[argyllcms] Re: colour management and... photo lab
- From: Nikolay Pokhilchenko <nikolay_po@xxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 23 Jun 2011 11:07:52 +0400
Hi, Marek!
I can propose You the next strategy:
1. Find a labs with stable results (without mention why the prints looks).
2. Find a most contrast and most wide gamut lab from stable ones
(without visual comparison).
3. Build color profile for chosen best lab.
4. Do the conversion of images to printing from image space to Your custom lab
profile.
5. Enjoy.
1. At first stage You can print the same testchart, generated by Argyll CMS, in
different labs and several times (3..5 times) during prolonged period (a month,
for example). If You have spectro, You can measure the stability between
different chart prints from every tested lab. The chart should be exactly the
same. I recommend this command for generation:
targen -v -d2 -A1 -c"sRGB.icc" -G MiniLabTest
You should choose appropriate patch number for Your preferred paper format. It
can take one or two paper sheets and consist of 250..1000 patches.
2. At second stage You can use the chart measurements from first stage to build
preliminary profiles for every lab. You can average several chart from one lab.
After profile building You can check color gamut and contrast of lab.
3. Generate the final target for Your best lab, using preliminary lab profile
from second stage, for example:
targen -v -d2 -A0.8 -c"MiniLabTest_xx.icc" -f 1500 -G FinalMinilab
The patch number I recommend is between 800 and 2000 for classic silver process.
High number of patches can be useful because the lab do the LUT color
compression (the most labs do).
For redundancy, You can print target several times with several day pause.
After measuring and combining of several charts data into one final. You
shouldn't use averaging, just add data from all targets, include step 2, into
one final.
Build the MiniLab profile.
4. Convert Your photos to print in Minilab profile. If You want ultimate
results, You can use collink and build links with image-depended optimal gamut
mapping. It's not in the scope of my present recommendations.
5. Enjoy.
I hope You'll be happy!
Other related posts:
