• 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: Inconsistencies with large WAV file
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Inconsistencies with large WAV file


  • Subject: Re: Inconsistencies with large WAV file
  • From: Doug Wyatt <email@hidden>
  • Date: Fri, 04 Dec 2009 08:36:11 -0800

I tested AIFF and it doesn't have this problem.

Doug

On Dec 4, 2009, at 3:21, Richard Dobson <email@hidden> wrote:

Brian Willoughby wrote:
If this turns out to be a bug in the AudioFile source, then you should probably look for something similar with AIFF, in addition to WAV. When I first started working with AIFF (long before AudioFile), I made the assumption that the chuckSize should be UInt32, since there's no meaningful need for negative sizes. Apparently, however, the AIFF specification states that the chunkSize is only 31 bits, and is technically undefined for 32 significant bits.



AIFF/AIFF-C has always been a bit of a mess. In the spec, the chunk size is defined as a signed 32bit field; which does indeed mean that in principle an AIFF file as a whole cannot exceed 2MB. However, the specs (Apple draft, 1991) also propose that the best source of information on the size of the audio data is not the chunksize but the numSampleFrames field in the COMM chunk - and that is an unsigned 32bit long. BUT: the chunk size in the SSND chunk is of course a signed value - go figure. Way back then, the chances of anyone wanting to make a file that long must have seemed vanishingly low. I have to confess in CDP we simply treat all sizes as unsigned, Just In Case. Looks like AudioFile needs to do the same. Of course, ~now~, everyone should be using CAF instead anyway.


The great irony here is that in the (IMO) ill-conceived RF64 64bit- extended WAVE proposal adopted unquestioningly by the AES and EBU (and meant to be backwards-compatible with WAVE; which it isn't), they depend on setting a size of 0xFFFFFFFF in the main chunk size in order to indicate that the new larger 64bit "ds64" chunk should be used to obtain the size.

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:
@apple.com


This email sent to email@hidden
_______________________________________________
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: Inconsistencies with large WAV file
      • From: tahome izwah <email@hidden>
References: 
 >Re: Inconsistencies with large WAV file (From: Doug Wyatt <email@hidden>)
 >Re: Inconsistencies with large WAV file (From: Howard Moon <email@hidden>)
 >Re: Inconsistencies with large WAV file (From: Doug Wyatt <email@hidden>)
 >Re: Inconsistencies with large WAV file (From: tahome izwah <email@hidden>)
 >Re: Inconsistencies with large WAV file (From: Doug Wyatt <email@hidden>)
 >Re: Inconsistencies with large WAV file (From: Brian Willoughby <email@hidden>)
 >Re: Inconsistencies with large WAV file (From: Richard Dobson <email@hidden>)

  • Prev by Date: Re: Inconsistencies with large WAV file
  • Next by Date: Re: Inconsistencies with large WAV file
  • Previous by thread: Re: Inconsistencies with large WAV file
  • Next by thread: Re: Inconsistencies with large WAV file
  • Index(es):
    • Date
    • Thread