[Cin] "Unbounded" floating point image manipulation

[Andrea paz]

Thank you for the detailed explanation. The sentence "So, unless one
deals with sequence of tiff/exrs there always will be at
least one step between what libavcodec outputs (bunch of integers) and
what cinelerra-gg can accept (32fp at best)" together with what Georgy
explained, indicate that not much can be done, not only because of
CinGG and its plugins, but especially because of ffmpeg.
Out of curiosity I tried using CinGG's internal engine instead of the
usual default ffmpeg. Tests with the two Blend Program give the same
results as with ffmpeg. I should try a pipeline with EXR sequences and
see if you can keep the color consistent.

Looking at the ffmpeg-devel mailing-list, I understand that it is
impossible to bring ffmpeg to work in float, they would probably have
to rewrite everything from scratch (I saw that you reported our thread
with no response...).

Arch:
startx requires you to install Xorg (I don't know about Wayland).
Maybe that is the problem? Why did you put Arch in anyway, isn't that
just a problem with Cmake?