• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: "Native" sample format
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: "Native" sample format


  • Subject: Re: "Native" sample format
  • From: Jens Bauer <email@hidden>
  • Date: Thu, 11 Dec 2003 00:35:00 +0100

Hi James and Jeff,

Thanks for the information on the native sample format.
I've been thinking about it, and yes, it'd probably be best, if I use a float, rather than integers, imagining that I mix a few 0.33333 values into one sample value. When doing that as integer, it'd probably be much more precise using floats. =)

On Wednesday, Dec 10, 2003, at 21:12 Europe/Copenhagen, Jeff Moore wrote:

The canonical sample format is 32 bit float. We chose that because it strikes the right balance between memory bulk (only take a third more space than the original 24 bit sample) and computational speed (floating point headroom management and Altivec acceleration speed things along quite nicley).

On Wednesday, Dec 10, 2003, at 20:41 Europe/Copenhagen, James Coker wrote:

The native format is 32-bit floating point, but you can always
move to double-precision when necessary (filter coefficients).
Of course, you'll take a big performance hit for that, unless
you're running on a G5....

On Wednesday, December 10, 2003, at 06:37 AM, Jens Bauer wrote:

What's the native format of a sample ?
-Is it a 32bit signed integer ?

I'm asking this question, because I'd like to write my code, so CoreAudio doesn't have to convert my data, before feeding them to the soundcard.

I've been thinking about using floats, but since that floats are only 32bit, they're getting unprecise (but ofcourse, as long as I need 24 bits only, I don't know whether the data are actually ruined).
I'd prefer doubles, so I'm sure that I won't get that problem.
Unfortunately, there's no such format, so is the most precise format 32-bit ?

I know that most people can't hear these differences, but if I'm going to sample some data for other purposes than sound, it could be quite neat to have them as exact as possible, however, I'm also going to write sound applications. ;)



Love,
Jens
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.

References: 
 >Re: "Native" sample format (From: James Coker <email@hidden>)

  • Prev by Date: Re: Documentation confusing
  • Next by Date: Re: Documentation confusing
  • Previous by thread: Re: "Native" sample format
  • Next by thread: Re: "Native" sample format
  • Index(es):
    • Date
    • Thread