[ogg-dev] Regarding applications implementation of Skeleton's Content-Type

[Ivo Emanuel Gonçalves]

I noticed this first in vorbis-tools and later in ffmpeg2theora.  The
Skeleton bitstream contains a header called Content-Type.  It lists
the media types of all the data in Ogg.

In vorbis-tools and ffmpeg2theora the output is normally Vorbis and
Theora.  The Skeleton implementation there reports those media types
as video/x-theora and audio/x-vorbis.

I would like to suggest that it would be better to use video/theora
and audio/vorbis as those media types are in the middle of being
registered through their respective RTP RFCs.

Why urge this change instead of waiting?  Because of interoperability
as the Skeleton parsers will be expecting one media type for each of
those codecs and not two (the x- and non x- versions).  It would be
pretty bad if there were two kinds of Ogg+Skeleton files out there.
It's already bad that there are those with and without Skeleton as it
is.

The less confusion the better for everyone and, considering the media
types are used internally, no one's going to complain.

-Ivo