Page 1 of 2

Bits and sound quality

Posted: Sun Dec 09, 2007 10:20
by YGoh
Assuming that the sample rate is high enough for both scenarios, may be >192 kHz, so is audio quality, such as transparency, better reverb, smoother, and etc.... depend on, the number of bits representing the audio resolution, such as 32 or 64 bits floating points, or does it depend on the number of bits the audio software is written at.

For example: Is a 64bits audio software that run on a 64bits OS will sound a bit better than a 32bits audio software run on a 32bits OS, assuming both softwares have the same algorithm and has 32bits floating points audio processing engine?

Posted: Sun Dec 09, 2007 13:39
by Graemme
Impossible (or at least a waste of time) to discuss without experience.

Must be experienced and then maybe we can talk about it.

More dependent on the skill of the software writer than pure bit resolution alone.

I don't mean to be rude, just pragmatic.

Graemme

Posted: Sun Dec 09, 2007 14:00
by Bernhard Guettler
Just to add to Graemme's comments, which are so true, you might also want to learn first about following issues:

the influence of the engineers haircut on the sound when sitting behind him

temperature in the control room and it's influence on the sound quality

what is the difference in the perceived sound quality from the producer's perspective, if the score is very big rather than small. :wink:

Sorry, we don't want to be mean, we just are clueless ourself...

Re: Bits and sound quality

Posted: Sun Dec 09, 2007 16:59
by Bernhard Guettler
YGoh wrote: so is audio quality, such as transparency, better reverb, smoother, and etc....

and on a more serious note: The day you give us a mathematical description for "audio quality, transparancy, better reverb and smoothness" we will give you an answer.

bit depth/accuracy beyond the human abilitiy for discrimination (~20-24 bit) in some cases is useless, in some cases is nice to have and in some cases is essential. Follow Graemme's advice and you will eventually find out.

-B

Posted: Sun Dec 09, 2007 19:39
by YGoh
Sorry, may be I'm not understanding the messages here, but what's Graemme's advice?

Yes, I agree that the algorithm matter, skill of the recording engineers, and etc...., but I was asking if all things are equal.

Posted: Sun Dec 09, 2007 21:08
by klaukholm
I think you are getting some concepts confused.

The bits of an OS have nothing to do with the bit depth of audio.
an 8 bit OS could theoretically process 32bit audio. It would simply take more cpu cycles to perform each calculation.

Posted: Sun Dec 09, 2007 22:53
by YGoh
I know.... but that's not reason why I'm asking. The reason is that with 64bits software that run on 64bits OS, you're able to take advantage of the larger RAM and perform more task in parallel, so therefore it might able to execute intensive algorithm, such as Reverb more efficiently, therefore it might sound better.

Posted: Sun Dec 09, 2007 23:16
by Bernhard Guettler
You go to a sushi restaurant. They serve 32 pieces. Next day you go to another sushi restaurant. They serve 64 pieces. Which sushi tastes better?
:D

Posted: Mon Dec 10, 2007 07:17
by Graemme
YGoh,

It's *all* in the implementation, not so much the theory...

You seem stuck at a point that will never yield a real-world answer and I only really care about what I think/feel about how something sounds (and whether or not it sounds like what I want it to sound like!)

Things are rarely equal, especially once they are implemented in real-world designs.

Graemme


YGoh wrote:I know.... but that's not reason why I'm asking. The reason is that with 64bits software that run on 64bits OS, you're able to take advantage of the larger RAM and perform more task in parallel, so therefore it might able to execute intensive algorithm, such as Reverb more efficiently, therefore it might sound better.

Posted: Wed Dec 12, 2007 15:21
by fgoupil
Hi Ygoh,

You seem to mix two different things here : wordlength and software coding.
The fact that a software is coded to run on a 64 bit processor has nothing to do with the fact that it can do its internal audio processing at 48 bit, 64 bit or any other precision. I'm not an expert in processors but I believe that the main difference between a 32 and a 64 bit processor is the amount of memory they can access.

Audio software running on a 32 bit processor can do its calculation in precisions that exceed 32 bit. Therefore, running on a 32 or 64 bit processor has no impact on audio quality.

Best,

Fran├žois

mix

Posted: Wed Dec 12, 2007 17:49
by charlienyc
fgoupil wrote:You seem to mix two different things here : wordlength and software coding. The fact that a software is coded to run on a 64 bit processor has nothing to do with the fact that it can do its internal audio processing at 48 bit, 64 bit or any other precision. I'm not an expert in processors but I believe that the main difference between a 32 and a 64 bit processor is the amount of memory they can access.


open up this month's Mix to page 5 and find a huge Sonar ad that says SONAR: 64, ProTools: 48. Just Sound Better. the first point under the heading is SONAR offers the best audio quality in the industry: 64-bit end to end

while the ad goes on to show more detailed differences between the DAWs, none of those points proves anything about audio quality :roll: the worse thing is the top of the ad is misleading. so it's not surprising this confusion is happening and will probably soon become widespread. like so many misconceptions in the past, we have marketing people to thank!

don't believe everything you read :)

Re: mix

Posted: Wed Dec 12, 2007 18:24
by Bernhard Guettler
charlienyc wrote:open up this month's Mix to page 5 and find a huge Sonar ad that says SONAR: 64, ProTools: 48. Just Sound Better. the first point under the heading is SONAR offers the best audio quality in the industry: 64-bit end to end
...
don't believe everything you read :)


64-bit end to end... RIGHT! from 64bit A/D converter at recording all the way to linear 64bit D/A in the iPod :P

more mix

Posted: Wed Dec 12, 2007 18:57
by charlienyc
oh yeah, a few pages after that ad you'll see a typo in regards to CD quality audio at 44.1k/160 bits! wouldn't that change everything?

Posted: Wed Dec 12, 2007 21:06
by dperry
With all these bits flying about, why is the best sound quality still one bit?

Posted: Sun Dec 16, 2007 07:01
by nanohv
dperry wrote:With all these bits flying about, why is the best sound quality still one bit?
Hi. Yes, I couldn't agree more. By converting to DXD (352.8kHz PCM), do you think it will loose some of those high resolution of DSD, especially at 128fs