On Mon, Apr 10, 2006 at 07:31:17PM +0200, Nicolas MAUFRAIS wrote: > On Mon, Apr 10, 2006 at 07:14:17PM +0200, Nicolas MAUFRAIS wrote: > > On Mon, Apr 10, 2006 at 09:35:29AM +0200, Stefan M. Fendt wrote: > > > Stefan M. Fendt schrieb: > > > > > > >Some additional tests: > > > > > > > > > > > This becomes weird... > > > > > > Now, I did use _exactly_ the same commandline as you for the sequence > > > you sent me... but, no golfball, no green... ?!? Could you pls. mail me > > > one example frame (TGA,with framenumber) out of "04.yuv" processed by > > > that commandline on your machine? I repeated the test by using my > > > neutral gray-chart and still... no green... weired... > > > > Hello, > > > > I confirm there's a really strong green tint over my encoded video. > > Unfortunately, I didn't find any suitable extract to put on the > > internet (there're kids on the picture). I'm searching for a part of the > > video and I'll upload it. > > I think I probably over-denoised the chroma. > > > > Nicolas. > > OK. I put on the internet the encoded extracts. > For those of you who are interested in the original YUV file, it can be > downloaded here: > http://www.europephoto.com/info/montage_video/04.yuv.tgz > The AC3 stream if you want to mplex the streams: > http://www.europephoto.com/info/montage_video/04_extract.ac3 > > For my test, I did use different parameters for yuvdenoise and > mjpegtools versions. The CVS mjpegtools are installed on my computer > with the _cvs_20060409 suffix. > > Normal encoded video (no green tint): > yuvcorrect -v 0 -T INTERLACED_BOTTOM_FIRST | yuvscaler -v 0 -I > ACTIVE_702x560+8+8 -M BICUBIC | y4mspatialfilter -L 4,0.94,4,0.94 -C > 3,0.94,3,0.94 | yuvmedianfilter -v 0 -t 0 | yuvdenoise -v 0 -s 1,2,2 -g > 0,0,0 -t 2,3,3 | mpeg2enc -v 0 -K tmpgenc -r 32 -4 1 -2 1 -D 10 -E 10 -g > 15 -G 15 -q 5 -b 9500 -f 8 -o $1 > The result is: > http://www.europephoto.com/info/montage_video/test_v18.m2v > > The encoded video with a green tint: > yuvcorrect_cvs_20060409 -v 0 -T INTERLACED_BOTTOM_FIRST | > yuvscaler_cvs_20060409 -v 0 -I ACTIVE_702x560+8+8 -M BICUBIC | > y4mspatialfilter_cvs_20060409 -L 4,0.94,4,0.94 -C 3,0.94,3,0.94 | > yuvmedianfilter_cvs_20060409 -v 0 -t 0 | yuvdenoise_cvs_20060409 -v 0 > -m6,24,24 -t12,24,24 -M3,24,24 | mpeg2enc_cvs_20060409 -v 0 -K tmpgenc > -r 32 -4 1 -2 1 -D 10 -E 10 -g 15 -G 15 -q 5 -b 9500 -f 8 -o $1 > The result is: > http://www.europephoto.com/info/montage_video/test_v19.m2v > > As I said before, I think the problem comes from over-denoising the > chroma channel. I read this in the Mjpeg howto: > Sometimes, depending on the source material, median filtering of the > chroma can cause a slight color shift towards green. Filtering on the > luma component (disabling the chroma filtering) is the solution to that > problem. > > As just put on the internet the encoded videos to show you the problem. > I'm doing some tests right now. > > Thank you. > Nicolas.
OK. I did my tests. I used various versions of the mjpegtools, and I confirm the green tint comes from yuvdenoise CVS. The other tools are not responsible. The green tint appears even with low values for yuvdenoise, such as: yuvdenoise_cvs_20060409 -v 0 -m2,6,6 -t2,6,6 -M2,6,6 However, I just observe the green tint. I'm not quite sure it's a bug! I read in a paper magazine video CCD noise is often made of RED pixels. Wouldn't yuvdenoise remove those red dots, and therefore increasing the overall green tint? Moreover, even if the encoded "green" videos are "greener" than the version I got with the "non-CVS" yuvdenoise, I recognize the picture in the original footage is a bit too red! That's probably due to those "red dots". Is there a way to reduce the green component in the final video, while still using the CVS version of yuvdenoise? Because that version is REALLY efficient in removing the noise. I really like the result (excepted for the green tint). Nicolas. ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 _______________________________________________ Mjpeg-users mailing list Mjpeg-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mjpeg-users
