24bit recording exported as 32bit

[PortaStudio]

Hello,

I see descriptions of recordings that say they were recorded in 24bit and exported in 32bit. What benefits do 24bit recordings exported as 32bit have? My DAW (cubase) allows me to work in up to 64bit, while my audio interface only allows for 24bit recordings. What are the pros and cons?

Thanks in advance!

 

[twsecrest]

I'm not the expert, but I believe going from 24-bit to 32-bit makes the audio a little louder.

 

[PortaStudio]

I'm not the expert, but I believe going from 24-bit to 32-bit makes the audio a little louder.

Yes, it does increase the overhead in relation to the noisefloor. So you get more amplitude in relation to the noisefloor out of it.

When mixing 24bit material in 32bit, it should be a cleaner result, right?

 

[twsecrest]

Yes, it does increase the overhead in relation to the noisefloor. So you get more amplitude in relation to the noisefloor out of it.

When mixing 24bit material in 32bit, it should be a cleaner result, right?

Nope, it would not make it any cleaner than the original 24-bit recording.

 

[PortaStudio]

Nope, it would not make it any cleaner than the original 24-bit recording.

Okay, it's a bit odd, because I stumble across recordings that say they were recorded at 24bit but exported at 32bit. I don't understand the thought-process behind that...

 

[yurqqa]

I'm not the expert, but I believe going from 24-bit to 32-bit makes the audio a little louder.

Have you ever received a product in a nice tight box that is packed into a bigger cardboard box ?
Basically the same thing - takes more space with no benefit for the original product.

Ok. In case of the box it at least protects the original packaging, which function was to protect the original product.
So, again, not too useful.

 

[levimax]

Hello,

I see descriptions of recordings that say they were recorded in 24bit and exported in 32bit. What benefits do 24bit recordings exported as 32bit have? My DAW (cubase) allows me to work in up to 64bit, while my audio interface only allows for 24bit recordings. What are the pros and cons?

Thanks in advance!

You can't "improve" anything by up sampling ... what ever information was recorded is what it is. The advantage to higher bit depth has to do with processing and mixing, it can be done better (less added noise and other issues) at high bit depth. So recording at 24 bit and then up sample and process at 32 or 64 bit and then down sample to 24 bit or even 16 bit for distribution makes perfect sense.

 

[PortaStudio]

You can't "improve" anything by up sampling ... what ever information was recorded is what it is. The advantage to higher bit depth has to do with processing and mixing, it can be done better (less added noise and other issues) at high bit depth. So recording at 24 bit and then up sample and process at 32 or 64 bit and then down sample to 24 bit or even 16 bit for distribution makes perfect sense.

Interesting. I'm considering doing this as well. I will have to do some testing. I know that a higher bit depth means lager file sizes, but will it be heavier on the CPU while processing?

Cubase asks me whether or not to upscale imported audio if it is at a lower sampling rate than the project. I don't know if Cubase will do the same for different bit depths? If yes can I leave them as is, or should I let Cubase rerender them?

Another question I have is, if I have 24bit material, is mixing it in 64bit overkill in contrast to 32bit?

 

[mkt]

Proverbial grain of salt if not the whole shaker

Lower Resolution, Better Sound Quality?

TRPTK • Recording the extraordinary. High-resolution, immersive recordings of contemporary artists in classical music and jazz. Buy hi-res downloads up to DXD 32bit, Dolby Atmos TrueHD, DTS:X, Auro-3D, Hybrid SACDs, and vinyl.

trptk.com

 

[twsecrest]

Okay, it's a bit odd, because I stumble across recordings that say they were recorded at 24bit but exported at 32bit. I don't understand the thought-process behind that...

My two cents.
How about, lets say you own a music streaming business and offer CD-audio quality (16-bit/44.1k) audio.
(you bought a bunch of music CDs and copied them to a HHD, on your music server)

And you want offer higher than 16-bit/44.1k quality music, for higher paying customers.
So you take the 16-bit/44.1k audio and rerecord it as a 24-bit/96k music file and put that on your music server.
So now you charge customers a high price for listening to those 24-bit/96 music file, even thou it's really the same quality as the original 16-bit/44.1k recording.
Or even better rerecord the 16-bit/44.1k at 32-bit/192k and charge even more

Also could be the person who exported 24-bit audio at 32-bit, really had no good reason (beneficial) for doing it in the first place.

 

[sarumbear]

There is some confusion on this thread. Here is a good article that explains the bit depth in layman terms.

What Is 32-Bit Float Audio, and Should You Record In It?

Is it the end of setting levels? Just another recording gimmick? Well, it's somewhere in between.

www.wired.com

 

[levimax]

Interesting. I'm considering doing this as well. I will have to do some testing. I know that a higher bit depth means lager file sizes, but will it be heavier on the CPU while processing?

Cubase asks me whether or not to upscale imported audio if it is at a lower sampling rate than the project. I don't know if Cubase will do the same for different bit depths? If yes can I leave them as is, or should I let Cubase rerender them?

Another question I have is, if I have 24bit material, is mixing it in 64bit overkill in contrast to 32bit?

Sorry I am not a Cubase expert. I know Audacity which is a free program automatically upsamples to 32 bit float for processing and then you chose how to export. I would check some Cubase forums for ideas and of course test them out.

 

[PortaStudio]

Proverbial grain of salt if not the whole shaker

Lower Resolution, Better Sound Quality?

TRPTK • Recording the extraordinary. High-resolution, immersive recordings of contemporary artists in classical music and jazz. Buy hi-res downloads up to DXD 32bit, Dolby Atmos TrueHD, DTS:X, Auro-3D, Hybrid SACDs, and vinyl.

trptk.com

That is a great article! They record in DSD256, mix and master in PCM 352.8kHz (at 705.6kHz editing software causes issues) 64bit floating point, and sell the product as such (with additional downgraded versions).

Since DSD256 is (as far as I know) superior to 192khz 24bit by a lot, I wonder what the best case scenario in my case would be. I can record in 192khz 24bit at best. Cubase (as far as I remember) allows me to mix my project in 32bit floating point, 32bit fixed point and 64bit fixed point. Would it be 64bit fixed point? Coming from a 192khz 24bit source is there a point in going that high? If not should I pick 32bit floating point or fixed point?

If Cubase asks me to upscale the 24bit recordings, should I do so or can I leave them as is?

Sorry I am not a Cubase expert. I know Audacity which is a free program automatically upsamples to 32 bit float for processing and then you chose how to export. I would check some Cubase forums for ideas and of course test them out.

Yes, I should ask there if processing is upscaled within the DAW like you've described. But if I process the audio in a higher bit depth and there are vst instruments (not sample based) in the mix, it might be better to do the whole project in a higher bit depth anyway and exported as such.

 

[MaxwellsEq]

Doing processing in 32 or 64 bit minimises rounding or truncating errors. The finished file can then be downsampled to a usable store or transmit format. Upsampling can't make the source file better, though.

 

[HarmonicTHD]

That is a great article! They record in DSD256, mix and master in PCM 352.8kHz (at 705.6kHz editing software causes issues) 64bit floating point, and sell the product as such (with additional downgraded versions).

Since DSD256 is (as far as I know) superior to 192khz 24bit by a lot, I wonder what the best case scenario in my case would be. I can record in 192khz 24bit at best. Cubase (as far as I remember) allows me to mix my project in 32bit floating point, 32bit fixed point and 64bit fixed point. Would it be 64bit fixed point? Coming from a 192khz 24bit source is there a point in going that high? If not should I pick 32bit floating point or fixed point?

If Cubase asks me to upscale the 24bit recordings, should I do so or can I leave them as is?


Yes, I should ask there if processing is upscaled within the DAW like you've described. But if I process the audio in a higher bit depth and there are vst instruments (not sample based) in the mix, it might be better to do the whole project in a higher bit depth anyway and exported as such.

Read what sarumbear linked. Most DAWs work at higher bit depths. In very simple words one avoids the mathematical rounding errors of all the calculations when filtering, effects adding etc basically all you do in a DAW before going back to 24 bit. It does not however „improve“ sound quality. 32 vs 64 bit is not gaining you any sound quality either. It just taxes your CPU which can lead to much more (and even audible) problems. So stay at 32 bit and rather focus on the creativity of it all ;-)

(BTW. Hardly anyone if at all is able to distinguish btw 16 and 24bit. But 24bit just gives you the ultimate peace of mind. Do the math what 24bit DR really means)

 

[PortaStudio]

My two cents.
How about, lets say [...]

A commercial CD is a mastered digital format and there is no further processing to be done.

Doing processing in 32 or 64 bit minimises rounding or truncating errors. The finished file can then be downsampled to a usable store or transmit format. Upsampling can't make the source file better, though.

VST instruments should benefit. Maybe delay and hall effects as well? If it put a considerable amount of weight on the CPU, it might not be worth it to run the project in 32bit for me anyways.

In very simple words one avoids the mathematical rounding errors of all the calculations when filtering, effects adding etc basically all you do in a DAW before going back to 24 bit. It does not however „improve“ sound quality. 32 vs 64 bit is not gaining you any sound quality either. It just taxes your CPU which can lead to much more (and even audible) problems.

Okay, I will test how much heavier a higher bit depth is on the CPU, just to know the difference.

So stay at 32 bit and rather focus on the creativity of it all ;-)

I assume you mean 24bit in my case.

 

[HarmonicTHD]

A commercial CD is a mastered digital format and there is no further processing to be done.

VST instruments should benefit. Maybe delay and hall effects as well? If it put a considerable amount of weight on the CPU, it might not be worth it to run the project in 32bit for me anyways.

Okay, I will test how much heavier a higher bit depth is on the CPU, just to know the difference.

I assume you mean 24bit in my case.

24bit output. The DAW (at least mine) does 32bit for internal calcs anyhow. You can play with the sampling frequency if CPU becomes bogged down and causes excess delay or even audible glitches (meaning 48khz is plenty. No need for 96 or even higher. ).

 

[PortaStudio]

24bit output.

Okay, I will keep the project at 24bit then.

The DAW (at least mine) does 32bit for internal calcs anyhow.

I'm sure cubase is doing this as well.

You can play with the sampling frequency if CPU becomes bogged down and causes excess delay or even audible glitches (meaning 48khz is plenty. No need for 96 or even higher. ).

Thank you very much!