubase is capable of recording audio files in 32-bit float format.

Link to comment
Share on other sites

24 bit recording is all you need. You would be wasting hard drive space trying to record more bits.

 

If you want a quick example: let's say you have 1 quart of water you want to bring to me. You can choose between a 3 cup recipient (slightly less than a quart), a gallon recipient, or a 10 gallon recipient. Which do you choose?

 

Well the 3 cup will not allow you to transport the whole quart of water. That's the 16 bit setting.

 

The gallon is more than you need for 1 quart, but at least you're not losing any water. That's the 24 bit setting.

 

The 10 gallon recipient is too large, heavy to handle, difficult to maneuver... it still allows you not to lose any of your quart of water, but doesn't have any advantage over the 1 gallon recipient.

 

That's why Logic doesn't offer you a 10 gallon recipient (32 bit floating point audio files).

 

Freeze files, however, are 32 bit floating point. That's because you don't necessarily optimize the levels of a freeze file, which could be peaking at more than 0 dBFS, or less than -50 dBFS. But when you record or bounce, you don't want to clip (above 0 dBFS), and you usually optimize the level so it doesn't have a super-super weak level (like peaking at -60 dBFS). So in that case, 24 bits are more than you need already.

Link to comment
Share on other sites

To migrate all my Cubase files into Logic in OMF I had to convert all the files from 32 bit to 24 bit... Also I can't simply import any of my Cubase files from the finder if they are 32 bit so I tend to agree with Ploki that perhaps different apps use the extra 8 bits in different ways or perhaps it is just that Logic doesn't support saving the files this way even though the internal engine uses the 32 bit float... At this point I am not concerned anymore since I just finished migrating all my Cubase projects over to Logic this weekend... lot's of work and the mix settings and automation are all gone but that's OK ... just wanted to have them ... gonna get rid of Cubase altogether now... :D

Logic is too much fun to use ... and I find it easier on the eyes the way it is laid out...

Link to comment
Share on other sites

  • 1 month later...

Having said that, what is the best (highest audio quality) way to transfer audio from Live 7 to Logic 8?

 

1) Set live to record in 24 bit (instead of 32) then export all audio files using the 24-bit option with triangular dithering and then open these files in logic?

 

2) or do the same but leaving the Live recording setting to 32-bit and export to 24-bit using triangular dithering?

 

3) or rewire and bounce each track from inside Logic (which takes a long time and is not very stable having both applications on sometimes)

 

4) Another way I do not know about?

 

What is the best quality wise? There always seemed to be a little loss of quality when exporting files from Live to Logic. Logic does not accept Lives 32 bit files, so should I record in 24bit in Live and export, or record in 32bit and export in 24bit, or rewire (in this case should I have recorded live in 24 bit or 32 bit before?)...

 

This is something people do a LOT, transferring from Live to Logic. However I have not seen a definitive answer to this so far... What is best please????

 

Thanks

Link to comment
Share on other sites

1) AFAIK no software records 32 bit audio files. Surely no audio converter create 32 bit audio signals.

 

2) No idea what "triangular dithering" is, but you shouldn't use dithering to go from 32 bit floating point to 24 bit. I assume we're talking about 32 bit floating point?

 

3) No!

 

Basically, if you want to export audio files to be used in Logic, export them as 24 bit fixed audio files. That's it. No dithering (dithering is to convert from 24 bit fixed to 16 bit fixed).

Link to comment
Share on other sites

Well, imagine that the 32 bit is for internal calculations within a program itself. And that can be helpful for things like reverbs (especially convolution reverbs) as well as EQ and the rest. These things are basically mathematical calculations and so by increasing the bit depth while doing those calculations you will get more exactly what it is you want.

 

But there would be no real purpose for making a 32 bit audio file for a lot of reasons: 1) Its main advantage would be to have more dynamic range, but 24 bit already covers sounds so quiet you can barely hear them to things so loud you could barely stand to be around them... more of that sort of thing wouldn't really help you, 2) I don't think there are many things that can handle a 32 bit integer audio file at the moment, CD's are of course 16 bit, God knows why in 2008, and DVD's are I believe 24 bit. I know they aren't 32 bit or any higher than 24. I wonder what bit depth a Blu Ray disc is? 3) So you would have to downgrade your audio quality once or maybe twice (possibly more) to accommodate what people are going to play it on, and every time you do that you add noise and lose fidelity.

 

I would never use 16 bit except if I knew something was going to end up on the web or something as an mP3 anyway, and just hell with it. Or if I was just doing human voice, no singing. Or recording my girlfriend at night to prove to her she snores... oh wait, I don't have a girlfriend. If anyone can think of why let me know!

 

And as far as Live, why don't you just export it as 24 bit and import it into Logic the same way? Nothing is getting out of Live at 32 or 64 bit there either... that's all internal to Ableton as well. I think when you see a program willing to save at 32 bit floating point, or 32 bit integer even, then it is designed to be opened with another professional program at some point, not made to be a consumer format.

Link to comment
Share on other sites

Thank you, basically:

 

1) 32 bit record - this is an option in Live's preferences, which can either be 16/24 or 32 bit recording. I used to leave this setting at 32 and then export the files at 24bit, however, since I am exporting it in 24 bit I have now changed the recording setting to 24bit also

 

2) triangular dithering - any audio exported out of Live below 32 bit requires you select a dithering 'mode', and for some reason triangular is recommended. So yes, even for exporting at 24bits the program asks the user to choose a dithering mode

 

I tried different things to get the best results. Having everything set at 24bit has been ok. However, it is not yet clear to me if you get the best audio quality by exporting the audio files from live to logic, or by rewiring both programs and bouncing each track from inside Logic. Is one better than the other for audio quality?

 

Like I said before, the export function in Live forces the user to choose a dithering mode EVEN in 24bit. Does this mean it is better to Rewire?

 

Thanks

Link to comment
Share on other sites

1) OK I'm not familiar with Live, but I'm not sure WHAT you're supposed to record at 32 bit, since there are no existing 32 bit A/D converters. Doesn't make any sense to me?

 

2) You can't dither from 32 bit floating point to 24 bit fixed, period. Doesn't make any sense. Think about it that way: "What's the most precise way to record the value '01101001" using 8 bits? Hmmmm... how about... '01101001'?? See, no dithering required. Dithering is a mathematical "trick" software can do to lessen rounding errors when trying to record a value with LESS bits than the original. From 32 bit floating point to 24 bit fixed point, you're not changing the number of bits coding the value of the signal. From 24 bit fixed to 16 bit fixed, you are changing the number of bits, hence the option to dither.

 

I guess you could dither from 32 bit fixed to 24 bit fixed, but again I've never ever heard of 32 bit fixed audio recording.

 

Like I said before, the export function in Live forces the user to choose a dithering mode EVEN in 24bit.

My guess is Live only use the dithering choice when you're exporting to 16 bit.

 

So keep it simple, export as 24 bit, don't worry about the dithering settings - they most probably don't do anything (or do your own tests, my guess is you won't hear any difference).

Link to comment
Share on other sites

Thanks, maybe that explains why I did not like the results when recording in 32bit and converting to 24bit with dithering during the export. It seemed the sound lost a bit of punch. When I found this bit option in recording preferences I imediately changed it to 24 bit, and the sound seems more faithful to the original thing when exporting. Now it is all 24 bit. Between doing that (exporing/bouncing) and rewiring not sure what sounds best, hope it is the same like ploki suggests, however technically they are two different processes since bouncing the program asks to dither (even to 24bit I guess because lives engine playsback at 32 bit floating) and in rewiring there is no such thing. Anyway, its cool it already sounds better recording in 24bit and exporting the same. Thanks for the feedback. I love both programs and want to get the most out of working between them. Met Logic recently and cannot live without it
Link to comment
Share on other sites

Basically, Ableton is able to do work with 32 bits, again, I'm sure this is all internal to Ableton itself. I can't see if this is supposed to be 32 bit floating point or fixed integer. My assumption is that this is floating point, and will have to be made into a fixed integer 24 bit file on export or render- whatever Ableton calls it.

 

And there won't be any way of actually recording any sort of real world audio this way since nothing at the moment can actually record anything over 24 bit. There could be some sort of reason to record a software instrument and effects at the Ableton 32 bit level, and it's nice they give that option and have an eye for the future maybe, but it certainly isn't being exported at that level. Because nothing except a computer can work with 32 bit files, floating point especially.

 

So it's a little disingenuous of Ableton possibly to let everyone think they are recording at 32 bits when they in fact are not, but I still think they have a great product in every way... hope this was helpful, L

 

The picture below is from Ableton's data sheet, by the way.

163735522_Picture1.png.9c8336d47f64ec80313d3c4acb16a228.png

Link to comment
Share on other sites

thank you for the feedback, so it needs to be fixed 32 to make sense, according to your explanation and david's example with the 00100s... ok, will just be using 24 bit, will experiment rewiring against exporting more often to see if the results are any different for better or worse.. tks
Link to comment
Share on other sites

  • 2 weeks later...

It seems there are a lot of misconceptions here concerning internal bit-depths and dithering.

 

Firstly, yes, any audio you record will have no more than 24 bits per sample, as that's the maximum that any normal ADC can provide. Logic, Ableton and most other DAWs work in at least 32bit resolution though, because as soon as you start manipulating that audio with EQs, reverb, etc you need the extra internal resolution to prevent rounding errors and distortion during the processing steps. Some DAWs offer the option of exporting the full 32bits if you expect to process the audio further later on, though in the vast majority of cases a 24bit export is more than sufficient.

 

I don't know where people got the idea that you can't dither from 32bit float to 24bit. You certainly can and should apply dither each time you move from a higher bit-depth to a lower one if you want the best results. (David: 'triangular dithering' refers to the triangular probability density function that is commonly used to generate the dither noise. Other PDFs (square, Gaussian) have been tried in the past, but triangular is accepted to sound the best.) When you dither you will also be presented with the option to apply noise shaping. It's best to use only a light amount of noise-shaping (or none at all) when exporting a 24bit file. This is because the dither noise sitting at the bottom of 24bit audio is truly inaudible and that audio will almost certainly be converted to 16bit down the line for distribution, when the final dithering and noise-shaping will be added. (The much lower dynamic range of 16bit audio means that dither noise can sometimes be audible, which is why noise-shaping was introduced to move it away from the frequencies to which the ear is sensitive.)

 

So, in summary:

1) Yes, check the '24 bit recording' checkbox

2) If you just want to save your recording, and haven't applied any sort of processing to it, then export as 24bit with no dither, as the audio is still pure 24bit.

3) If you've added some EQ, reverb, or any other form of processing, then your audio has become 32bit and you need to compensate for that, so set Resolution to 24 Bit and Dithering to POW-r #1 (Dithering) (Or use the dithering on a mastering plugin, as long as it's the final plugin in the chain on your master output and the faders on the master ouput are set to 0).

 

The articles STILL linked earlier in the thread include a good overview of bit-depth and dither.

Link to comment
Share on other sites

Thanks Charleski for the helpful info... I would like to hear your comment on what is best considering the following workflow:

 

When recording and arranging with efx and eq in Live and mixing and finalizing with further efx and dynamics in Logic what is best?

 

1) need to choose the bit depth to record in live (from 16bit to 32bit), I currently record in 24bit

 

2) after recording with eq and efx in live and finalizing the 'first draft arrangement' (as things might change a bit in loigc) I need to send the files to logic. What is the best to do this: rewire to Logic or render and then open the files in Logic?

 

3) in case rewire is best ok,

 

4) in case render is best or equal (which would be great as it is more practical than rewiring), what is the best rendering setting to match the decision in point 1. In other words, if you find it better to record it in 32bit, what is the best rendering option to send to logic remembering logic doesn't accept 32bit rendered files, I assume render to 24bit using triangular dithering. However is this better than having recorded in 24bit in the first place and just keeping it 24bit without dithering when sending it do Logic?

 

My goal is to work with both applications keeping the audio quality the highest possible and necessary (I don't record over 48)...

 

Thank you

Link to comment
Share on other sites

1) need to choose the bit depth to record in live (from 16bit to 32bit), I currently record in 24bit

Stick with that (assuming that your ADC is 24bit). Your audio will be coming in at 24bit, it only turns into 32bit once you apply effects, change gain, etc.

[Edit]I just realised that, because I haven't used Live much, I may be misinterpreting your question here. If you're exporting your recording without any processing, use 24bits and export without dither. If you're doing EQ etc on the take before exporting it, make sure that Live is operating in 32bit-mode, and then dither down to 24bits on export, as discussed below.

 

2) after recording with eq and efx in live and finalizing the 'first draft arrangement' (as things might change a bit in loigc) I need to send the files to logic. What is the best to do this: rewire to Logic or render and then open the files in Logic?
I must confess I haven't used ReWire much at all, and would only think of using it if I needed to have two different programs processing audio in a chain with both running at the same time. I find getting things right is hard enough with just one program though, and try to avoid increasing the complexity and chance of glitches. I had a brief look around but couldn't find any firm specs on the ReWire channels, though I suspect they steam output at full 32bit resolution. So if you're transferring audio that has been processed in Live you probably would retain some quality using ReWire.

 

BUT:

If you're recording tracks, it's always a good idea to keep a copy of the untouched recording, if you can. I've often found myself wanting to go back and redo things, and that's a lot easier if I have an archive copy of the original, untouched by whatever mangling I might have drunkenly applied at the time ;).

 

So, when you do your recording, I'd advise recording at 24 bit without any effects, gain-change or fading, and saving each take out as a 24bit file with no dither. Once you have your takes in the can, reimport them into Live, apply whatever processing you want, then export as 24bit with triangular dither and low or no noise-shaping. This audio can then be imported into Logic for further work. You could use ReWire at this stage, but the difference between a properly-dithered 24bit file and a 32bit stream will be imperceptible - 24bit has a dynamic range of 144dB, which is well beyond that of our ears. Without dithering the 24bit file you get distortions which can mount up, especially if you go through the process a few times.

 

Of course, there will be times when you need to apply the processing to the audio as it's recorded, because of time- or space-constraints. In that case, save the takes as 24bit files with dithering as above (since the processing converted the audio into 32bit). Remember, it's only when you process the audio in your DAW that it becomes truly 32bit, but once that's happened, it's best practice to dither the audio on exporting down to 24bit. 'Process the audio' here means any form of fade, gain-change, EQ, reverb, compression, etc - basically, unless the audio is passing straight through your DAW with faders on 0, no inserts and no mixing, it's been processed, and contains 32bit information.

 

In a sense, we are really splitting hairs here, as the distortion introduced by simply truncating 32bit audio to 24 bits will be extremely low-level. The distortion is much more noticeable if you work with truncated 16bit intermediates. But if you want to be sure you're retaining as much quality as possible, dithering is easy to apply on export and won't harm your signal as long as you don't apply extreme forms of noise-shaping.

Link to comment
Share on other sites

Thanks so much for your reply Charleski.

 

I do a lot of processing in Live before going to logic. Diferent perspectives really, where the processing in Live for my workflow has more to do with soundshaping and finding the right elements, whereas in Logic I also process these same sounds in a Mixing perspective... its not true for all of my projects since some are 100% live or logic... however it is true for the ones that raised the topic...

 

Given your feedback, I understand the best thing for me is to record straight at 32 bit and then export at 24 bits with triangular dithering. From your explanation, I understand this is better than recording processed sounds at 24bits and sending at 24bits to logic without dithering.

 

For some reason it seems the sound looses a bit of punch when I open it in Logic, regardless of the method. I have also heard this from some popular DJs with a similar workflow. The same is not true when doing a project end to end in both DAWs. Something happens between them which we do not understand. I will try rewiring also to see if it is any different. However, like you said, rewiring can be tricky depending what you have connected to your computer. I find it only really works with stability when used only to record tracks from one DAW to other, and once they have been 'bounced', better to shut rewire off before opening new instruments and plug ins...

 

Thank you...!

Link to comment
Share on other sites

For some reason it seems the sound looses a bit of punch when I open it in Logic, regardless of the method. I have also heard this from some popular DJs with a similar workflow. The same is not true when doing a project end to end in both DAWs. Something happens between them which we do not understand.
It might be an idea to check the levels, a small lowering of sound level often won't be obviously softer, but will have less punch. Get a sound-level meter ($50 from RadioShack) and generate a pure sine-wave test tone (Audacity can generate the tone for you and save it as a 32bit .wav). Play the tone back in Live and measure the sound-level from your speakers, then save it as a dithered 24bit file and play that back in Logic. The sound levels should be exactly the same, if there's any difference you'd need to compensate. There might be something else going on, but this is the first thing I'd check.
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...