QTR-Quadtone RIP

On Dec 11, 2007 10:14 AM, Howard Shaw <glassman@...> wrote:

> Ernst Dinkla wrote:
>  >>> ...
> >>> the profile refinements became available so it is understandable.
> Adding
> >>> numbers for perceptually correct density ranges and calling them Ideal
> >>> curves would help to show the difference and add another control for
> >>> users. Is this correct or did I miss updates of the manual ?
> >>
> >> Here, I think you are suggesting that there be another table (or set of
> >> columns) that provide "perceptually correct density values". Do I
> understand
> >> this correctly? If so, what are perceptually correct density ranges and
> how
> >> are they determined?
> >
> > There's a chapter:
> >
> > Profiling the Advanced Black and White Mode
> >
> > on this page
> >
> > http://www.outbackphoto.com/printinginsights/pi045/essay.html
> >
> > that I thought sums up the linearisation stage above and the
> > next step the profiling on top of it.
> >
>
> I'm none the wiser. I can see the profiling step described on that page
> (for use with Epson's ABW mode) but not a table of "perceptually correct
> density values". Could anyone explain further please or, better still,
> provide the table?
>
> Howard
>

I don't think the notion of "perceptually correct density values" really
makes sense.

I agree with Ernst in that "ideal densities" is a misleading label -- they
were meant
as the densities for a QTR linearized output so "linearized densities" is a
better name.

The word "perceptual" is used in ICC profiles as a conversion intent.  I.e.
as a goal
in converting one set of data numbers into another such that perceptually
they are
as good as one can get.   The working spaces that we use such as gray gamma
2.2
or dot gain 25 have theoretical density ranges of 0 to infinity or
luminosity of 100 to 0.
When converted to output -- paper or screen - they must be compressed to the
range that the media can give.  So 0...infinity is compressed to
dMin...dMax, or
with luminosity units from 100...0 to say 96...16.   In QTR linearization
the compression
here is a simple linear conversion -- 100,95,90,... 10,5,0 becomes
96,92,88,...24,20,16.
Notice that what was a straightline before is a straightline afterward, just
a lower slope.
But if you look at this in density units they aren't straightlines so it's
much harder to
visualize.  So that's what the "ideal densities" chart is -- just
straightlines in luminosity
converted to density units.  The idea in QTR is to make all QTR curves
consistent
so they all are linearized to the same standard.

When we move up and look at color management and perceptual intent they are
a
lot more variables.  Each working space (embedded profile) has a different
mapping of
pixel numbers into luminosity.  They are virtually no straightlines.  Take a
stepwedge
and assign different profiles and you'll see that each one results in a
different set of
luminosities for the steps.  Color management with ICC profiles is the one
and only
tool to do conversions that map one luminosity into another.  The conversion
has to
be custom created based on both an input and an output ICC profile.   The
input comes
from the embedded profile and the output comes from the printer profile you
pick in
Print Preview or Convert-to-profile.   The QTR-Create-ICC tool is a way to
make a
print profile based on a particular output driver be it QTR or Epson ABW.

Perceptual intent is a conversion from one luminosity range to another that
does it's
best to give the same perceptual "feel" for an image.  Mathematically it
uses yet
another set of units called Y - luminance.

QTR linearization is special in that all curves meet this linearized
luminosity so it's
possible to make a general ICC profile that works on many setups -- this is
the
QTR Gray Matte Paper and QTR Gray Photo Paper print profiles.

Roy


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