[vorbis] Re: [vorbis-dev] rehuff [source attached]

[Christian.Buchner]

> > This thread taken from vorbis-deff  [sic!]
> 
> why not leave it on vorbis-dev.  thank you.

too late ;)
 
> > So if rehuff can do that kind of magic, why can't Ogg Vorbis 1.0
> > generate a better coded stream?
> 
> Because you need to know all data in advance, to calculate optimal
> codes.  libvorbis can stream, so that's a no-no.

Wouldn't it make sense to integrate this functionality into libvorbis(file)?
When setting a particular flag in the API, the lib would then re-compress
the output file it just generated during an encoding run. As you indicated
this would only work when the encoded data was saved to a file -- not
for other data sinks like an Internet audio stream.
 
> > Why is there a need for such a tool?
> 
> Not much need on normal streams; it only saves a few percent.
> 
> But on otherwise recoded streams (peeled streams, stereo-to-mono coded streams,
> etc.) it saves a bunch.  Also, you get much smaller headers this way
> (which is
> very helpful for fast-start network streams).

Oh, by the way. Would vorbis support changing the codebooks on the fly
within a single audio stream? Yeah, with chained streams this is already
possible -- but what if I encoded a full CD in one stream. Could I still
change codebooks between individual tracks? Like before a transition from
classical music to death metal? ;)

<p><p><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-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.