Glitch/Dropout Removal

[TruthOverFacts]

https://www.mediafire.com/file/o4h7geg2t5vnrd5/GFclip.avi/file

Has anyone seen glitches like these before? What are the best options for removing them?

Any suggestions would be appreciated.

[Selur]

how about: TIVTC() + ChromaShiftSP(X=2.5) + QTGMC(InputType=3) ? (maybe also add DeSpot())
If machine learning stuff is an alternative also throw 1x-BleedOut-Compact (https://openmodeldb.info/models/1x-BleedOut-Compact) at it.

Cu Selur

[johnmeyer]

What glitch am I supposed to be seeing? I see no artifacts.

[TruthOverFacts]

They are thick, transparent, and look like bars.

https://i.ibb.co/S449Nmxg/vlcsnap-2026-06-13-04h47m05s459.png

[Selur]

best upload image to i.e. https://imgbb.com/ or similar and post a link

[TruthOverFacts]

how about: TIVTC() + ChromaShiftSP(X=2.5) + QTGMC(InputType=3) ? (maybe also add DeSpot())
If machine learning stuff is an alternative also throw 1x-BleedOut-Compact (https://openmodeldb.info/models/1x-BleedOut-Compact) at it.

Cu Selur

Thanks, are you recommending QTGMC/DeSpot specifically for the horizontal dropout lines, or are TIVTC and ChromaShiftSP addressing separate issues you noticed in the sample? Just want to understand what's addressing what is all.

[johnmeyer]

I downloaded and zoomed into those pics. I still don't see much of anything. It should doesn't look a glitch of any kind. Certainly not a dropout.

[Selur]

I added ChromaShiftSP and BleedOut since the chroma seems to be shifted and bleeding (see the right side of the face)
https://i.ibb.co/VYvPDGCm/grafik.png (https://ibb.co/0yfdZL9t)