[x264-devel] Summer of Code: RDO B-Frame decision algorithm optimizations and psychovisual optimizations for mode decision

[Robert Deaton]

Hi all,

I will be working on SoC this summer. The meat of my project's
proposal is pasted below, as requested. :). Enjoy!

The goal of the project is basically covered in the abstract. The
following is a basic timeline, designed to be flexible depending on
how long it takes to learn and adapt to this project.

The Community Bonding Period would be used to do the required
background reading and learning on video compression and specifically
x264. This would include reading the relevant parts of the H264
specification and familiarizing myself with the x264 codebase and
general video coding techniques. If this is completed early, then work
on the project may begin early, or as I am learning.

The first interim period up until midterm evaluations would consist of
working solely upon the RDO B-Frame decision algorithm. This is the
primary deliverable of the project, and my goal is to have working
code that shows some improvements by the midterm evaluations. Up until
this point, the focus will be on getting better results out of the
algorithm.

The second interim period after midterm evaluations and up until the
final evaluations would consist of tweaking the RDO B-Frame decision
algorithm to be as optimized as possible and to produce the best
possible results. Once this is sufficiently complete and if time
permits, I would like to use the remaining time to begin research and
testing for the psychovisual optimization part of the project. This
part of the project will only come in with time permitting and should
not in any way affect the delivery of the B-Frame improvements.

I look forward to working with you all this summer,

-- 
--Robert Deaton