Remove Checkered Colored Bleeds & Semi Diagonal Waves From My Digital Captures

[levi]

This comes from an antenna signal connected via coaxial cable to my capture card. I don't see it on HDTV captures - only on SDTV captures broadcast on the digital channel.

I'd like suggestions on how to handle these via filters


http://img367.imageshack.us/img367/7118/69764537hj9.th.png (http://img367.imageshack.us/my.php?image=69764537hj9.png) - Semi Diagonal Waves


http://img396.imageshack.us/img396/6981/25348059pl1.th.png (http://img396.imageshack.us/my.php?image=25348059pl1.png) - Checkered Colored Bleeds


Is this an issue with my card? Thanks :o

[MadRat]

That's called dot crawl (or dotcrawl or whatever) and it's notoriously hard to get rid of. I was just talking about dotcrawl to a guy who says he knows a lot of the script writers here and he says this can help...

LanczosResize(width*2,height*2).TurnLeft().SangNom().TurnRight().SangNom().LanczosResize(width,height)


I've never tried it. I get this a lot from my video capture card and normally use tcomb(), however tcomb only works on NTSC video before it's deinterlaced.

[levi]

I tried this suggestion, but it didn't help with my source.

Here is what did work (maybe it will help someone else):

I was able to remove the incorrectly placed coloring with the FixChromaBleeding() script.
I was able to mostly address the dot crawl with checkmate()

I did not find anything to address the diagonal waves - If anyone has a suggestion for that, I'd appreciate any advice.

[R3Z]

Try F2Quiver (from vcmohan) or DeFreq (from Fizick). I prefer F2Quiver using butterworth point symmetry band ("bp"). Essentially you need to call F2Quiver like so (the author recomends using it after separatefields() );

SeparateFields()
F2Quiver(test=true, Gamma=have to play with it usually < 1.0)
Weave()

That will give you a frequency spectrum and you should look for white dots such as these;

http://www.users.on.net/~inet_s51/pref/f2before.jpg

You may need to change the gamma value or step through the footage to find the white dots.

Its then a case of patching those white dots with Horizontal and Vertical degree points. So your script would look like;

SeparateFields()
F2quiver("bp3", 100,100)
F2Quiver(test=true, Gamma=your value) # This must be here to show your "patched" points.
Weave()

BP denotes the size of the point to use in the spectrum, the higher the greater.

Heres one i did;

http://www.users.on.net/~inet_s51/pref/f2after.jpg

See all the black dots ? they are points notched out of the spectrum. The code i used was;

f2quiver (
\"bp3",98,194,
\"bp3",96,-192,
\"bp3",96,64,
\"bp3",96,-64,
\"bp3",0,128,
\"bp3",195,0,
\"bp4",50,-96,
\"bp4",50,160)

I hope thats pretty clear, i had no clue when i first started trying to tackle the issue and no matter what i read, i couldnt get my head around it.

R3Z

[MadRat]

That's a great tutorial R3Z, thanks!

[Dogway]

I cant quite understand this filter, I know I have to patch, but how do I patch a diagonal line?
EDIT: I discovered, blb1 makes it.

Now I cant make it work with F2QuiverUV, it doesnt work, at least in an image Im working on.
I keep the flags like these:

F2QuiverUV (
...
\"bp3",95,210,
\"bp3",95,235,
\"bp3",33,145,
\"bp3",47,-35,
\"blb1",195,170,
\"blb1",30,-200)
F2Quiver(test=true, Gamma=0.08)