Re: Delta E calculator
Re: Delta E calculator
- Subject: Re: Delta E calculator
- From: Robin Myers <email@hidden>
- Date: Tue, 01 Jul 2003 22:11:39 -0700
- Organization: Robin Myers Imaging
tom vanderlinden wrote:
>
>
good night - - -
>
>
On Tuesday, July 1, 2003, at 07:28 PM, Steve Upton wrote:
>
>
>
>
> Beware "simple" Lab calculations. Unless you use delta-E 94 or CMC you
>
> will find that delta-E number values do not correspond with actual
>
> perceived color differences.
This is not completely accurate. The delta-E 1994 and CMC equations are
recommended for small delta-Es only. They are inaccurate for large
delta-Es. The CIE 1976 delta-E equations are still recommended for large delta-Es.
>
>
>
> it's true that Lab was defined so that the simple 3D distance between
>
> any two points is a good method of comparison but in reality it turns
>
> out that the current version of Lab that we are all using is
>
> non-isotropic - that is it is warped, and warped more in some areas
>
> than in others. Lab, when compared with perceptually accurate systems
>
> like Munsell can introduce errors of 20% or more in certain areas.
CIE 1976 L*a*b*, commonly referred to as "Lab", was designed as a
representation for qualitative comparisons of color differences. It was
pressed into service as a PCS by the ICC, but it was not designed for
the job. Once in use it's warts started to show, which are the
non-uniformities of perceptual mappings in the L*a*b* space. The main
reasons for this usage are that there are equations for converting
between XYZ and L*a*b*. Color spaces derived from direct human
perceptual evaluations, such as Munsell, do not have any equations that
easily convert between their perceptual coordinates and XYZ. XYZ is
important because it is a direct correlate of the cone responses of the
human visual system and are easily derived from spectral measurements.
There are no programs available that allow one to convert between XYZ
and Munsell with user defined lighting (there is one program available
from GretagMacbeth but it is basically a table lookup system and I
believe it only works in CIE Illuminant C).
As a historical note, several "Lab" color spaces existed before the CIE
standardized the 1976 L*a*b* system. There are currently shipping
instruments that can report values in these earlier coordinates (e.g.
the Eye-One).
>
>
>
> so... beware.. if your tools do not use newer flavors of delta-E, they
>
> are prone to error.
Remember, the newer delta-E equations are only valid for small delta-E
values. I would not use them with delta-Es greater than 5.
Robin Myers
_______________________________________________
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.