[FFmpeg-user] tinterlace broken - SAR & DAR wrong

[Mark Filipak (ffmpeg)]

This just keeps getting weirder and weirder. setdar seems to be a workaround for tinterlace's SAR 
change for some circumstances but not for others. I'm going to explore further until I figure it out 
and can present some simple cases.

Opinion: tinterlace changing SAR is not a good idea. Anything changing SAR is not a good idea unless 
the 'anything' is something that differentially changes samples/line and lines/frame (i.e newPAR) 
_AND_ newDAR/newPAR != oldDAR/oldPAR. ...But no matter. I'll address that issue in the future.