QTR-Quadtone RIP

Thanks for checking this out.
I'd much prefer the consistency that this provides.  I'd agree that the
"Simulate Ink Black"
while a nice option is not nearly so important.  I would not be concerned
about the
spikes -- variations in how to treat 8 bit rounding is well within a
reasonable result.

It appears that an extra "round-trip" through the profile is being done.
 This is usually
done for soft-proofing.  Hopefully this is not causing additional problems
but my guess
is that it is OK.

Roy

On Jan 1, 2008 2:14 AM, ferdinand_paris <ferdinand_paris@...> wrote:

> To prove that I've gone compulsive-obsessive, here are results with
> the modified QTR-Create-ICC-BPC.exe using my standard test image:
>
>   http://www.ferdinand-paris.com/BPC-Qimage_V4.png
>
> It confirms what Roy says - that profiles created with it are
> well-behaved in Qimage with or without BPC. Well, well behaved up to a
> point - the Qimage CM still produces spikes in the darker end of the
> histogram, but I don't think that's the fault of the profile.
>
> @Roy:  I have been thinking about what I said in my last post - that I
> don't find the softproof with my custom profile (the one made with the
> normal version of QTR-Create-ICC) particularly reliable.  I did some
> tests.
>
> I opened in PS an output file from Qimage that shows the big dMax loss
> after conversion to a "normal" custom profile.  I also opened the
> original sRGB file in PS and soft-proofed using the the same custom
> profile.  I turned off simulate paper colour for this comparison.  As
> I was using perceptual, BPC did not make an impact.  If I turned on
> "Simulate Black Ink" then the two images appeared to be the same on
> the screen.  Thus it appears to me that Qimage is reading the
> information required to simulate the ink black in a soft proof as a
> BPC to be applied in conversion.
>
> Of course the same test can't be done with new custom profiles made
> with QTR-Create-ICC-BPC.exe, because you've turned "Simulate Black
> Ink" off, and so Qimage is not confused by it.  I assume that's why
> you turned it off.
>
> I tried this test also with a 21 step wedge with the same results.
> The output from Qimage (perceptual + BPC conversion to a "normal"
> custom profile) looks the same on screen as the original step wedge
> soft-proofed from 20% grey into my "normal" custom profile with
> "Simulate Black Ink".  I.e.  it appears that Qimage is again applying
> the soft-proof black point simulation when converting with BPC.  Black
> 0 becomes luminance 26.
>
> Getting back to where I started, I don't find this soft-proof all that
> convincing.  It's convincing in terms of what Qimage produces with a
> "normal" custom profile + BPC, i.e. of a large dMax loss.  But not of
> what I get if I print from PS, nor if I convert in PS and print from
> Qimage without CM.
>
> In summary, the "Simulate Black Ink" soft-proof adjustment seems too
> large and seems to be misinterpreted by Qimage if it's in the profile.
>
> I hope this helps and makes sense, and doesn't demonstrate too many
> fundamental CM misunderstandings.
>
> F_P
>
>
> --- In QuadtoneRIP@yahoogroups.com, "Roy Harrington" <roy@...> wrote:
> >
> > I've got a new version of Create-ICC that might help in this.
> >
> > Download:    http://www.quadtonerip.com/QTR-Create-ICC-BPC.exe
> >
> > This maps black to black on both directions of the profile.  What
> > this gives up is soft-proofing with Simulate Ink Black.  The benefit
> > though is my tests with Photoshop show identical results for all
> > combinations of Perceptual & Relative Intent and BPB on or off and
> > with different CMSs on the Mac.
> >
> > I'm interested in any findings you guys get.   I highly recommend
> > using a 21step wedge to do this.  Cut out just the steps so you have
> > a perfect comb in the histogram.
> >
> > Then you can see if the results are the same.  (+/- 1 level is OK).
> > With PS I used 16bit and the histo-results were 100% identical.
> > I'd also go with a simple made up custom profile such as "20 96"
> > which shows significant movement in the histo.
> >
> > Roy
>
>
>


[Non-text portions of this message have been removed]