Bill Nye - Top fields are sharp but bottom fields are blurry. Why?

[Katie Boundary]

Obviously, the way to deal with it is with a purely spatial bobber like NNEDI3 in single-rate mode, but I'm interested in what could possibly cause such a bizarre phenomenon. Bad telecine hardware?

[wonkey_monkey]

You've been here long enough to know that no-one's going to be able to even attempt to answer that without a sample, or at the very least the provenance of your source. Being more specific than "Bill Nye" might also be useful.

[Katie Boundary]

Found a really spectacular example today basically by accident...

https://www.mediafire.com/file/p94bc3d6yc6vg1q/blur.demuxed.m2v/file

[hello_hello]

I'm guessing, but if you try to bob de-interlace (it's progressive isn't it?), the lines in one field might be in a better position to be interpolated into a full frame than the other, and therefore it looks sharper. Plus there's a lot of noise, which is probably the main culprit, and with the movement it sometimes results in one field looking more blurry, but not always the same one, so I'm not sure one field is inherently blurred, at least not in this case. If you bob de-interlace with something like Yadif, you'll definitely end up with some sharp and blurry frames, but it's not always the same interpolated field that looks blurry. QTGMC handles it pretty well.

Stacked fields

https://i.ibb.co/tTTHkznp/A.png

Yadif(mode=1).Subtitle("Yadif")
Frame()
A = last
B = A.Trim(1,0)
StackHorizontal(A, B)

https://i.ibb.co/7txk0qLB/B.png

QTGMC().Subtitle("QTGMC")
Frame()
A = last
B = A.Trim(1,0)
StackHorizontal(A, B)

https://i.ibb.co/FqhRrqqd/C.png

[Z2697]

My guess is the video is fake interlacing but low quality deinterlaced from an interlaced source

[Katie Boundary]

My guess is the video is fake interlacing but low quality deinterlaced from an interlaced source

Not likely. Other parts of the episodes are true 60 fps.

[huhn]

that does not dismiss his guess what so ever.

[Katie Boundary]

that does not dismiss his guess what so ever.

If bob-deinterlacing had been performed at any step on the journey to DVD, I'd expect it to be after the episodes were edited together but before they were put on DVD. As in, the DVDs wouldn't be broadcast accurate. I'm not even sure that bob-deinterlacing existed in 1993 or that it could be done with analog equipment.

[wonkey_monkey]

I'm not even sure that bob-deinterlacing existed in 1993 or that it could be done with analog equipment.

It did, and it could, and it was.

[huhn]

a bob deint is just a specialist scaling operation. so they could do such an operation at anytime they wanted in digital processing. they could take square film scans and make them anamorphic i'm sure they can also bilinear bob at that time...


the dvd is from 1996 so what ever they did in 1993 does not matter what matters is what they used as a source and them it was dumbed on a DVD like that.

the file is interlaced 30 that is not truly PSF while been PsF now. the number of soft trash fields is also relative low. the file is also full with weave artifacts while been not interlaced "anymore".

a messes up deint and it been fake interlaced is a very very good guess. if there are parts that are true interlaced or parts that are not trash doesn't matter we have this sample nothing else and this is that fake interlaced/PsF from a weaved bad interlaced source where if deint some fields are low res.