How do I remove ghosting& screendoor/canvas effect from a DVD(cartoon show)

[dinoman]

I have rugrats dvds but there is a lot of ghosting and some sort of screendoor effect that looks like small diamonds shapes or similar to a canvas. I watched the Itunes version of the show and the frame rate is smoother and without ghosting. Plus Itunes version doesn't have that screendoor effect and has slightly better details too
Here is the comparison between the DVD and the itunes version:

DVD with annoying screen door effect:


[Attachment 55972 - Click to enlarge]

Itunes version:


[Attachment 55973 - Click to enlarge]

Thank you

[dinoman]

Here is the ghosting that I was talking about:


[Attachment 55974 - Click to enlarge]

[jagabo]

You need to provide a short sample of the source in its original MPEG 2 format. But the noise looks like residual PAL chroma carrier. My guess is you have a PAL DVD. You'll probably have to use AviSynth.

[dinoman]

I used handbreak to "cut" the VOB file and make the sample and I chose MPEG-2 option as "video codec". I need to mention that I have no idea what I'm doing...
https://files.videohelp.com/u/297545/VIDEORUG3.mp4

[jagabo]

Handbrake messed up the video. You can use DgIndex to open a VOB file and snip out a small section. Download the DgMpgDec package:

https://www.videohelp.com/software/DGMPGDec

Extract DgIndex.exe to a suitable location -- say, your Desktop. Run it. Open one of your VOB files with File -> Open. Use the slider and the mark-in and mark-out tools ([ and ]) to mark a selection. Include medium speed panning shot. Use File -> Save Project And Demux Video. Upload the resulting M2V file. That will be an copy of the original MPEG 2 data from the DVD.

[dinoman]

When I open the file nothing happens. I also tried an older version of dgindex.exe and the same thing: File->Open-> I select the vob file but then nothing happens other than the file list window appearing which is also empty.

[jagabo]

When I open the file nothing happens. I also tried an older version of dgindex.exe and the same thing: File->Open-> I select the vob file but then nothing happens other than the file list window appearing which is also empty.

After selecting a VOB the File List dialog should look something like this:


[Attachment 55977 - Click to enlarge]

Then press the OK button to return to the main window. I've never seen this fail.

Another program you can try is Mpg2Cut2. Open a VOB, select a section, File -> Save This Clip (as MPG).

[dinoman]

The Vob file doesn't appear in the file list and even if I press ADD and select the vob file it still won't show up. I will try mpg2cut2 next. Thank you

[jagabo]

Are you trying access the VOB file directly from an encrypted DVD? You have to decrypt to a VIDEO_TS folder first. Then access the VOB from there.

[dinoman]

It was because of the name of the folder. I had to change the folder's name in order to work. https://files.videohelp.com/u/297545/VTS_01_2.demuxed.m2v

[dinoman]

Are you trying access the VOB file directly from an encrypted DVD? You have to decrypt to a VIDEO_TS folder first. Then access the VOB from there.

What is your opinion? What can I do to fix the ghosting and the screendoor/canvas effect? Thank you

[davexnet]

Are you trying access the VOB file directly from an encrypted DVD? You have to decrypt to a VIDEO_TS folder first. Then access the VOB from there.

What is your opinion? What can I do to fix the ghosting and the screendoor/canvas effect? Thank you

I couldn't find a obvious way to clean the blends so here's an attempt when all else fails,
reduce the frame rate to the clearest 12 1/2 frames per second, then double the frame rate back to 25.
But the fluidity does suffer

Fanfilter is documented to help with this interference pattern but it's not well documented.
Here it reduced it a little
http://avisynth.nl/users/vcmohan/FanFilter/FanFilter.html


Is this a commercial DVD? This quality is poor. Which episode is this?

[jagabo]

One way to reduce the residual dot crawl is to downscale the frame width, then scale it back to full width. It doesn't hurt the resolution too much, especially with a little sharpening.

Davexnet is right about the blending -- there's too much to remove it all. Most of the character animation in cartoons is at 12 fps so reducing the frame rate to 12 fps is only a problem for the panning shots (which are at 24 fps). But at normal playback speed the residual blending of 24 fps is much less annoying than the jerky panning of 12 fps.

Mpeg2Source("VTS_01_2.demuxed.d2v", CPU2="ooooxx", Info=3) 
BilinearResize(440, height) # blurs away most of the residual dot crawl
QTGMC(EZDenoise=1.0) # double rate deinterlace, mild noise reduction to further remove residual dot crawl
SRestore(frate=23.976) # restore original 23.976 NTSC frame rate, try 11.99 if you want to see the no-blending option
aWarpSharp(depth=5) # sharpen edges a bit
Spline36Resize(720, height) # restore the frame width
Sharpen(0.5, 0.0) # sharpen horizontally

[davexnet]

One way to reduce the residual dot crawl is to downscale the frame width, then scale it back to full width. It doesn't hurt the resolution too much, especially with a little sharpening.

Davexnet is right about the blending -- there's too much to remove it all. Most of the character animation in cartoons is at 12 fps so reducing the frame rate to 12 fps is only a problem for the panning shots (which are at 24 fps). But at normal playback speed the residual blending of 24 fps is much less annoying than the jerky panning of 12 fps.

Mpeg2Source("VTS_01_2.demuxed.d2v", CPU2="ooooxx", Info=3) 
BilinearResize(440, height) # blurs away most of the residual dot crawl
QTGMC(EZDenoise=1.0) # double rate deinterlace, mild noise reduction to further remove residual dot crawl
SRestore(frate=23.976) # restore original 23.976 NTSC frame rate, try 11.99 if you want to see the no-blending option
aWarpSharp(depth=5) # sharpen edges a bit
Spline36Resize(720, height) # restore the frame width
Sharpen(0.5, 0.0) # sharpen horizontally

Yes this is what I saw. I was going to upload my result, but I'm having connection problems today.
I may do so later, when I'm able, as a comparison

[jagabo]

Attached is a 11.98 fps version for comparison.

It might be possible to detect panning/zooming shots and use 23.976 fps for those and use 11.98 fps with frame doubling for the rest. Off the top of my head I don't know how to do that though.

[frlev120]

One way to reduce the residual dot crawl is to downscale the frame width, then scale it back to full width. It doesn't hurt the resolution too much, especially with a little sharpening.

Davexnet is right about the blending -- there's too much to remove it all. Most of the character animation in cartoons is at 12 fps so reducing the frame rate to 12 fps is only a problem for the panning shots (which are at 24 fps). But at normal playback speed the residual blending of 24 fps is much less annoying than the jerky panning of 12 fps.

Mpeg2Source("VTS_01_2.demuxed.d2v", CPU2="ooooxx", Info=3) 
BilinearResize(440, height) # blurs away most of the residual dot crawl
QTGMC(EZDenoise=1.0) # double rate deinterlace, mild noise reduction to further remove residual dot crawl
SRestore(frate=23.976) # restore original 23.976 NTSC frame rate, try 11.99 if you want to see the no-blending option
aWarpSharp(depth=5) # sharpen edges a bit
Spline36Resize(720, height) # restore the frame width
Sharpen(0.5, 0.0) # sharpen horizontally

Yes this is what I saw. I was going to upload my result, but I'm having connection problems today.
I may do so later, when I'm able, as a comparison

MPEG2Source(chemin, idct=5, info=3)

### Deinterlace-Match Fields-Decimate ###
Function FieldMatch(Clip Cee) {
  Global PeePee = Cee.DuplicateFrame(0)
  Global CeeCee = Cee
  Global NeeNee = Cee.DeleteFrame(0)
  Pee2 = PeePee.SeparateFields()
  Cee2 = CeeCee.SeparateFields()
  Nee2 = NeeNee.SeparateFields()
  Global PeeCee = Interleave(Pee2.SelectEven(),Cee2.SelectOdd()).Weave()
  Global CeePee = Interleave(Cee2.SelectEven(),Pee2.SelectOdd()).Weave()
  Global CeeNee = Interleave(Cee2.SelectEven(),Nee2.SelectOdd()).Weave()
  Global NeeCee = Interleave(Nee2.SelectEven(),Cee2.SelectOdd()).Weave()
  Global Deintee = QTGMC(CeeCee).SelectEven()
  Return ScriptClip(CeeCee, \
    "!CeeCee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? CeeCee : " + \
    "!NeeNee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? NeeNee : " + \
    "!CeeNee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? CeeNee : " + \
    "!NeeCee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? NeeCee : " + \
    "!PeePee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? PeePee : " + \
    "!CeePee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? CeePee : " + \
    "!PeeCee.IsCombedTIVTC(CThresh=12,Chroma=True,BlockX=16,BlockY=32) ? PeeCee : Deintee")
}

AssumeTFF()
RoboCrop(LeftAdd=6, TopAdd=2, RightAdd=4,  BotAdd =2, align=True)
TFM(last, Order=-1,Mode=5,PP=2,Clip2=FieldMatch( last ),Slow=2,MChroma=False,Ubsco=False,CThresh=12,Chroma=True)

vinverse_avsi().DFMDeRainbow(maskthresh=10)

source=last
source.frfun7(lambda=1.1, T=6.0, Tuv=2.0).minideen(radiusY=3, radiusUV=3, thrY=10, thrUV=12, y=3, u=2, v=23).Extracty() #converttoy(matrix="Rec709")
Tweak(hue=0.0, sat=1.00, bright=-50, cont=1.50, coring=True, sse=False, startHue=0, endHue=360, maxSat=150, minSat=0, interp=16).frfun7(lambda=1.1, T=6.0, Tuv=2.0).FluxSmoothST(temporal_threshold=3, spatial_threshold=7, luma=true, chroma=true)
a = last.converttoy().mt_edge("cartoon",thY1=4.5,thy2=4.5).Removegrain(27).Blur(1.58)
b = last.converttoy().mt_edge("cartoon",thY1=20,thy2=20.0).Removegrain(27).Blur(1.58)
masque = mt_hysteresis(b,a).Santiag().mt_deflate()

denoised_clip = source.GaussianBlur(varY=0.75,VarC=0.75).MiniDeen(radiusY=2, radiusUV=2, thrY=6, thrUV=7, y=3, u=3, v=3)
denoised_clip = denoised_clip.MCTemporalDenoise(settings="very low", radius=2, Chroma=False, edgeclean=true, ecrad=6, stabilize=true, maxr=2, AA=True, useEEDI2=True, maxd =24, protect=True)
denoised_clip = denoised_clip.SMDegrain(TR=3,ThSAD=600,ContraSharp=True,RefineMotion=True,Plane=0,PreFilter=2,Chroma=False, DCT=5, pel=2)
denoised_clip = denoised_clip

clip_dark =  source.GaussianBlur().Hysteria(2).ConvertToY(matrix="Rec709").LSFmod(defaults="slow", strength=100,soft=-2,  ss_x=1.25, edgemode=0, preblur="FFT3Dfilter(sigma=4,plane=0)")
denoised = mt_merge(denoised_clip,clip_dark , masque).LSFmod(defaults="slow", strength=100,soft=-2,  ss_x=1.25, edgemode=0, preblur="FFT3Dfilter(sigma=4,plane=0)")
denoised.HQderingmod().EdgeCleaner(strength=8, rep=True, rmode=17, smode=1, hot=False)
Hysteria().LSFmod(defaults="slow", strength=100,soft=-2,  ss_x=1.25, edgemode=0, preblur="FFT3Dfilter(sigma=4,plane=0)")

EDI_RPow2(4,CShift="Spline36Resize",FWidth=720,FHeight=576, nsize=0, nns=4, qual=2, LSB=True, lsb_out=False, taps=4, etype=1)
neo_f3kdb(preset="low", sample_mode=2, range=10)

Not sure that script is optimal as it's a script for another cartoon I'm working on.

It might be to intense. I marked in blue the parts I think are helping the most. I think it could be a little sharper. But that's my disease.

You can play a bit with the mask (masque) by changing thY1 and thy2 in "a" and "b", you can see the mask result by putting "return masque" after the line "masque = ..."


Here's more on mt_hysteresis https://forum.doom9.org/showthread.php?t=128458