Ok ... will stick to imagemagick then since I work on RGB or sequence of
TIFF files of film frames.
Denoising needs to be done before color correction and since I have
either 8 or 16 bits TIFF files, I might as well use what all the data
before reducing to the YUV colorspace
Cheers
E
On 04/08/2011 07:06 AM, [email protected] wrote:
Question 2 :
- yuv4mpegpipe is a YUV-only output format.
- I strongly doubt that hqdn3d is able to process RGB data. I didn't find any
clear information about that for ffmpeg, but avisynth's port of hdqn3d
processes only YUV video (http://akuvian.org/src/avisynth/hqdn3d/hqdn3d.txt).
- the only filters I know that are able to denoise RGB video are working with
avisynth (see http://www.aquilinestudios.org/avsfilters/spatiotemp.html)
- but, in the last command you provide as an example, you finally end up with
YUV-family video (dnxhd), so is there really much point in denoising RGB ?
Question 1 :
I compared different denoisers, only in the YUV colorspace family. The best
(but slower) ones work inside avisynth :
speed<---------------------> quality
hqdn3d< fft3dfilter< mctemporaldenoise
MCTemporalDenoise( preset="Medium" ) does some king of magic on noisy sources.
You can input cinelerra-rendered mov video into avisynth with the QTSource
external filter.
Since I work almost only with YUV-video, I usually render from cinelerra using the
"YUV4MPEG Stream" option, piping this stream to mencoder to produce ffvhuff in
an avi container, and then I process this avi inside avisynth.
D
_______________________________________________
Cinelerra mailing list
[email protected]
https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra
