I am not doing any scaling at all. I am leaving all of the math up to
Little CMS, which is my purpose for using it.

I know from experience that a 1.5 delta E is subtle. The difference in
RGB values (displayed RGB versus lcms RGB calculated from measured Lab)
that I am seeing are 6+ delta E off, which is in no way subtle. I have
to assume that either I am doing something wrong, that Little CMS is
seriously flawed, or that I am completely misunderstanding what a color
management system does.

For instance, if I display a swatch on my LCD with an RGB value of 85,
85, 85 and measure it with a spectrophotometer, I get an Lab value of
42, 0, -0.6. If I ask lcms, given my monitor's recent profile and the
default Lab profile, to transform this Lab value to RGB, I expect to get
85, 85, 85, or something close to it. I don't. I get 96, 98, 99, which
is off by 13%, 15%, and 16%.

If it is something that I am doing wrong, I would be grateful to find
out what it is so I can fix it, because having this much error in the
results is unacceptable.

Thanks for your reply,

Mike.


Kai-Uwe Behrmann wrote:
> Am 08.12.07, 18:36 -0600 schrieb Michael A. Litscher:
>
>   
>> The short version:
>>
>> The issue I am having is that if I display an RGB value on the screen
>> and measure it's Lab value, I should be able to feed that Lab value into
>> LCMS, along with the monitor's profile, and get the original RGB value,
>> or something very close to it visually.  However, the RGB value I get
>> back from LCMS is off, significantly, and lighter in almost every case.
>>     
>
> Does you proper scaling for luminance (lumi tag ~CIE*L 116)?
>  
>   
>> The input profile I'm using is the built-in Lab4 profile.  The output
>> profile I'm using is created from the monitor's profile file.  The
>> transform I am doing is from Lab double to RGB 8, Absolute Colorimetrix
>> Intent, with a No Precalc flag.
>>     
>
> Seems ok.
> The profile looks internally ok. ICC Examin reports a DE of ~1.3.
>
> The spreadsheet shows not too much, as it has only results not the math.
>
>
> regards
> Kai-Uwe Behrmann
> --
> developing for colour management 
> www.behrmann.name + www.oyranos.org
>
>   

-------------------------------------------------------------------------
SF.Net email is sponsored by: 
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Lcms-user mailing list
Lcms-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

Reply via email to