[argyllcms] Re: Fakeread using ti3 and profile

Hi Martin,

Martin Weberg wrote:

Hi,

I'm creating IT8.7/4 data from ECI2000 data or ECI profile using
fakeread and I'm getting big differences using ti3-data and
ICC-profile.

logo2cgats FOGRA29L.txt FOGRA29.ti3

fakeread -l FOGRA29.ti3 IT8-7.4-chart-data

fakeread -l ISOuncoated.icc IT8-7.4-chart-profile

verify IT8-7.4-chart-data.ti3 IT8-7.4-chart-profile.ti3

Verify results:
Total errors: peak = 31.279488, avg = 11.027247
Worst 10% errors: peak = 31.279488, avg = 22.069879
Best  90% errors: peak = 20.484855, avg = 9.797758


I just made a quick cross validation:

cp argyll/ref/FograStrip.ti1 FograStrip-f29.ti1
fakeread -l F29.ti3 FograStrip-f29
cp argyll/ref/cp FograStrip.ti1 FograStrip-f29p.ti1
fakeread -l /Library/ColorSync/Profiles/ISOuncoated.icc FograStrip-f29p
verify FograStrip-f29.ti3 FograStrip-f29p.ti3

Verify results:
  Total errors:     peak = 0.945529, avg = 0.255433
  Worst 10% errors: peak = 0.945529, avg = 0.656237
  Best  90% errors: peak = 0.480497, avg = 0.206555



What could be wrong?


I guess FOGRA29L doesn't contain all of your IT8-7.4-chart-data patches.
(AFAIK FOGRA29L is based on ECI2002 and has 1485 while IT8-7.4 has 1617 patches.)
<http://argyllcms.com/doc/fakeread.html>: "If a .ti3 file is specified instead of an ICC or MPP profile, then the *closest matching measured points* in the ..ti3 are substituted for the test values in the .ti1 file on output" 

Klaus

Other related posts: