scanback profiling question (rendering intents)
scanback profiling question (rendering intents)
- Subject: scanback profiling question (rendering intents)
- From: "Tom Lianza" <email@hidden>
- Date: Sat, 10 May 2003 00:42:44 -0400
Hi to all,
This issue of rendering intent on the input side is very important and it will
be the subject of a great deal of discussion as our understanding of the needs
of the digital photographer becomes more advanced. The latest definition of
the PCS includes an assumed illumination level and a Black level. The PCS is
luminance range limited (As I believe that it should be). I believe that
current spec provides for a maximum luminance ratio of 288:1. Hence, most
transparency films will need to be mapped to this range in future profiling
efforts.
This new definition was done because it was impossible to take into account
scene luminance adaptation in the prior specification. In most imaging
applications outside of the ICC world, these assumptions are applied on the
input side. For instance, the work done by a number of early researchers in
photography, demonstrated that the mismatch between absolute luminance levels
of taking and viewing complicated the image reproduction goals. They found,
as the illumination level increases, both the scene contrast and colorfullness
seem to increase. Traditional slide films have very high gamma and very high
saturation because the images are usually viewed at illuminations levels that
are orders of magnitude less than those during the aquisition of the scene.
The only way to make the scene appear natural under these conditions is to
introduce the gamma and color distortion into the film. If the manufacturers
lowered the gamma for "more accurate" reproduction, the scene would appear
"less accurate" when viewed on the projection screen.
Another example of this can be found in nearly all video cameras. Although
the working gamma is assumed to be 2.2, the actual system gamma:
scene->signal->display is quite a bit higher. The reason for this is exactly
the same as the reason for increasing the gamma of the transparency film: the
viewing environment has a much lower level and our adaptation level is very
different than it was when we viewed the original scene.
For an ICC profile to be "accurate", this adaptation must be taken into
account. The raw data must then be "rendered" into the PCS. The exposure
information from the digital camera exposure system can be used to correct for
these factors and the next few years will bring that technology to light. The
big problem is that there is a real gap in the research and the big research
organizations that did all the fundamental work at the major film
manufacturers are gone and there really isn't a good substitute out there.
Hopefully Kodak and Fuji will bridge that gap, but they are certainly not
market leaders in the digital world and I have real doubts that they will have
the support of management to perform and publish the basic research that would
be necessary to advance the art.
So there you have it, more problems for color scientists to justify their
exsistance to their employers.....
Take care,
Tom Lianza
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.