Datacolor User to User Support Group.

I am new to all this, so I am trying to find out how it all works.  I 
printed the 150 patch target thru Qimage with no color management 
either in Qimage nor the Canon i9900 driver.

The paper is Red River's River Linen, with OEM Canon inks.


Then I read the page in Printfix Pro, and saved the resulting profile.

Then I printed the target tiff again with Qimage using this new profile.

Then I read this new result page in Printfix Pro.  I expected to find 
that the target printed with the new profile would show very little 
need for correction - the two halves of each patch would look about the 
same after reading.

There does not seem to be a good way to compare two results sheets side 
by side, but as far as I could tell, the profile did not make much 
difference.  There were still marked differences between the measured 
and desired triangles on many of the patches.

Is there something wrong with my method here?  Or am I expecting too 
much?  Is it just that River Linen cannot produce all the colors called 
for by the target?

Bryan Stone

>I am new to all this, so I am trying to find out how it all works. I
>printed the 150 patch target thru Qimage with no color management
>either in Qimage nor the Canon i9900 driver.

First of all, use the 225 patch target. It's "better", to the point that
spending another minute reading the additional 75 patches is worthwhile.

>The paper is Red River's River Linen, with OEM Canon inks.
>
>Then I read the page in Printfix Pro, and saved the resulting profile.
>
>Then I printed the target tiff again with Qimage using this new profile.

The target tiff isn't a good test image. It represents colors spanning
all of RGB space; many of them are out-of-gamut.

>Then I read this new result page in Printfix Pro. I expected to find
>that the target printed with the new profile would show very little
>need for correction - the two halves of each patch would look about the
>same after reading.

I think you're probably the first person who's ever tried doing this.
It's not a bad concept, but no, this isn't how to go about testing the
profile. The "pure" colors in the upper left of the patch squares in the
Target window view aren't color managed; we show them only as a reference,
so that you can see a "raw" version of the RGB patch that you're reading;
you still won't get reflective readings from properly profiled/printed color
patches that will match these perfectly. (This would be the case with any
profiling software on the market, including packages that cost thousands of
dollars).

>
>There does not seem to be a good way to compare two results sheets side
>by side, but as far as I could tell, the profile did not make much
>difference. There were still marked differences between the measured
>and desired triangles on many of the patches.

Right; this isn't a good way to test the profile.

>Is there something wrong with my method here? Or am I expecting too
>much? Is it just that River Linen cannot produce all the colors called
>for by the target?

Yes, among other things.

You want to test your profile by printing a "real" image through it; not
the 225 color squares.

Print PFP's "PDI" test image. (A standalone copy is in the Test images
subfolder). Print it through QImage, using the same output color management
settings you had in effect when printing the target; and do whatever else
you need to do, in QImage, to use the profile that you built. I'm not
familiar enough with QImage to tell you how to configure this.

You should get a "good" print. If QImage lets you softproof through the profile,
as well as print through it, that's one way to compare: the print should be a
very close match to the softproof. You can also softproof in PFP immediately
after building the profile; or in Photoshop (View:Proof Setup:Custom) if you
have that installed.

-- 
David Miller
Senior Software Developer, Digital Color Solutions
ColorVision

David
Thank you for your thoughtful response to my questions about testing a 
profile.

For the less experienced amongst us, the idea of printing a test image 
and eyeballing it to see if we are getting optimum results causes some 
uneasiness, since we purchased the equipment to get away from 
eyeballing.  I realize that in the end it is the subective experience 
which counts.

I wonder, would it be possible to provide a profile testing image with 
readable patches matching the colors in the test image, so that one 
could evaluate a profile.  This would be especially useful in 
evaulating papers.

Best wishes,  Bryan Stone