| Author: | Carsten Haese |
|---|---|
| Contact: | carsten@xiph.org |
| Date: | July 7, 2003 |
The monthly meeting for July 2003 took place on July 5, 2003, at midnight UTC in the IRC channel #xiphmeet on irc.xiph.org. The following people were present, in alphabetical order:
The meeting was moderated by Jack Moffit.
Since Dan had a scheduling conflict, the Theora part was pulled to the top of the agenda. Therefore, Dan started off the meeting with an update on his work to write the Theora bitstream specification. The beginnings are checked into CVS and are also available via FTP from ftp.vp3.com. With the help of Mauricio, Ralph has started a Wiki version which is available at http://wiki.xiph.org/TheoraSpec.
Dan's expected date for completing the spec is mid September, unless there are people willing to help him. Mike Melanson, who is very active on the theora mailing lists, was suggested as an obvious candidate.
One open item that needs to be resolved is the fact that VP3 is encoded upside-down, whereas Theora is not. This needs to be addressed so that existing VP3 streams can be converted into Theora streams in the future. This may be done by including an orientation flag into the Theora stream.
Regarding the question of further bitsream changes, it was decided that interlacing would be implemented in a new revision of the stream format.
Jack took us back to the beginning of the agenda with an update about the Xiph.org Organizational status. He was present at the Xiph.org board meeting on June 16th, 2003, together with 3 other Xiph.org board members, their lawyer, and an advisor. Minutes of the board meeting are available.
Jack then gave us a brief contract status update. Monty's Neuros firmware implementation is moving along, and the first beta for Ogg Vorbis playback was released recently. There are a few outstanding known problems, but those will be taken care of in the next few weeks.
Jack reported that he, Monty, and Ralph have agreed to migrate the Xiph CVS repository to Subversion. Initial skepticism was quickly overcome, and there was a broad consensus that this move will be a good thing. The move will happen once the cvs2svn export-import script properly supports branches and tags.
Since Subversion depends on Apache 2, it was decided to migrate the web server to Apache 2, instead of maintaining two server instance.
Next up, Nathan presented the proposed design for the unified Xiph.org websites. Zope will be used for the content management system; Plone doesn't seem ready for prime time yet.
Brendan has been pushing to get a release of libshout2 out of the way. As he put it, it is currently in perpetual "just one more beta" stage. He is also working on perl and python bindings which are very close being done. The sticking point is documentation, and Brendan would appreciate if he could get some help with that.
Since Monty was absent, Jack relayed Monty's list of near-term goals that he will work on once the Neuros contract is completed. One is the oggfile API which will provide a vorbisfile-like API for multiplexing and demultiplexing Ogg files. Next up is a Vorbis 1.0.1 bugfix release, backporting some Tremor changes to the reference implementation, including the libogg2-zerocopy memory management, splitting decoding into its own library.
Finally, Monty is planning to make a Vorbis 1.1 AAC+SBR killer release. The exact plans are not known, but Vorbis 1.1 is going to include encoder tweaks to be more competitive in the low bitrate arena. There is also a possibility that some of Garf's high bitrate tuning will be incorporated.
In the last part of the meeting, the branding issues surrounding Ogg were discussed. It was decided that OggSquish will be dropped, as it is no longer in use. Also, Ogg as the name for the multimedia umbrella project will be dropped in favor of the Xiph project. This leaves Ogg unambiguously as the container format that the Xiph project employs.
The discussion then moved towards acceptable use policies for Xiph logos. On one hand, there is the desire not to allow inferior products to create a bad impression of Xiph technology. On the other hand, nobody has really the time to define and enforce strict quality guidelines. As a compromise, it was decided that the logos shall be usable with only minimal restrictions that are to be determined, and that the logos shall not imply quality. A few possibilities for logos that don't imply quality were pointed out by foolip and Jack.
Before the meeting adjourned, Jack squeezed in a small update from Phil Kerr. The draft specifications for the Speex RTP payload and Vorbis-over-RTP are scheduled to be discussed at the next IETF meeting in Vienna. The Vorbis bitstream specification didn't make it, but Phil is confident that it will be ready for the following IETF meeting.
Meeting Minutes for the July 2003 Monthly Meeting