[HELP] Telecine Vatiations Problems, Fixes and Questions...

[shijan]

The thing is those artifacts don't looks like normal combing interlaced artifacts. They not related to source video itself, but seems somehow produced by tape media itself during duplication or recording process. They look like separate additional layer of artifact.

Maybe i am not deeply learned Vinverse but it gives me two "looks" depending of "Strength" setting:
Strength 2.7 -  comb interlace artifact is partially removed and image oversharpened and jagged edges.
Strength 0...0.5 - comb interlace artifact is fully removed, but everything is softened.

---

Also AA filter looks essential for animation. It helps to fix jagged lines that almost always appears in line animation due low amount of pixel resolution on DVDs.
DAA also slightly softens image, so it is not perfect.
Here is side by side test Vinverse vs DAA.

---

Here is how my updated filters order look:
QTGMC Deinterlacer or TIVTC+QTGMC deteleciine
Color - Matrix Rec.601 to Rec.709
Color - Range Conversion (use only for video source that have incorrect levels)
DeBlock - QED
DeGrain - SMDegrain TR2 (helps to remove poorly compressed noise and film grain that looks more like mess of pixels. Useful for animation. Helps to make solid areas "like new")
AntiAliasing - DAA
DeRing - HQDering
DeHalo - (i prefer do not use it at all, because it adds too many "damage" to lines)
DeBand - GradFun3Mod
Resizer - NNEDI3 8x4 N128 N2
Sharpen - CAS 0.8 (sharpen applied after resize looks cleaner. Seems CAS strength is depends of resolution, to 0.8 applied at HD resolution produce same sharpness amount as 0.6 at SD resolution)
AddGrain - GrainFactory3 (This helps to emulate original grain structure, cleaned earlier by SMDegrain. It also helps to improve compression in solid color areas.)
Color - Tweak Chroma 0 (use for black and white films)

[Selur]

Okay, then why are you posting this in this thread which you started to speak about telecine and interlaced artifacts?
That does not make any sense!

-> you should may be spend at least some time to figure out what artifacts you got and then look think about it.
RemoveDirt and similar stuff might be more suited.

Cu Selur

[Selur]

Color - Range Conversion (use only for video source that have incorrect levels)

I really would not recommend using that since I'm pretty sure it doesn't work the way it should.
Using Color Matrix and if nececessary before it overwriting the input color matrix in a custom addition would be the better approach.

Color - Tweak Chroma 0 (use for black and white films)

Why not use it as first filter instead of last?

If the rest of the order works fine for you that's okay.

Cu Selur

[shijan]

So yeah, i find two DVDs with incorrect levels. So i used Levels filter manually set "out min/max set to 0-255" or Range Conversion. Both produced the same result. So Range Conversion seems works.
Levels filter have advantage - it could be placed to any order and applied after filters. So filters may be applied to uncorrected "dull" source video and may produce better quality.
Also in x264 settings i manually force set to VUI ot "tv".
Still not sure about DAA and other AA filters. As you noticed they improve aliased parts of the image but same time may smooth and soften other fine lines that are not real aliasing. Will test more how DAA behave when applied after resize...

"Color - Tweak Chroma 0 Why not use it as first filter instead of last?"
I don/t know really. Thought maybe filters may work better with color pixels data instead of pure greyscale. But maybe not. Need to test it...

[shijan]

I noticed that a plenty of animation DVDs use some sort of special uncommon telecine logic. In some scenes used real 2:3 Pulldown pattern (3 progressive, 2 interlaced), but in some other scenes sometimes pattern changes and it looks like 4 progressive frames 1 interlaced. Sometimes pattern changes to 1 progressive 1 interlaced. In many cases progressive frames in those uncommon patterns have damaged jagged lines.

Usually there is no 2:3 Pulldown tags in those original VOB files.
MediaInfo detect files as:
29.970 (30000/1001) FPS NTSC
Scan type: Interlaced
Scan order: Top Field First

If i manually set input in Hybrid as Progressive and use TFM of VFM filters, i can output to 29.97 and in worst case scenario i can see that in those problematic dynamic scenes pattern consists of one normal deinterlaced nice looking frame and one similar original frame with jagged lines.

If i add sRestore - it randomly use those jagged frames instead of normal looking duplicates.

If i manually set input in Hybrid as telecine and use VIVTC or TIVTC - i can see that some random pattern sometimes jagged frames may appears instead.

VIVTC/TIVTC can remove most of those problematic frames, but some amount of jagged unfortunately remains. I tried different modes and settings and none of them really helps. Changing Block (x/y) and MI settings to large amounts helps to drop more jagged frames, but same time it adds blending artifacts to other normal frames. So it is not a real solution.

So here are some questions.
Why sRestore and VIVTC or TIVTC select those low quality progressive jagged frames instead of nice looking deinterlaced frames?
Is it some bug in telecine logic related somehow to Hybrid? Maybe some script could be fine tuned to fix this problem?
Or is it some global bug with plugins or with some settings not related to Hybrid that should be discussed in other forum?

Here is small example of source clip https://drive.google.com/file/d/1gIPkJMV...sp=sharing
If use 29.970 framerate problem starts from frame 2342
If use 23.976 framerate problem starts from frame 1873

Input: Progressive


Input: Progressive + TFM


Input: Telecine + TIVTC

[Selur]

Not knowing your settings and not having a source sample to reproduce the problem.
-> Ask the author of the filters.

[shijan]

Seems you missed something in my post. I posted link to source clip sample https://drive.google.com/file/d/1gIPkJMV...sp=sharing

No any special settings used. Default VIVTC/TIVTC as well as any other special preset will give the same result.

---

Just in case if it is not a Hybrid related problem, i also started thread at doom9 https://forum.doom9.org/showthread.php?t=183265

[Selur]

Yeah missed the link.

I noticed that a plenty of animation DVDs use some sort of special uncommon telecine logic. In some scenes used real 2:3 Pulldown pattern (3 progressive, 2 interlaced), but in some other scenes sometimes pattern changes and it looks like 4 progressive frames 1 interlaced. Sometimes pattern changes to 1 progressive 1 interlaced.

So it's no clean telecine and just using IVTC won't fix it?

No any special settings used.

And there is you problem, assuming that defautl settings or presets are good are any source.

# Deinterlacing using QTGMC
clip = havsfunc.QTGMC(Input=clip, Preset="Slower", InputType=0, TFF=True, TR2=3, SourceMatch=0, Lossless=2, EZDenoise=2.00, NoisePreset="Slow", MatchEnhance="10.00", Denoiser="dfttest") # new fps: 59.94
# make sure content is preceived as frame based
clip = core.std.SetFieldBased(clip, 0)
# adjusting frame count and rate with sRestore
clip = havsfunc.srestore(source=clip, frate=23.976, omode=6, speed=-25, thresh=16, mode=2)

seems fine -> https://filebin.net/gblgumvvh949wvdx

Cu Selur

[shijan]

Yes, it's no clean telecine. It is some variable pattern. Just using IVTC mostly fix it, but in some places it keeps those original jagged frame duplicates instead of normal looking ones. This is probably due some variable telecine pattern.

Deinterlacing telecined animation using QTGMC bob is a bad idea because it damage original lines a lot and lowers the resoilution.

AA filters may help a little but they don't fully fix those jagged frames and also soften normal frames.

As you may see from provided gif animation "Input: Progressive + TFM" - There are plenty of normal (non-jagged) frames that may be normally used, but for some reason they are dropped. And instead jagged frames duplicates used by sRestore or by IVTC.

[Selur]

And instead jagged frames duplicates used by sRestore or by IVTC.

not surprised that sRestore would prefer those lines, since it assumes it gets properly progressive input and it's about avoiding ghosting not combing.
Point is I doubt there are settings for either VIVTC (VFM+VDecimat) or TIVITC (TFM + TIVTC) using:

clip = havsfunc.QTGMC(Input=clip, Preset="Fast", TFF=True,FPSDivisor=2)
clip = core.tivtc.TDecimate(clip=clip)# new fps: 23.976

might be better.

"Input: Progressive + TFM" - There are plenty of normal (non-jagged) frames that may be normally used

Yeah, normally ivtc selects 4 out of 5 frames. you unless you find 4 good in 5 problematic frames ivtc has no chance.
as a side note: AnimeIVTC might work, but then again that would do tons of filteringe etc.

Cu Selur