24 bit.....where to begin??

[audiobot202]

I'll hopefully be installing a new M audio 2496 card soon and as I've never worked in anything other than 16/44.1 before...I'd like to know what would give best sound results with the new card?
I've heard people using 24/48k and those using more than that. What are the pros and cons?

Steve

[C00kie]

Pros:
* more accurate recordings at higher bit rates (increased dynamic range) and higher sample rates (increased performance for high frequencies).

Cons:
* you have to convert it back to 16/44 if you want to put it on a standard (redbook) audio CD. Sample rate conversion can be a problem, for bit depth you can truncate or dither.
* with a S/N ratio of 100dB on the audiophile 2496 there is only one bit of the extra 8 filled with significant info (16bits has theoretically a S/N ratio of 96dB, 6dB for each bit)
* 24bits files take up twice as much disk space as 16bits. 24bits is actually stored in 32bits.

[IIRs]

If you are working to a CD, stick to 44.1KHz, but use high bit depths (24 or 32 bit) until the final stage.

That's my advice anyway.. feel free to try upping the sample-rate and trying to hear a difference. (I haven't really been able to do this as my cpu can't handle a full mix at high sample-rates)

If you do go higher though, and you still intend the end release for CD, consider running the card at 88.2KHz instead of 96KHz, as the sample-rate conversion of the finished mix will be less damaging..

[audiobot202]

So, what would be my best bet for this particular card taking into account the S/N ratio issue?
Rendering to 24/48k and then mastering in Audition using the Ozone plugin which contains a dithering algo...then saving from Audition to a 16/44.1k file...do you think there'd be any conversion problems or is there a better workaround?

[audiobot202]

If you are working to a CD, stick to 44.1KHz, but use high bit depths (24 or 32 bit) until the final stage.

That's my advice anyway.. feel free to try upping the sample-rate and trying to hear a difference. (I haven't really been able to do this as my cpu can't handle a full mix at high sample-rates)

If you do go higher though, and you still intend the end release for CD, consider running the card at 88.2KHz instead of 96KHz, as the sample-rate conversion of the finished mix will be less damaging..

Yes. Finished work is for CD. OK...so 24/44.1 would provide theortically more headroom and more dynamics? The only difference I can hear betwen 44.1 and 48k is a very very slight sheen at 48k at the top end of the sound.

[headquest]

Hi

As a general rule its good to use the highest quality your hardware/software will allow for recording in. There will be some deterioration of sound so it's good to start with the best you can get.

When rendering ready for CD or Mp3 it is essential to "dither", a process that prevents unwanted artifacts creeping in when you reduce the sample and bit rate back down to 16/44.

If you burn a CD in Adobe Audition (which is basically what I know and use) then this happens automatically as part of the burning process (and you can select all the parameters according to what you want). I believe that most top-drawer software has this option.

Hope this helps

[valley]

Yes. Finished work is for CD. OK...so 24/44.1 would provide theortically more headroom and more dynamics? The only difference I can hear betwen 44.1 and 48k is a very very slight sheen at 48k at the top end of the sound.

If you are mastering to 44.1 don't work at 48. Any small benefits you may gain from teh higher sampling rate will be lost when you down sample, and you'll be introducing interpolation noise that needn't be there.

As platinumears says above, if you must work at higher sample rates, choose one that is a direct multiple of your target rate. In this case, 88.2.

[UltraJv]

If you are working with vsti, forget anything other than 44.1 (going higher wont guarantee any better quality and your cpu usage will increase) If you are recording audio, do a few test wavs of the intruments/vocals you will be using, one set at 16 bit/44.1 and the others at 24 bit 96. If you cant hear or dont like the change then stay with 16 bit. Some people claim a difference but that dosnt mean its more accurate at higher rates. All this is pointless of course if you render to mp3s. You will be throwing it all away.

[audiobot202]

Thanks all. The masters will be wave files, not mp3s. I'll do some tests as you suggest and see if I can hear a real world difference.
I aim to render the project from Tracktion without dither or conversion and leave that final process for Audition.
I believe that some VSTi support higher sample rates....couldn't name them though.

[IIRs]

If you are recording audio, do a few test wavs of the intruments/vocals you will be using, one set at 16 bit/44.1 and the others at 24 bit 96. If you cant hear or dont like the change then stay with 16 bit.

The main point to recording at 24 bit is to allow headroom without loss of resolution. Recording at 16 bit with 6db of headroom means you will only use 15 of your bits..

Some people claim a difference but that dosnt mean its more accurate at higher rates.

Higher resolution is more accurate by definition.

[audiobot202]

So, theoretically then I could record all audio at 24/44.1 and have the benefit of that increased headroom? Even at 44.1, will it make a real difference in terms of dynamics?
It's the catchment of detail and dynamics that I'm after here but not looking to gum up the processing power too highly.

Steve

[UltraJv]

Higher resolution isnt more accurate, its the old marketing numbers game - it depends on the quality of your AD/DA section and has nothing to do with bits. Do you think Creatives cards are accurate-they do run at 24 bits....?

[valley]

So, theoretically then I could record all audio at 24/44.1 and have the benefit of that increased headroom? Even at 44.1, will it make a real difference in terms of dynamics?
It's the catchment of detail and dynamics that I'm after here but not looking to gum up the processing power too highly.

Yup.

The great advantage of 24 bits is that you can give yourself stacks of headroom to elminate any risk of clipping, whilst still producing an audio file that is greater than 16bits in resolution.

[audiobot202]

So, theoretically then I could record all audio at 24/44.1 and have the benefit of that increased headroom? Even at 44.1, will it make a real difference in terms of dynamics?
It's the catchment of detail and dynamics that I'm after here but not looking to gum up the processing power too highly.

Yup.

The great advantage of 24 bits is that you can give yourself stacks of headroom to elminate any risk of clipping, whilst still producing an audio file that is greater than 16bits in resolution.

So, in practice does this mean that I can pump the levels up in Tracktion and have no fear of clipping at the master output? And in this case, I would be able to preserve the full range of 44.1 and produce a render that has more detail?
Does 16 bits limit even the audio detail of a sampling rate of 44.1?

[audiobot202]

Just found this interesting quote:

"The PCM format provides its optimal resolution when signal levels are at their very highest. As signal levels decrease to lower levels, resolution deteriorates, leaving quiet cymbals and string instruments sounding typically sterile, dry, harsh, and lifeless. The more bits you have available to you in the process of quantizing the amplitude of a waveform at any given sampling, the more accurately a lower level signal can be represented." (24 bit recording FAQ)