Weird "shadowing" problem- help please?

[Chaobunny12]

It's not really working, and all I'm getting with these new filters is severe color distortion. The filter makes the image about ten times darker than it was before.

Here's my script:

Code: Select all

Loadplugin("C:\Program Files (x86)\AviSynth 2.5\plugins\TIVTC.dll")
mpeg2source("F:\D2V files\haru.d2v")
TFM(order=-1, field=-1, mode=5, PP=7, slow=2)
tdecimate()
clean = nnedi2(field=1)
amask = mt_makediff(clean, -1)
mt_adddiff(amask)

[Chaobunny12]

Oh, and now I remember why I couldn't use the FFT3D filter- when I try and load it I always get this error.

Avisynth open failure:
FFT3DFilter: Can not load FFTW3.dll !

The .dll file is in my plugins folder, and I even downloaded the .dll for this FFTW3, which I assumed was part of the FFT3D but apparently isn't. o_O

[mirkosp]

ah see, you got the code wrong. Here's what you should do:

Code: Select all

Loadplugin("C:\Program Files (x86)\AviSynth 2.5\plugins\TIVTC.dll")
mpeg2source("F:\D2V files\haru.d2v")
TFM(order=-1, field=-1, mode=5, PP=7, slow=2)
tdecimate()
original = last
clean = nnedi2(field=1)
amask = mt_makediff(clean, original)
mt_adddiff(amask)

As for FFT3DFilter, aside from having fft3dfilter.dll in avisynth's plugins directory, you'll need to place fftw3.dll in system32 (syswow64 or whatever is it called if your OS is 64bit).

[Chaobunny12]

Okay, thanks! I wasn't sure what "original" meant in the suggested script. But I got it to work now! The shadowed frames seem to be gone.

I had to change the field number to 0 instead of 1, so here's what finally worked:

Code: Select all

Loadplugin("C:\Program Files (x86)\AviSynth 2.5\plugins\TIVTC.dll")
mpeg2source("F:\D2V files\haru.d2v")
TFM(order=-1, field=-1, mode=5, PP=7, slow=2)
tdecimate()
original = last
clean = nnedi2(field=0)
amask = mt_makediff(clean, original)
mt_adddiff(amask)

Thanks very much to all of you for your help! ^_^

[Mister Hatt]

You could also do better on your TFM line if you want slightly less blur etc. See how TFM(mode=1, slow=2, micmatching=2, pp=7) goes. Alternatively to nnedi2 (read: if you want more speed) is to have TFM run clip2 against tdeint(mode=0, field= 0, type=3) and you should get a similar result.