>| So, what's the deal? Is this an actual error in >| yuvdenoise/medianfilter, or an expected artifact from denoising the >There are definitively errors/inaccuracies in the lav-tools concerning the >range of YUV-values. yuvdenoise tends to broaden the [16-235]-range of Y.
Not just the lavtools... these kinds of errors are rampant in all the digital video code I've looked at over the last year or two. It's one of those funny/sad things: so many people copy other people's code, but they never do the research to see if that original code actually did the right thing to begin with. Anyhow, Stefan Fendt wrote yuvdenoise, and I think he is the only person who is going to be able to comment on the Y'CbCr fidelity of his code. >assumes [0-255]-ranges! At least mpeg2enc needs the correct >(small) range to encode. It clips out-of-range values. For me this results I was under the impression that mpeg2enc does *not* clip out-of-range values and does simply encode them. Although values outside of [16,235] do not map to R'G'B' codes, they are still valid values --- and necessary even, if you want to encode things like standard colorbars. Someone please correct me if I am wrong. -matt m. ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ Mjpeg-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/mjpeg-users
