[FFmpeg-soc] avfilter: what data ought to be available for configuration?

[vmrsss]

On 17 May 2008, at 22:17, Michael Niedermayer wrote:

>>> The problem is that the way the pixel_aspect is handled in lavf,  
>>> it is
>>> a field of each one of the pictures (I wonder what is the use of  
>>> that,
>>> if you can't change dimensions on the fly, why would you need to
>>> change the pixel aspect ratio?).
>
> Well, maybe we one day will fully support changing dimensions on the  
> fly.
> Adding a sample_aspect_ratio to AVCodecContext would surely be  
> possible
> if this solves the problem.


Unless I am mistaken, there is already a sample_aspect_ratio in  
AVCodecContext. What am I missing?

My question is whether we need to add a sample_aspect_ratio field to  
AVFilterLink and make it available at filter configuration-time. More  
generally: what codec parameters are likely to be needed in order to  
**configure** (not run) a filter chain? Width, height, PIX_FMT, SAR  
for sure (any others? fps? interlaced flags?).

All these are probably best packed inside AVFilterLink and passed  
through the chain (and possibly returned to ffmpeg) the same way  
width, height and PIX_FMT are currently. (There will be an amount of  
duplication of information, but isn't that the same with the frame's  
width and height?)

-vmrsss