AnimeMusicVideos.org

The separatefields/selecteven stuff is NOT a fix, it will break sync. It is a way to determine frame structure so you can see what the blending problem really is. I wanted you to run only that, without the tdeint or srestore to ruin it even more. Doing it after deinterlacing is even more pointless a...

Re noticing and/or telling the difference, facts: to notice the difference at all, the desync has to be 100ms or longer. This is due to the human eye being incapable of noticing motion changes faster than 1/10th of a second. At regular framerates (24fps), this equates to 2.5 frames. As for the diffe...

Going by that, it looks like your DVD was created by decimating a telecined NTSC release. Definitely no soft telecine, and given it's flagged as interlaced coding that sounds about right. See what happens when you do say, separatefields().selecteven(), does it still have blending?

VDM is the problem, use regular VirtualDub, it's actually up to date and I don't think you could provide any real benefit that VDM has over it other than being super outdated and slow, if you're into that. Also, if FFMS2 doesn't work then even moreso the problem is the player and not the decoder.

ffvideosource("your-filthy-stolen-clips.mp4")

Whoa, for some reason I thought Europe was R3, for like 10 years. Mind kinda blown what...

There are a few ways around it but probably, yeah. R2 isn't PAL so no idea how you got something that borked unless it wasn't PAL to begin with. What is the actual frame rate/composition?

DSS uses your system codecs so you probably /did/ break something there. It's also super unreliable so better not to use it, same for vdubmod.

The easiest way really is to find a pan and just try different pulldown() combos for a succession of 5 frames. That's what I do for 1PN clips.

Sounds like an Australian DVD, also obligatory lol semen. Something I once tried is to re-interlace each field separately after bobbing and then deint them again but it's kinda messy as you need to manually 4 frames now. Something you could try just to confirm field blending (and not PAL retardation...