[Ffmpeg-devel] [PATCH] Basic H.264 encoder

Panagiotis Issaris takis.issaris
Fri Apr 21 15:33:01 CEST 2006


Hi,

Op woensdag 12 april 2006 22:31, schreef Michael Niedermayer:
> [..]
> then again i wont accept or reject patches due to political reasons so if
> your encoder is clean and otherwise fit for cvs ill apply it
We're obviously happy to hear this. Hopefully, we'll get it fit for CVS soon.


> On Wed, Apr 12, 2006 at 04:01:49PM +0200, Panagiotis Issaris wrote:
> [...]
> > ?- motion vectors are set to zero
>
> that is a reason to reject it unless you convice me that it will be fixed
> soon
We are currently trying to get the motion vectors working. Unfortunately,
we are having some problems with figuring out how to use the motion 
compensation API (or even if there exists any). It seems only the Snow codec
is using the motion compensation code, without being integrated in the
mpegvideo.c file. All the other codecs seem to be intertwined with mpegvideo.c
code.

So, here are some questions we are struggling with:
- Is it better to modify mpegvideo.c for using motion compensation for our 
H.264 codec, or can we do it without touching that file? (To us, it seems the
latter is cleaner.)
- Will integrating in mpegvideo.c make it easier to get rate control working?
- In the sourcecode there's a remark that insinuates that the motion control 
code will be separated from the MpegEncContext struct. Is that planned for 
the near future?

> > ?- the image width and height are reduced to a multiple of 16
>
> why?
Actually, just because it made our initial implementation effort easier. In 
our updated patch, this is fixed.

> > ?- for debugging purposes, the NAL units are written to a file
> > ? ?/tmp/teststream.264 and reconstructed frames are appended to
> > ? ?the file /tmp/teststream.yuv
>
> obviously not ok for lavc, but you didnt submit it for inclusion into lavc
> yet so everythings fine ...
In the process of cleaning the updated patch, the above debugging code has 
been removed.


With friendly regards,
Jori and Takis





More information about the ffmpeg-devel mailing list