[whatwg] The issue of interoperability of the <video> element

Maik Merten maikmerten at gmx.net
Tue Jun 26 10:24:46 PDT 2007

Jerason Banes schrieb:
> If that's true, then I'm greatly relieved. VP3 (the source of Theora) is
> generally compared to MPEG1, a standard far exceeded by H.263. I have
> not seen any publicly available Theora benchmarks that would overturn
> such impressions. (Do any exist?)

Most public benchmarks are usually outdated and using VP3. Albeit the
claim is often made that VP3 performed no better than MPEG1 this -
concluding from my tests - isn't correct (at least not for the version I
tested - actually there were different VP3 versions, the "latest" one
being VP3.2 IIRC).

Actually the original encoder xiph.org received from On2 did have a
bunch of bugs that impacted on the output. Theora's current encoder is
generally in a better shape than what was VP3.

I'll gladly send you small test clips showing H.263 and Theora at low
bitrates (which would be the major use for web video) if you wish (below
one megabyte, even suitable for attaching to an email).

Maik Merten

> On 6/26/07, *Maik Merten* <maikmerten at gmx.net
> <mailto:maikmerten at gmx.net>> wrote:
>     Jerason Banes schrieb:
>     H.263 is seriously outperformed by Theora. I don't know where all that
>     "Theora is high bitrate and low quality" talk comes from. It's not as
>     good as H.264, but it's for sure not worse than H.263 and from my tests
>     it's consistently better at low bitrate video.

More information about the whatwg mailing list