Re: Of colorful scepticism
Re: Of colorful scepticism
- Subject: Re: Of colorful scepticism
- From: email@hidden
- Date: Thu, 27 Sep 2001 14:22:14 -0500
Henrik Holmegaard wrote:
>
This is one of the most common metaphysical labyrinths
Maybe. I'm not a professional filosopher, so my reasoning may not be
perfect.
But I'd prefer some practical arguments over comments on my style of
reasoning.
>
Fact is: I spend a lot of my time looking at photo's and making judgements
about their colours. I can tell you I tried using camera profiles and it
simply didn't work. And thinking about it afterwards I came to the
conclusion camera profiles simply don't make sense. Maybe my reasoning is
flawed, but the facts aren't. Is that clear enough?
It appears that your reasoning "is" flawed. Camera profiles don't work
well, but not for the reasons you mention. Because many if not most of the
digital cameras CCD arays or Sensor arays (CMOS) have the ability to
capture a much larger color gamut than even the visual spectrum, it is a
project in futility sometimes to accurately profile this type of device
with "normal" tools available in most profiling packages today. To do this
profiling properly, we must be able to measure the spectral response of the
sensor aray in the device through the lens.
If some clever person comes up with the "Place over the Lens profiling
device" then only white point ajustments are needed.
>
btw:
>
In the course of thinking about the subject I once did a little
experiment.
>
I took a standard A4 NCS colour sample and within minutes I took some 20
>
shots on slightly different locations, using one and the same film, using
>
the best possible adjustments of the camera (a nikon F601). Then I scanned
>
the film, all images with the same adjustments. Just for fun I measured
the
>
deltaE values in the middle of the photographed samples, compared to the
>
CIElab value of the NCS sample. Minimum deltaE was 7.5, maximum was 39.2.
>
The fact that the photographed ncs sample differed from the original
should
>
come as no surprise. But from the fact that the photographed samples had
>
very large differences among eachother, has led me to the conclusion that
a
>
profile would be useless.
>
I'll be happy to mail the detailed results to anyone interested.
I would be supprised if this method generated accurate results. :)
>
There are literally hundreds of papers that will help you out of this
>
misconception about color
Ok, if you can point me some on the web, I'll certainly look into it.
I'll do my best to prevent myself from waisting mine and your bandwith with
'descriptive logic', if I possibly can.
:-)
Ok start here:
http://www.inforamp.net/~poynton/notes/links/color-links.html
Igor Asselbergs
Best regards,
Bob Hallam
Color Specialist
Que-Net Media
A Division of the Quebecor World Group
135 East Algonquin Rd.
Arlington Heights
IL 60007
USA