[gst-devel] vorbis / theora caps

[Benjamin Otte]

While we had this recent headers in vorbis/theora discussion, I had an idea:
What about specifying 2 different formats for vorbis data? one, named "audio/x-
vorbis" for the stuff as it comes out of oggdemux, and another (for lack of a 
better name) "audio-x-vorbis-
stream,header_data=...,meta_data=...,setup_data=..." where the first 3 buffers 
are already put into the caps?
It should be dead easy to write a converter between these 2 formats (just put 
the first 3 buffers in caps / get them out of caps) and it would buy us the 
ability to easily get vorbis audio out of Matroska, AVI and frends?

Any opinions?
Ronald?

Benjamin