[theora] Theora/Thusnelda features questions.

[xiphmont at xiph.org]

I'll also answer, which is not meant to imply that Gregory's answers
were somehow lacking :-)

> First of all I don't know if this is the right place to ask but I haven't
> found any Xiph/Theora forum or something similar...

Within the company we don't use a forum because it's useful for
chatting, but not really for getting work done.  It might be useful
for user support, but we don't really have the resources to offer such
a thing on a continuous basis.  We do help folks who pop into IRC or
ask questions on the mailing list.

For general codec discussions, there are already other good places to
go to, eg, HydrogenAudio for audio and doom9 for video.

> I don't even know if this is the right way of "posting" in mailists... But
> anyway...

Yep.  You send mail.  Pretty simple :-)

> Of course, I did a lot of testing with the 1st Thusnelda alpha of
> ffmpeg2theora and results are very good! I'm very pleased with results.
> Compression, quality, ease of use and stability are really great even when
> there's a good amount of new code since Theora 1.0 version (for what i've
> read).
> Congratulations to the devs!

I'm glad you've had good results.  In general the tools have a long way to go.

> *- Is it possible to support a "Two-Pass" or even "multipass" feature in
> Theora?

Yes.

> *- Is there any plan to support that or a similar feature in Theora?

Originally I hadn't considered two-pass to be all that compelling.
The biggest reason you're getting better results using two pass in
Xvid is because, right now, xvid is still a better codec (not
surprising, Thusnelda is not finished.  There's still a major piece of
infrastructure missing in the encoder).

That said, two pass makes some interesting optimizations useful and we
may well implement a two-pass mode in our own encoders.

> - knowing that your "goal" with Theora is/was (?) to achieve similar quality
> than MPEG-4 Part-2/ASP (?) -

We will keep up for a period.  On paper, MPEG part 2 and part 10 are
considerably more advanced codecs.  However, a codec's quality
performance is only as good as the encoder, and we've got alot of room
for improvement yet (as does MPEG of course).  Eventually, MPEG4 will
surpass Theora for good, but we'll probably still have comparative
lightweightness working for us.

> *- Do you plan to implement more advanced features comparable with MPEG-4
> Part-10/AVC? (h264)

If not in Theora than in a successor.  The only disagreement within
the org is which of those two paths to take.  I'd like to modestly
extend Theora, and Time would like to start fresh.

> And more than that...
>
> *- Can you implement such comparable features in current Theora/Thusnelda or
> that would require a whole new direction?

Well it depends what you mean.  There are features within theora we
don't yet use or use efficiently.  They wouldn't really be new.  Eg,
two pass would not require anything new, just a different way of using
what we already have.

If you mean 'new features' in terms of improvments to the format
itself, then you're talking about changes that are extensions by
definition.

Monty