[Theora] unified meta-data handling ogg/theora/vorbis

Ben Bogart bbogart at ryerson.ca
Mon Dec 5 10:50:00 PST 2005

Hey all,

I'm working on a media-database art project that basically stores a bunch of sound and video files, which are then navigated based on meta-data. I'm not sure where to start asking these, questions, here seemed like a good start. 

So I was thinking I should use an open format, like ogg/vorbis/theora. I also thought, will all that coll meta-data stuff in ogg I should be able to store the meta-data in the files and add the meta-data while recording. (or soon thereafter)

Of course now I know that ogg has no meta-data, vorbis does. I can read and write meta-data using vorbiscomment, great! but that is only for sound-files. :( Seems theora has meta-data also, but its not compatible with vorbis, and it seems there is no theoracomment to manipulate theora meta-data.

My grand dream was to be able to stream media from a client machine (running ices2 or oggfwd or oggcast [from PD] etc..) Have the ices2 client, or icecast server append the meta-data to it (location of recording, tone of voice, topic of recording etc.). Then using something like vorbiscomment to meander through the data-base based on meta-data content. The meta-data would probably have to be loaded into ram first for fast-searching. 

It would be nice if server did not care if it was theora or vorbis, and just put appropriate meta-data onto it. Also it would be great is the meandering was pan-media so that it could choose a sound-clip here and a video there etc.. 

So asked on the IRC channel and got to where I am now. I also looked a bit at the oggskeleton wiki. 

Since theora comments seem to be not as developed as vorbis comments, since there is no theoracomment (I know of), and since ogg does not provide unified meta-data itself, any ideas on approaching this problem?

Anything in SVN/CVS I could tinker with to see what the bleeding edge is? Anyone working on theoracomment or even some other way of tagging a ogm file that is not dependant on the audio/video codec?

Any ideas would be appreciated. 

Thanks for your time.

B. Bogart
Evolving Stories Lab
Ryerson University

More information about the Theora mailing list