[vorbis-dev] Decompression error with two streams.

[Lars Brubaker]

I am no longer having this problem in beta 4.

Thanks to all.

Lars.

> -----Original Message-----
> From: owner-vorbis-dev at xiph.org [mailto:owner-vorbis-dev at xiph.org]On
> Behalf Of Lars Brubaker
> Sent: Thursday, February 22, 2001 10:04 PM
> To: vorbis-dev at xiph.org
> Subject: RE: [vorbis-dev] Decompression error with two streams.
>
>
> > It could simply be a bug the other usage never tripped (e.g., Vorbis is
> > putting out variable sized buffers which .raw and .wav would probably
> > not).
> >
> > That's not to say it isn't a Vorbis bug, just that 'it never broke
> > before' isn't really a strong argument for anything.
>
> Your right, it absolutely isn't a proof that the bug is in Vorbis.
> But this code has at least seemed solid for a long time.
> Do you have a test in which Vorbis decompresses two streams in the same
> thread repeatedly in which you can listen to both sounds?
>
> >
> > > 	return (pOggSoundReader->m_pFile->Read(ptr, size * nmemb) / size);
> >
> > I'm sure it's the case, but nothing else is touching that FILE * pointer
> > under any conditions, correct?
>
> You are correct, nothing else touches these sounds.  I have tried fairly
> exhaustively to make sure that there was no error condition in the file
> code or any errors reported from Vorbis.
>
> >
> > Anyway, the code in the COggSoundReader class looked good.
> >
> > I assume that the content you're mixing plays back fine when decoded
> > one stream at a time and that the failure is not repeatable (e.g., same
> > streams, but not same failure point each time)?  Have you set up any
> > means to capture/inspect the results directly from decode?
>
> Absolutely correct.  I have never heard an error with only one stream and
> the error occurs intermittently and in different spots within the sound.
> This is an argument against some strange alignment problem in the read:
> Vorbis is always giving 4k chunks until the end of the sound and our file
> library handles loading everything else in the game regardles of their
> size.
>
> >
> > I've not seen any Vorbis failure along the lines you describe (yet),
> > and I doubt we'll find the problem simply by looking at code without
> > significantly narrowing the search scope first.
>
> If you would like me to test something I'll help in anyway that I can.
> At this point we are going to be code releasing very soon and will likely
> just live with any bugs that we have.  But I would love to help find this
> for the sake of the library (if it even is a bug in the library :).
>
> Lars
>
> --- >8 ----
> List archives:  http://www.xiph.org/archives/
> Ogg project homepage: http://www.xiph.org/ogg/
> To unsubscribe from this list, send a message to
> 'vorbis-dev-request at xiph.org'
> containing only the word 'unsubscribe' in the body.  No subject is needed.
> Unsubscribe messages sent to the list will be ignored/filtered.
>

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.