Path: utzoo!mnetor!uunet!husc6!hao!gatech!mcnc!ece-csc!ncrcae!ncr-sd!hp-sdd!hplabs!hpda!hpsal2!hpcupt1!hpfcdc!hpgrla!douglasg
From: douglasg@hpgrla.HP.COM (@Douglas Genetten)
Newsgroups: comp.graphics
Subject: Re: RGB to printer CMYK conversion
Message-ID: <4180010@hpgrla.HP.COM>
Date: 4 Feb 88 20:08:37 GMT
References: <10258@sgi.SGI.COM>
Organization: Hewlett-Packard, Greeley, CO
Lines: 26

Mr. Hue writes...

>...Then use
>a transmission-reflection densitometer and measure the densities of your
>17^3 RGB triplets and 17^4 CMYK quads... 

A common missunderstanding hides within this remark.  Densitomiters
DO NOT MEASURE NTSC RGB.  They typically measure using very narrow-
band filters which are designed to maximize sensitivity to variations
in process ink sets.  These filters block out most of the visible
spectrum---something you can't do if you want to measure color regardless
of the color space you use. These filters would have to have broad 
overlaping "standard observer" curves to make them useful for RGB 
measurement.

Using press room densities as estimates of NTSC RGB densities can cause
large objectionable color errors, especially when measuring strange 
pigments found in some color electronic printers.

A good alternative would be to use the Minolta Chroma Meter which is
a near-true "standard observer" colorimeter and measures using 
CIE XYZ and CIE LAB tristimuli.

Douglas Gennetten
Hewlett-Packard
Greeley Hardcopy Operation