theora dev - Feb 2010 - one feature to future theora which might have good impact

Hi All,

as https://trac.xiph.org/browser/experimental/derf/theora-ptalarbvorm goes
I''d like to share
one idea  on features to improve market penetration and  the acceptance of
theora codec.

there is a market - online web games.
but still most of the games are quite primitive as their expressive means
are not well developed.
rich interactive 3D rendering is not easy to make with flash and friends.

And for most of those games low latency video will be a real solution to add
extra life.

and there are approaches to achieve such modes

http://x264dev.multimedia.cx/?p=249

I did not test current theora on ability to provide low latency streaming.
Still few ideas like "Periodic intra refresh" ( from link above  )
might
improve the situation.

Now then, when it is in, it is easy to imagine someone sitting and pressing
keyboard with 3D game screen changing ( which is rendered on server ).

The penetration for theora ( as free codec compared to H.264 ( which is free
for free sites till some date in future but not otherwise )) might  really
increase.

Regards
Sergey
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/theora-dev/attachments/20100212/39584fc5/attachment.htm

Sergey Kurdakov <sergey.forum at gmail.com> wrote:
> I did not test current theora on ability to provide low latency streaming.
> Still few ideas like "Periodic intra refresh" ( from link above 
) might
> improve the situation.
Periodic intra refresh is planned for 1.2. I''m not sure what
we''re
going to do about communicating that to decoders yet, so that someone  
joining a stream in the middle doesn''t have to wait for a keyframe  
that will never come. There are IPR issues I need to work through  
here, as well.

After that the primary limit is going to be the accuracy of the rate  
control. We have not spent time to make accurate estimates of the  
number of bits that will be used by a given frame, and instead rely on  
somewhat coarse estimates that can be averaged over several frames and  
to which we can react quickly. Contributions in this area are welcome.

Hi Timothy,

great news on 1.2 features!
as for
accurate estimates - I will dig into code at some point,and if I can
contribute, I will, for sure.

Regards
Sergey

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/theora-dev/attachments/20100212/a37776cb/attachment.htm

Hi Timothy,

after accessing my skills to code  prediction of rate allocation ( I''m
game
developer and use theora as user-programmer ) I can admit, that I would
hardly add something useful.

Still I believe that it is possible, by the way of update like Monty''s
regular demo pages ( which I accustomed to read last year, that is why
I''m
here ) to call for arms for specific problems.

at least theora is already in firefox, and I believe that there are
corporate programmers which read updates and ready to help to push it
further but which are  still off dev list without invitation ...

Regards
Sergey
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/theora-dev/attachments/20100213/3b3dc6fe/attachment.htm

On 02/13/2010 01:11 PM, Sergey Kurdakov wrote:
> Hi Timothy,
>
> after accessing my skills to code  prediction of rate allocation (
I''m
> game developer and use theora as user-programmer ) I can admit, that I 
> would hardly add something useful.
>
> Still I believe that it is possible, by the way of update like
Monty''s
> regular demo pages ( which I accustomed to read last year, that is why 
> I''m here ) to call for arms for specific problems.
>
> at least theora is already in firefox, and I believe that there are 
> corporate programmers which read updates and ready to help to push it 
> further but which are  still off dev list without invitation ...
>
> Regards
> Sergey
I would like to add my voice to this e-mail, because, although there''s 
little I can personally do to benefit libtheora''s code aside from 
testing, I was still drawn in to be a part of the Xiph community through 
Monty''s updates, which really served to increase my interest.  Yes, we 
have Planet Xiph, but I didn''t even know about that until some time 
after becoming a regular on #theora.  So, I just thought I''d share my 
own agreement with the above recommendation based on my own personal 
experience.

-- 
       Basil Mohamed Gohar
abu_hurayrah at hidayahonline.org
http://www.basilgohar.com/blog
basilgohar on irc.freenode.net
GPG Key Fingerprint:  5AF4B362