Home Products Download Order Contacts


Subject: Re: ACR Calibration


>> The target values specified by the script do not match the values specified
in the Lindbloom (ProPhoto) chart I downloaded.

You are correct in stating that there are some differences in the target values. While I have confidence in my numbers, it would be brash to opine that one was right and another was wrong. So I will only explain the rational behind my numbers.

First, I found different credible sources listing different target values. Then, I found differences in Bruce Frazer’s numbers in different revisions of his books. Then I discovered that Bruce Lindbloom’s CC calculator gave different Lab values at different color temperatures, but the RGB values never changed. This just doesn’t compute with me. Gretag gives the numbers in xyY at illuminant C. Roy Berns gives XYZ values at illuminant D50. Lots of different values to chose from. Which one is right? For what conditions?

With a little more research I discovered that from the same Lab values, Adobe produced different ppRGB values than Bruce Lindbloom’s color calculator. It turns out that this is logical since the Adobe color engine is different than Bruce’s. And it will be influenced by user preferences such as conversion intent, black point, and such. Thus, I decided to go with target values in Lab mode and use Adobe to convert them to the desired RGB color spaces. These should match what (calibrated) ACR would do, in theory at least.

The white point (illuminant) affects the XYZ to Lab conversion, so I applied this as appropriate. I validated my math with Bruce Lindbloom’s. I selected Lab target values for four illuminants from different sources and generated corresponding targets in each of the four ACR RGB color spaces. This is the source of my target values. These same tables are used in ReadMacbethPatches. I validated the Lab D65 values against Bruce Lindbloom’s but the ppRGB values are as you state, different. I am using the Adobe generated values.

Tom Fors uses one set of ppRGB values for the target. He computes XYZ, Lab and such on the fly for some of the difference comparisons. He does not use the Adobe engine. His results are good for some cameras, bad for some others, in my experience.

It sounds like you have done a pretty good job of calibration. Congratulations. If you wanted to use different target values, it should be a simple matter to update the appropriate tables in both or either of my scripts.

As others have said, these calibration tools are nor precisely color science. Close is the best you can get. How close is the real question.

Due to these and other discrepancies, I have been doing some serious research the past several months. It is unfinished at this time. I find some serious differences in the basic color temperature chromacity calculations (using CIE CCT or Kelvin) and various tables. They vary by as much as 5% depending on who you believe. This is simply white balance, a starting point, before any color matching. So far, the color-matching dragon has me pinned under a rock.

It is no surprise that we get different numbers from different tools.

I firmly believe we need the spectral response data from the sensor manufacturers for accurate color matching. Until this is generally available, we will be constrained to somewhat limited calibration tools for raw data.

For the record, Mac versus Windows does not affect the target values at all.

Cheers, Rags :-)


View All Messages in adobe.photoshop.camera.raw

ACR Calibration =>

Re: ACR Calibration

Copyright © 2006 WatermarkFactory.com. All Rights Reserved.