• 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: 64bit processing - Why?
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 64bit processing - Why?


  • Subject: Re: 64bit processing - Why?
  • From: Richard Dobson <email@hidden>
  • Date: Fri, 31 Jul 2009 11:32:00 +0100

Brian Willoughby wrote:
..

I basically disagree, although perhaps you can find a flaw in my description of why.


Audio samples are expected to be in the range where all values are less than 1.0, with few exceptions. This fact alone generally avoids exactly what you described. Digital audio is not exactly the broadest use of float.


I am with you for the case of final track mixing (though if channel counts rise into the 1000s we may start to see problems!); but there are other summing situations where width helps - phase accumulation for additive synthesis (e.g. in pvoc etc) being the one I have spent most time with. Running a real-time soak test with a float accumulator loses accuracy remarkably quickly (e.g. in generating a sawtooth additively, the shape degrades within 30 mins using floats as phase accuracy reduces). Plain pvoc and FFT generally also involves summing of many values, Pvoc is pretty good with floats, but doubles is definitely cleaner-sounding and the extra precision helps in all sorts of other ways too. I suspect many physical modelling applications (e.g. where 1000s of nodes may be summed) also need more precision than the float can provide. So in the specific case of final audio mixing (with "normal" channel counts) clearly there is plenty of margin, but this does not mean that in the general case summing is without problems.


Nevertheless, it still would not surprise me very much if summing over a 48bit fixed-point bus did not sound just a tad cleaner than summing over a 32bit float bus. I have to take such things on trust, since I can't afford the gear necessary to prove it one way or the other!

Richard Dobson


_______________________________________________ Do not post admin requests to the list. They will be ignored. Coreaudio-api mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: This email sent to email@hidden
  • Follow-Ups:
    • Re: 64bit processing - Why?
      • From: Roman Thilenius <email@hidden>
References: 
 >Re: 64bit processing possible? (From: Evan Olcott <email@hidden>)
 >Re: 64bit processing possible? (From: William Stewart <email@hidden>)
 >Re: 64bit processing possible? (From: Brian Willoughby <email@hidden>)
 >Re: 64bit processing possible? (From: David Duncan <email@hidden>)
 >Re: 64bit processing possible? (From: Brian Willoughby <email@hidden>)
 >64bit processing - Why? (From: Ethan Funk <email@hidden>)
 >Re: 64bit processing - Why? (From: Brian Willoughby <email@hidden>)
 >Re: 64bit processing - Why? (From: Ethan Funk <email@hidden>)
 >Re: 64bit processing - Why? (From: Brian Willoughby <email@hidden>)
 >Re: 64bit processing - Why? (From: Richard Dobson <email@hidden>)
 >Re: 64bit processing - Why? (From: Brian Willoughby <email@hidden>)

  • Prev by Date: Re: 64bit processing - Why?
  • Next by Date: Re: 64bit processing - Why?
  • Previous by thread: Re: 64bit processing - Why?
  • Next by thread: Re: 64bit processing - Why?
  • Index(es):
    • Date
    • Thread