[vorbis-dev] lossless ogg encoding with good compression rate

[Carsten Haese]

On Thu, 2003-06-26 at 08:55, Beni Cherniavsky wrote:
> Carsten Haese wrote on 2003-06-26:
> > [snip]
> > Different decoders (e.g. on different platforms) are not
> > guaranteed to produce the same wav from the same Ogg Vorbis file.
> > [snip]
> This statement worries me.  The spec specifies the matematically
> "right" decoding when you have infinite precision FP, doesn't it?  So
> is this just a matter of FP errors in actual implementations?  It's an
> acceptable tradeoff to pay with small noise for effecient decoding but
> it is still desirable to have the "one true" decoding defined and
> possible.  If you decode with libvorbis to 24bit and then round,
> wouldn't you get it right?  Perhaps if you replace all floats with
> doubles in libvorbis and recompile?

Yes, the specs specify the correct way to decode, but I have yet to see
a computer that performs infinite precision math. The difference comes
down to different rounding strategies between different platforms. Since
Vorbis is lossy, this is not a problem in practice, because those errors
are drowned out by the quantization error that the encoder produces. It
only becomes a problem when you try to make a lossless process from an
inherently lossy process.

Carsten.

<p>--- >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.