openexr-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Openexr-devel] Blender + OpenEXR


From: Bob Friesenhahn
Subject: Re: [Openexr-devel] Blender + OpenEXR
Date: Sat, 4 Mar 2006 00:54:08 -0600 (CST)

On Fri, 3 Mar 2006, Kai-Uwe Behrmann wrote:

ask Bob Friesenhahn. He has the probably best dpx/cineon implementation
out there in GraphicsMagic including the 16-bit colour management via lcms
on the output side.

I am interested in the answer so I am listening. :-)

The more I learn about color management for film, the more I see how ICC CMS does not adequately address the current methods. There is a bunch of magic employed, but it always seems to include a viewing LUT (e.g. to a Rec.709 calibrated monitor) which maps the film colors to the display. So the data is output referenced after it has been graded. If the viewing monitor can not be sufficiently well calibrated by itself, then an extra black box can be used to convert the SDI data so that correct colors are displayed for Rec.709 input.

For DCI digital cinema, colors are expressed in XYZ space so they are absolute and don't use a LUT. But a LUT is necessary to transform from the film colorspace into XYZ space for the DCI master. In this case, Rec.709 is the reference point to make that possible since the work is graded on a Rec.709 monitor. So ICC CMS could be used to implement the translation.

MPEG digital cinema is presumably Rec.709 and can use the same LUT as was used for viewing on a Rec.709 display except that the data has to be transformed into Rec.709 for delivery rather than just viewed through a LUT.

Lastly, DVDs are presumably in Rec.601 space which is yet another LUT. This could be handled via a simple ICC CMS transform, but it is also likely that the colorist will want to make additional adjustments based on what he sees in the output or based on experience.

None of the above includes OpenEXR which yet another variable to the mix.

The elephant in the closet is that since Rec.709 is the standard, it is impossible to properly evaluate colors which fall outside of the Rec.709 gammut.

If I have gotten anything wrong, please feel free to correct me.

Bob
======================================
Bob Friesenhahn
address@hidden, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]