Hi - Asking for the purpose of making more compact back-ups... I'm keeping the source of course! For normal video editing I've been keeping to mod-8. But recently, for DVD archiving I moved to mod-16 (it seemed a "safer" choice"...). So I crop pointless black / cover noisy edges leaving up to 6/8 pixel borders at edges.
However... I recently discovered x264 blurs edges if there are black bars + picture contained within a 16x16 macroblock... the black borders I retained or added.
(...apparently x264 assigns a higher quantizer to the whole macroblock due to the static black, so the valid picture content gets encoded with a lower quality and becomes slightly blurred within those 16 edge pixels. This is more noticeable when the picture at the edges has a uniform color/texture etc).
So... asking some pros... what are the actual benefits of cropping black/noise to anything above mod-2 (in 2025)?
I'm aware encoding just pads "invisible" pixels to make any video back up to mod-16 - at some expense to efficiency - but is it worth just going a mod-2 route while cropping, if genuine quality is lost by cropping noise and adding borders to keep mod-8/16 instead?
(How does this padding work with the macroblocks? Any downsides at all?)
* What I'm cropping varies between huge black 4:3 letterboxing, VHS capture noise, to just MPEG-2 DCT edge blurring on high quality pro DVDs. But it seems doing so adds its own edge issues.
Thanks in advance!