24 bit vs 16 bit

krizrox · Post by **krizrox** » Fri Feb 27, 2004 8:27 am

No this isn't a question about which is better.

Yesterday, I was doing a recording session for a client with Sonar and at some point during the session (near the end) I realized I had Sonar configured for 24 bit recording. I had been playing around with it prior to the session and didn't set it back to 16 bit. I didn't see it as a problem per se, but then I noticed I was using the 16 bit ASIO drivers in Pulsar. I started to worry at that point.

I thought it was strange that Sonar would even let me record anything in 24 bit unless there was a corresponding 24 bit ASIO driver loaded. From what I can tell, it didn't really seem to matter.

It seemed as though Sonar was actually recording at 24 bit but I suspect the resolutiuon was limited to 16 bit because of the ASIO drivers I was using.

What am I confused about? I admit stupidity on this subject. Any comments?

Oh by the way: I had to export the files out as 16 bit WAV files for the client to take with him. What did I dither? A 16 bit waveform recorded at 24 bit? I listened to the WAV files and they sounded fine although I don't know if I actually might have caused some degredation of the sound in the process.

Post by **garyb** » Fri Feb 27, 2004 12:56 pm

afaik,you made 16bit wave files with an extra 8bits of headroom in sonar. the comversion to 16bit should have been painless....

Shayne White · Post by **Shayne White** » Fri Feb 27, 2004 2:19 pm

I've compared files with dithering and files without dithering, and frankly, I can't hear any difference. If there is, it's negligible. I wouldn't worry about it!

Counterparts · Post by **Counterparts** » Mon Mar 01, 2004 5:20 am

I had a similar thing happen to me in Cubase.

I was using the 24-bit ASIO drivers in SFP, and in the 'device setup...' dialog, Cubase reported using these 24-bit drivers.

I was playing around with looking at various menus and things when I noticed that Cubase was actually set-up to record 16-bit wav files. Worse than that, 16-bit 'Broadcast' Wav files (urgh!) - this seems to be the default. The situation was quickly corrected!

Royston

samplaire · Post by **samplaire** » Mon Mar 01, 2004 6:39 am

Worse than that, 16-bit 'Broadcast' Wav files (urgh!) - this seems to be the default. The situation was quickly corrected!

AFAIK the broadcast one has just a different header and the quality is the same as 'plain' wav. ProTools on Mac produces broadcast wavs only (no option for normal wavs)

<font size=-1>[ This Message was edited by: samplaire on 2004-03-01 06:40 ]</font>

Counterparts · Post by **Counterparts** » Mon Mar 01, 2004 9:45 am

That's right.

As I mentioned recently in the "Pulsar Creativity" forum, it's a file format which has suffered somewhat from "too many cooks" - there are lots of different 'chunks' which can be present in the file, usually amounting to lots of different ways of acheiving the same result!

The sample data itself is always contained in a chunk labelled 'data', which consists of interleaved sample frames.

Royston

Neutron · Post by **Neutron** » Sat Mar 06, 2004 1:19 am

I think increasing to 24 or more bit over 16 makes more difference than higher sampling frequency..but thats just me

esp when it comes to mixing 16 bit sources into the creamware mixer and recoord in 24 or 32 bit. sounds much better.

<font size=-1>[ This Message was edited by: Neutron on 2004-03-06 01:20 ]</font>

scary808 · Post by **scary808** » Sat Mar 06, 2004 3:10 pm

Dynamic range of your signal increases as bit depth increases. I feel it's WAY more important than samplerate.

When my projects start getting too full, often times I'll need to free some DSP up. I will take samples of simple hits & drums(if they're synths) with the STS4K at 32 bit. There is virtually no audible loss in quality. If I record into Logic at 24 bit I can hear a slight degradation. You can hear more depth to the sound especially reverbs. I know it will eventually have to be reduced to 16 bit but the outcome seems much better when beginning at higher bit depths. Try this experiment:

Record a 16 bit file & a 24 bit file of the same bassline with a reverb on it in your sequencer. Then take both of those files & load them into a sampler. Record the output of the sampler again at 16 bit. Some may not hear the difference but it's there.

Every time you record digital audio even if it's in the digital domain it degrades the quality. Not much but it does. Most of that degradation occurs in the bottom levels of the dynamic range. Since the dynamic range moves straight down from zero that means the degradation is at a much lower level when using higher bit depths.