Avisynth_challenged
5th July 2008, 01:26
Hi everyone. In my attempts to filter out rainbows in an analog capture, I seem to have inadvertantly found a good, simple way to denoise the chroma (U and V) parts of a YUV source.
Basically, I'm using the following line in my Avisynth script, which was intended to remove rainbows on an interlaced source, YUY2, prior to IVTC (Tcomb, chroma processing only):
Tcomb(mode=1,fthreshC=127,othreshC=127)
So basically, I've put the settings for fthreshC and othreshC right in the center of their possible range (0 - 255).
When I ran this on my raw capture, I noticed two things:
1. Most rainbows were substantially reduced, but the severest cases were still noticably present in the output. I think that these severe rainbows are in the actual source itself, which is a laserdisc, i.e. stored composite analog video, and thus cannot be filtered out with rainbow removal filters. (And I've tried a number of them!)
2. To my surprise, however, the Tcomb line in my script smoothened out the chroma quite nicely without a perceptible loss of detail. Here are some screenshots to show the results:
Image 1 of 3: Original unprocessed frame: (http://i252.photobucket.com/albums/hh18/FoolsTalk/chromatest1.png)
Image 2 of 3: Original unprocessed chroma mapped to luma: (http://i252.photobucket.com/albums/hh18/FoolsTalk/chromatest2.png)
Image 3 of 3: Filtered chroma mapped to luma: (http://i252.photobucket.com/albums/hh18/FoolsTalk/chromatest3.png)
So does anyone else here who deals with analog capped sources use Tcomb in this way, as a chroma denoiser? I hesitate to continue my work without consulting this amazing forum, because the Tcomb Readme file states:
"Once you've got values for mode, fthreshL/fthreshC, and othreshL/othreshC, add the necessary tcomb() line into your script and run through part of it. If you see any artifacts try lowering your fthresh/othresh values."
I don't see any artifacts, presumably because I'm using "mode=1" for chroma only. I searched the forums to see if anyone else was claiming Tcomb (mode=1) can be used for general chroma denoising (and not just for rainbow removal as described in the Readme). I found no postings to that effect.
So what does the community think? Any comments or feedback would be appreciated. Thanks :)
Basically, I'm using the following line in my Avisynth script, which was intended to remove rainbows on an interlaced source, YUY2, prior to IVTC (Tcomb, chroma processing only):
Tcomb(mode=1,fthreshC=127,othreshC=127)
So basically, I've put the settings for fthreshC and othreshC right in the center of their possible range (0 - 255).
When I ran this on my raw capture, I noticed two things:
1. Most rainbows were substantially reduced, but the severest cases were still noticably present in the output. I think that these severe rainbows are in the actual source itself, which is a laserdisc, i.e. stored composite analog video, and thus cannot be filtered out with rainbow removal filters. (And I've tried a number of them!)
2. To my surprise, however, the Tcomb line in my script smoothened out the chroma quite nicely without a perceptible loss of detail. Here are some screenshots to show the results:
Image 1 of 3: Original unprocessed frame: (http://i252.photobucket.com/albums/hh18/FoolsTalk/chromatest1.png)
Image 2 of 3: Original unprocessed chroma mapped to luma: (http://i252.photobucket.com/albums/hh18/FoolsTalk/chromatest2.png)
Image 3 of 3: Filtered chroma mapped to luma: (http://i252.photobucket.com/albums/hh18/FoolsTalk/chromatest3.png)
So does anyone else here who deals with analog capped sources use Tcomb in this way, as a chroma denoiser? I hesitate to continue my work without consulting this amazing forum, because the Tcomb Readme file states:
"Once you've got values for mode, fthreshL/fthreshC, and othreshL/othreshC, add the necessary tcomb() line into your script and run through part of it. If you see any artifacts try lowering your fthresh/othresh values."
I don't see any artifacts, presumably because I'm using "mode=1" for chroma only. I searched the forums to see if anyone else was claiming Tcomb (mode=1) can be used for general chroma denoising (and not just for rainbow removal as described in the Readme). I found no postings to that effect.
So what does the community think? Any comments or feedback would be appreciated. Thanks :)