[Tig] Color Grading in NLEs

Jean-Clement Soret jean-c
Fri May 16 12:46:54 BST 2003

Hi all,
I must say I do not share Craig's views about curves. Curves have really
opened a whole new world to me. When I was trying years ago to understand
how to replicate BW, Bleach BP or printing process using an OCN on telecine,
I took ramps (a linear gamma curve) recorded them on negative and looked at
how it was distorted by the process, and discovered it was all about curves.
Having only black, white and linear gamma control in these times gave me a 3
points curve which was not enough to replicate let's say a typical S curve
you get on a print, this was very frustrating and I had to rely on non real
time kit to get multiple points gamma curves. In those days sometimes we had
to do 2 gradings one for the highlights above 50% and one for the lowlights
then combine the 2 in post. Great results but very painful.
I understood that choosing a curve was like choosing a lab process, they can
be used to wrap the look of an image at the end allowing you to manipulate
the image underneath with a lot more range. Depending of the equipment you
use, curves will give you different results but they use the same
principles. The "creative package" on Spirit with B and W compression or
stretch is a curve with fixed points, so is the color corrector in DSX.
Colossus, IQ Fettle or Pandora Megagamma curves have huge possibilities, are
interactive and predictable when you get used to the translation between the
GUI and the effect on the image. 
I find pen and tablet far more adequate than using knobs when building a
curve because you create a transfer function that has to make sense and
avoid banding and solarisation which happens when your curve gets flat or
does not increment, with knobs you are never quite sure which part of the
curve you manipulate and end up doing a lot of trial and error.
I must admit in general when used too early in the chain of the color
correction process they tend to make things messy. Craig, I do respect your
opinion but I have the impression that your opinion might be slightly biased
by the kind of equipment you use.

Everyone now is talking about 3D LUTs, which is nothing else than a gamma
curve applied on an image but with its effect depending of the degree of
saturation. I would love to see a workable saturation dependent gamma
control in a color corrector.
Anyway, thanks for raising this subject.
Have a nice day.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://tig.colorist.org/pipermail/tig/attachments/20030516/3590a57f/attachment.htm 

More information about the Tig mailing list