Re: I Never Met A Profile I Liked!
Re: I Never Met A Profile I Liked!
- Subject: Re: I Never Met A Profile I Liked!
- From: Andrew Rodney <email@hidden>
- Date: Thu, 30 Aug 2001 14:36:44 -0600
on 8/30/01 1:16 PM, Robert M Eversole at email@hidden wrote:
>
> who's software writes the most accurate profile
>
>
I would say ProfileMaker and MonacoProfile (and subsequently CompassProfile
>
Pro) generate an equally accurate profile.
Accurate needs to be defined. Best gray balance? Best tonal range? Most
accurate match of an original (and then how was the object captured and how
does that affect the process)?
CompassProfile Pro output to a fairly wide gamut device (A Fuji
Pictrography 4000 looks a lot different from ProfileMaker Pro but which is
more accurate? I did this test about a year ago with a group of
photographers. We had a photo of a MacBeth (standard) Color Checker from a
profiled digital camera. I profiled the printer using both packages and
placed the resulting print under a D50 light box. Nearly all the
photographers preferred the print from CompassProfile. Then I placed the
actual Macbeth target next to each print. Well the profile from ProfileMaker
Pro was almost a dead match. So why did everyone prefer the print from
CompassProfile? Because it was more saturated and a bit more snappy. One
print matched and the other that didn't was preferred. Now I could have
gone into virtually ANY profile editing application and either bumped the
saturation of the ProfileMaker Pro a tad or dialed down the saturation of
the CompassProfile profile and ended up with about the same results in
print.
One exercise I'm working on for the Photo Plus Expo show in New York in
November is a profile package shoot-out in the Digital Learning Center on
the show floor. So far Monaco, GretagMacbeth and it looks like ITEC
(ColorBlind) will be there. Some of the issues that need to be considered is
not only how a well a standard test image appears but ease of use,
documentation and so on.
What will be interesting is seeing how the various packages produce output
(we will likely have both an RGB and CMYK output device to test) from the
same source image. I'm working on getting some other companies signed up to
participle in this shoot out. Anyone who has some good ideas about how to
set up some standards for testing, I'm all ears. We would of course use the
same measuring instrument (likely a Spectrolino/scan) so that there is a
level playing field there. I'm thinking that with packages that output
multiple test patches, we'd let each vendor decide how many they want to
measure (and account for that on the print). So if one package can make a
really great profile with 288 patches and the other package needs to measure
10000, that tells a story.