Re: Re: hdot264 - Developer

Hi. Im a phd student working on video coding. After working 
on an MPEG4 implementation I recently started looking at the JVT's 
source code. As someone already said, it is a big mess, it is old 
and contributed by many people. There is not a common naming 
convention. There are pieces of years old TML code around. 
And finally when I saw two functions each more than 1000 lines long 
I decided to work on my own H264 implementation. Therefore 
I am really interested in the hdot264 project and I am 
seriously willing to contribute.

Regarding the encoding/decoding complexity of codec, note that 
there are a multitude of optimization issues from the encoder's 
perspective. Therefore IMHO 8x encoding complexity vs MPEG2 is 
an optimistic figure unless you are not shooting for the best 
RD optimization. I think it is best to implement a realtime/suboptimal 
coding (e.g. for live streaming) and an offline/optimal coding (e.g. 
for storage) and let the user decide which option he/she would 
like to use. Decoder complexity is somewhat limited, therefore 
the highest C++ optimizing is the best that can be done for the 
decoder.

As stated, ref. software can and should be used for standard 
compliance. But it is not that mature yet, not a week passes 
by without someone noticing a new bug in the code. 

Anyway, I would like to help get this project going. 

Oztan



-------------------------------------------------------
This SF.Net email sponsored by: Parasoft
Error proof Web apps, automate testing & more.
Download & eval WebKing and get a free book.
www.parasoft.com/bulletproofapps