Re: LAB values
Re: LAB values
- Subject: Re: LAB values
- From: bruce fraser <email@hidden>
- Date: Sat, 2 Aug 2003 11:35:11 -0700
At 12:03 PM +0200 8/2/03, Peter Baumbach wrote:
>
>
Oh, I indeed forgot about the rendering intent setting in Photoshop (RWCM p.
>
335). Now the LAB values are the same.
I had misunderstood the question-glad Chris caught it.
>
I understand that the capture profile can almost never be perfect. I have
>
got an average delta-e of 1.22, on my Howtek 4500, gamma=2.5, Silverfast
>
software with calibration. The best 90% average delta-e is 0.64, the maximum
>
deviation is 15.56. That value seems a bit large. Any idea ?
Several ideas.
First, bear in mind that Lab isn't really perceptually uniform. One
of the nice things about the comparison plots in ColorLab or
MeaureTool is that they show you (within the limits of your monitor)
the actual color difference as well as the numbers, so look at how
visually significant that 15.56 E actually is. (It would be much
more significant in blues than in yellows, for example.
Second, if it's a dark color, you may want to try a higher gamma on
the scanner. I usually find that the sweet spot is somewhere between
gamma 2.7 and 3.1.
Third, you can, if you feel the need to, reduce that maximum error by
doing selective color edits on the profile, but in so doing, you will
raise the average of the best 90%. It's like adjusting a cat's
cradle-if you pull somewhere, it shifts somewhere else. Since right
now it's at 0.64, I'd say you have room to move-I'd try to keep the
best 90% average below 1. I've run into situations where I was able
to reduce a double-digit maximum E to around 4-5, while keeping the
best 90% in the 0.8-0.9 range.
Bruce
--
email@hidden
_______________________________________________
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.