[theora-dev] RFC: I would like to contribute with SoC

Lino Mastrodomenico l.mastrodomenico at gmail.com
Sun Mar 25 09:10:46 PDT 2007


[Resending since I sent my first mail before subscribing and I think
it didn't went through. I apologize if you got this twice.]

Hello everyone,

I would like to participate as a student in google Summer of Code and
I was very excited to discover that Xiph is one of the mentoring
organizations, because I am a big fan of both Vorbis and Theora.

So I would love to contribute to a Xiph.org project and I'm writing an
application (the deadline is Monday). I read all the ideas
from:

   http://wiki.xiph.org/index.php/Summer_of_Code

And there are several ideas that caught my imagination:
   * "Optimize Theora encoding/decoding speed", because I know the
MMX and SSE instruction sets, so I feel I can contribute here;
   * "Subtitle Definition", because I think this can be useful to a
lot of users (I currently use external .srt files with my Ogg Theora
videos);
   * "Intel to AT&T x86 assembly translation": I know both assembly dialects.

And I have an additional proposal: adding two or three pass support
for Vorbis and Theora encoding. The rationale is that I did a lot of
testing and I found that, after one or two tests to determine the
"compressibility" of the source, it's possible to predict with pretty
good accuracy the final size of a video or sound stream for a given
quality. So with a two pass encoding it should be possible to obtain
the best of both worlds: the full quality from a VBR encoding and at
the same time satisfy the people that want, e.g., a 700 MB rip of a
DVD. Comments?

But in the end I find almost all the ideas listed in your page very
interesting and challenging (in a good way ;-)). I think I am a good
programmer and I know C, x86 assembly (including x86_64), Python, C++
and Java.

So my question is: what are the ideas where my contribution (if
accepted) can be more useful? And, vice versa, what are the ideas that
I should give less priority because, e.g., you already have a lot of
students willing to work on them?

Since (again, if accepted) I will be getting paid, I'm not afraid of
doing hard work and searching and studying documentation for things
that I need to know. And, if everything goes well, I would love to
keep contributing even after SoC.

Thanks in advance for any suggestion.

-- 
Lino Mastrodomenico
E-mail: l.mastrodomenico at gmail.com


More information about the theora-dev mailing list