[Foundation-l] Theora and Vorbis support in Firefox 3.1a2

Gregory Maxwell gmaxwell at gmail.com
Wed Jul 30 23:41:16 UTC 2008


On Wed, Jul 30, 2008 at 7:01 PM, David Gerard <dgerard at gmail.com> wrote:
> 2008/7/30 geni <geniice at gmail.com>:
>
>> <humor>Of course now Theora isn't elitist any more it's time to switch
>> to Dirac</humor>
>
>
> H.120 is THE. CLEARLY UNENCUMBERED. VIDEO FORMAT.
>
> (I wrote the Wikipedia article. I shudder at the prospect anyone
> actually trying to use H.120 for anything, ever. But! It's
> unencumbered!)

Heh. Indeed.  So is ANIMATED GIF and UNCOMPRESSED YUV FRAMES (sure,
it's 1.1GBytes per minute (640*480*30fps) but it's unencumbered!)! :)
... There is a reason why I said "reasonably performing [...] codecs"

Dirac is a cool format, and I expect in the long term we'll get accept
submissions of content in very high bitrate I-frame only DIRAC and
transcode it down for web playback to browser-compatible
broadband-compatible low bitrate Theora.  (Dirac currently underforms
theora at 'web bitrates' by a healthy margin, and it's only somewhat
recently achieved real time operation on fairly beefy machines... so
Dirac really plays in a different space than Theora right now... but
it is good stuff too)

Speaking of performance, the Theora encoder has been undergoing a
rewrite: It's not done yet, but the preliminary results appear to be
getting roughly half the bitrate for the same quality (or,
alternatively, twice the quality for the same bitrate).
[http://web.mit.edu/xiphmont/Public/theora/demo5.html]   I'll
obviously be posting to the commons list telling folks to upgrade
their encoders once the new stuff is out of development and officially
released.

> That is good news although don't the Xiph.org standards now call for .ogv?

.ogg is the catch all, .oga for audio, .ogv for video. Of course all
should work. At some point we should switch.  "It's just a name" in
any case.. the clients won't care.

> Figuring out what to do with videos encyclopedia wise is going to be
> another potential interest

Yea.. Well there are lots of cases for use of video in a 'source'
material capacity. We're already doing a lot of that. But certainly we
write about TONS of things that move where a short video could improve
understanding.

I think the bigger challenge is that the text and imagery created by
Wikimedians is already at such high standards that making videos that
don't look out of place will be hard.  Wikimedians routinely produce
still photography and diagrams which could hold their own in any
professional publication. Creating a video which is both informative
and polished looking is simply a lot harder than creating a good still
photograph.



More information about the foundation-l mailing list