[FFmpeg-user] display size vs coded size

[Roger Pack]

> No, I'm looking for this functionality on the encoder side. I was hoping this would be available from the ffmpeg command line but if it is available via the libavcodec API, that'd be good too.
>
> What I want is to have video that is coded with a certain resolution, but to be able to additionally indicate in the container format that there is a different resolution that contains the interesting data. I've worked in closed multi-media systems before where this sort of thing to was used to indicate that even though the video needed to be encoded on a 16 pixel boundary, the actual valid input video might have had some different resolution (e.g. if the input video is 600x480, you might code the video at 608x480 but tell the player to only display 600x480). This obviously also requires you to be able to specify which pixels are the interesting ones.

I'm not sure which muxers/containers allow you specify that type of metadata...