Hi Peter, Ok, let's outline how this stuff works.
The profilers task is to compute the correspondence between physical colors and the digital values required by a device for rendering such colors. This correspondence is stored in a single file, we call it "a ICC profile". For computing such correspondence, the profiler needs to know the physical color and the device values. You have a card with a set of patches (the target) with known colors. Since you need to somehow tell the computer which are these colors, you need a way to measure them. So, this is the function of XYZ and Lab spaces. They are device independent spaces. They are not tied to any device, nor to computers in a more wide sense. If you measure a color by its XYZ value, then you have a reference to this absolute color. By convention, colors are measured under D50 illuminant. This avoid metamerism as well. Ok so far, now you want to profile, say, a monitor. The monitor is driven by the computer graphic card, which render colors by using RGB triplets. But since there are not two identical monitors, same RGB values can result in very different physical colors, so, two different monitors will give two different physical colors with same RGB values . A target, as said, is just a bunch of know physical colors, so, you cannot represent them by its RGB values, because this will tie the target to a specific device. But you can measure the colors using XYZ or Lab and then the colors will be described, since Lab or XYZ are device independent. Now to your questions. >Then why do none of the values in the files in targets/Kodak/Emissive/ >have values greater than about 85? All targets _reference_ comes with only Lab/XYZ measurement because the reason explained above. XYZ range goes from 0 to about 100 and Lab goes 0..100 for L and -128..+128 for a, b. > And further, can I create a reference file for my target using RGB? As said, not a reference but a measurement. The profiler will then create a profiler by comparing the RGB values and the physical (Lab) color obtained. You can just use Lab if wish so, the profiler only needs that. > Finally, is it necessary to have the standard deviation columns and the > mean and standard deviation dE values? No. The profiler will use them if available, but only for statistical purposes. Hope this helps, Marti Maria The little cms project http://www.littlecms.com [EMAIL PROTECTED] ----- Original Message ----- From: "Peter Gregson" <[EMAIL PROTECTED]> To: "Marti Maria" <[EMAIL PROTECTED]> Sent: Saturday, July 26, 2003 1:12 PM Subject: Re: [Lcms-user] Re: making a target reference Marti, Then why do none of the values in the files in targets/Kodak/Emissive/ have values greater than about 85? And further, can I create a reference file for my target using RGB? I notice that the other references are X,Y,Z and Lab. Finally, is it necessary to have the standard deviation columns and the mean and standard deviation dE values? Regards, Peter Marti Maria wrote: >Hi, > >The range of values for lprof is 0..255 too. No scaling or other transformation >required. > >Regards, >Marti Maria >The little cms project >http://www.littlecms.com >[EMAIL PROTECTED] > >----- Original Message ----- >From: "Peter Gregson" <[EMAIL PROTECTED]> >To: <[EMAIL PROTECTED]> >Sent: Saturday, July 26, 2003 3:55 AM >Subject: [Lcms-user] Re: making a target reference > > >Further to my previous message, the DryCreek data are RGB values from 0 >- 255 but I note that the values used in the reference files in the >lprof package are from zero to about 85. Is it a simple scaling, and if >not, what is the transformation to go from RGB to the required data? > >Regards, > >Peter Gregson > >\ > > > > >------------------------------------------------------- >This SF.Net email sponsored by: Free pre-built ASP.NET sites including >Data Reports, E-commerce, Portals, and Forums are available now. >Download today and enter to win an XBOX or Visual Studio .NET. >http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 >_______________________________________________ >Lcms-user mailing list >[EMAIL PROTECTED] >https://lists.sourceforge.net/lists/listinfo/lcms-user > > > > -- Peter H. Gregson, Ph.D., P.Eng. NSERC Chair in Design Innovation Director, iDLab Professor tel: (902) 494-2175 fax: (902) 494-2096 e-mail: [EMAIL PROTECTED] web: www.iDLab.dal.ca ------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 _______________________________________________ Lcms-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/lcms-user
