[Cin] Another book on digital video

Terje J. Hanssen terjejhanssen at gmail.com
Sun Sep 15 21:48:10 CEST 2024



Den 15.09.2024 19:01, skrev Andrew Randrianasulu:
>
>
> On Sun, Sep 15, 2024 at 7:41 PM Terje J. Hanssen via Cin 
> <cin at lists.cinelerra-gg.org> wrote:
> >
> >
> >
> > Den 15.09.2024 18:02, skrev Andrew Randrianasulu via Cin:
> > > On Sat, Sep 14, 2024 at 1:08 PM Andrea paz via Cin
> > > <cin at lists.cinelerra-gg.org> wrote:
> > >> Sorry Phyllis, I have tried to follow the various posts about ffmpeg
> > >> color problems, but I can't figure out how it works. I get the
> > >> impression that even the ffmpeg developers are not clear on how to
> > >> handle color in their program. I tried reading Poynton's excellent
> > >> book, recommended by Andrew, a couple of years ago but stopped early
> > >> because it was beyond my capabilities.
> > > there is simpler colorful intro from around 2006
> > >
> > > https://tech.ebu.ch/docs/techreview/trev_301-gierlinger.pdf
> > >
> > >
> > >> There are too many variables, beyond the problems of ffmpeg: lack of
> > >> color management (which I think all reduces to an initial 
> transcode to
> > >> the absolute X,Y,Z color space from which then the program could do
> > >> all the transformations it needs without too much loss of color 
> info),
> > >> lack of support for ICC profiles, not knowing how each and every
> > >> function and plugin of CinGG treats colors, interaction between CinGG
> > >> and ffmpeg, and who knows what else. As Adam said, “There is no color
> > >> management workflow.” It is no coincidence that colorimetry in 
> editing
> > >> programs is found only in the big, rich, specialized companies.
> > >> --
> > >> Cin mailing list
> > >> Cin at lists.cinelerra-gg.org
> > >> https://lists.cinelerra-gg.org/mailman/listinfo/cin
> >
> > Just in case something here is not already found:
> > https://trac.ffmpeg.org/wiki/colorspace
> > 
> https://medium.com/invideo-io/talking-about-colorspaces-and-ffmpeg-f6d0b037cc2f
> > 
> https://kdenlive.org/en/project/color-hell-ffmpeg-transcoding-and-preserving-bt-601/

In short, the second url above shows how to keep BT.709 (for HD-video) 
during transcoding with ffmpeg,
while the the third url keeps BT.601 (for SD-video)

The second url also has references to additional reading in two 
articles, part 1 and 2 (here the English editions):

*What are Color Models? | Wiggle pixel
https://www.wigglepixel.nl/en/blog/what-are-color-models

*What are Color Spaces, Color Profiles and Gamma-Correct?
https://www.wigglepixel.nl/en/blog/what-are-color-spaces-color-profiles-and-gamma-correction 



>
> In this specific case I found following sentence helpful^
>
> ========
>
> n 1928, the mineralogist
> Siegfried Rösch described a
> body (Fig. 6), whose surface
> is constructed of optimum
> colours. Optimum colours
> are those among the colours
> (trichromatic), which are the
> brightest of the same chro-
> maticity and the most satu-
> rated of the same brightness.
> The base of this body is the
> diagram published by the CIE
> in 1931. It was constructed
> as the result of expansive
> tests with emmetropes. This
> diagram enables calculation
> of the position of every colour in relation to its primary
> colour. *This two-dimensional diagram is only valid for one
> respective colour temperature,* and it is a section of a
> funnel-shaped body (Fig. 7)
>
> =====
> bold by me.
>
> I do not think I've encountered this bit of info/history before.
>
> >
> >
> > --
> > Cin mailing list
> > Cin at lists.cinelerra-gg.org
> > https://lists.cinelerra-gg.org/mailman/listinfo/cin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cinelerra-gg.org/pipermail/cin/attachments/20240915/efa59549/attachment-0001.htm>


More information about the Cin mailing list