Quote:
Originally Posted by henryho_hk
Would it be easier to for XviD's encode if we crop in a multiple of 8 (or even 16) to align with the MPEG2 macroblocks / block artifacts?
|
Coding with Xvid is most efficient when the picture is cropped to a multiple of 16 horizontally and vertically.
Worst case is to leave black borders which fall somewhere in between a 16x16 macroblock, because the sharp transition between the useful picture and the black border consumes extra bits for encoding.