Now that a few folks of the appropriate technical background are listening (and
that a couple more I'm expecting to join in a bit can grab this message from
the archive), it seems prudent to start talking about where Vorbis is and what
it needs. I'll assume at this point everyone knows what Vorbis is; stop me
if
I'm wrong ;-)
More of the code is complete than the web pages might suggest. That said,
libvorbis still needs at least a little work everywhere and alot of work in
several places. I'll talk about libvorbis mostly; Mike Whitson has been
handling the example encoder/player and I'll let him pipe up about that in
his
own time.
Because documentation is more scarce than code right now, I'll have to start
by
summarizing how Vorbis codes and at each step point out the relevant piece of
code. The CVS repository is now avaliable for all to see; take a look at:
http://www.xiph.org/cvs.html
for access information. I'll have CVS snapshot in tarball form on the
download
page for Vorbis, but CVS access is much more convenient. Having up-to-date
source around will help in following along; with luck, it it a model of clarity
;-) I also realize that many folks listening now will require read/write
access. Let me know when you desire it, and I'll get you access.
Unfortunately, the things to discuss/explain first (how the transforms fit
together) is the least-complete code; if you look at these parts of the
repository now (analysis.c and synthesis.c) you can see they're in a state
of
me mostly fooling around. The transforms are all there, but they've not yet
been glued together. The coder is not written, but all the framing and
syncronization is complete. I'll get to it all in the next few emails.
Monty
--- >8 ----
List archives: http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/