Re: Colorchecker SG and Scanner Profiling
Re: Colorchecker SG and Scanner Profiling
- Subject: Re: Colorchecker SG and Scanner Profiling
- From: Joseph Yates <email@hidden>
- Date: Wed, 18 Aug 2010 10:42:48 -1000
- Thread-topic: Colorchecker SG and Scanner Profiling
Since the originals we scan are on non-photographic medium, this was my
primary reason for wanting to use the ColorChecker SG - rather than the
IT8.7/2 target - to profile the scanner.
Robin, while we're on the subject - might I inquire about the RGB values you
state for the ColorChecker SG in your excellent "Repro Curve Guide V3"?
I've measured the CCSG chart with a Eye-One Pro spectro and converted the
L*a*b* values using Bruce Lindbloom's CIE Color Calculator - with BruceRGB
(2.2 Gamma, D65).
I also tried all the other RGB modes with a 2.2 gamma and most gave nearly
the exact RGB values as BruceRGB.
For half of the patches my RGB are exactly spot on with yours - and here I'm
talking about the Repro Curve 2.2 values for the CCSG - and several are only
1 or 2 RGB values way from yours. However there are several that are 6-7
values away - and I'm curious why?
Is it my device (Eye-One versus the spectro you used)? Or the method you've
used to calculate the Repro Curve 2.2 values for the SG target? Or a
variance in the tolerance of the SG target production? Or???
I also found LIBRAW's L*a*b* TIFF (which they state was measured with a
Spectrolina) - and again many of their 2.2 RGB values are spot on with yours
- and mine - or only 1 value away.
But patch E6 (Outer Black) is the one patch that varies the most - on LIBRAW
E6 = 28, Repro Curve = 32 and mine = 35.
Wondering if you might shed some light on this?
Regards,
Joseph Yates | Pacifica Island Art | Maui, Hawaii
> When scanning items other than photo materials, depending on the scanner
> hardware and the methods used in making the profile, the ColorChecker SG can
> make a better profile than the IT8.7/2 target. The reason is that the IT8
> chart is comprised of only 3 colorants, where the ColorChecker SG (96 unique
> patches) has, by some estimates, since X-Rite does not state, more than 10
> colorants. This results in better spectral sampling if your originals are not
> photo prints.
>
> If you are scanning only photo prints and transparencies then using the
> appropriate IT8.7/2 print is better. Appropriate here refers to Agfa IT8 for
> Agfa prints, Kodak IT8 for Kodak prints, etc. After all, the IT8.7 charts were
> specifically designed for scanning photo prints and transparencies, but they
> were never purposed for scanning anything else due to the limitations of 3
> colorant targets.
>
> Robin Myers
>
>
> On Aug 17, 2010, at 12:10 PM, Marco Ugolini wrote:
>
>>
>> The Digital ColorChecker SG has 140 patches (many of them repeats of white,
>> middle gray and black, hence with an even lower number of effectively unique
>> colors); the IT8 target has 288 patches, each of them unique.
>>
>> What is the rationale behind using a considerably smaller set of patches to
>> create a (supposedly) better scanner profile?
>>
>> Marco
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden