QTR-Quadtone RIP

Colin,

...

>I copied your qidf file, gave it a new name, modified the data for 
>my ink positions and deleted the linearization data. 

That sounds correct.

>In order to 
>print the step wedge using this file I also copied your original 
>quad file and gave it the same name as the qidf file. 
>Was this correct?

I don't think so.  My original *.quad file would have the linearization data
reflected in it, I believe.  If you'd saved the *.qidf file with the
linearization data removed, it'd make a new quad file.  This new file is
what should have been used for linearization.  On the other hand, if you
saved the new qidf file, it probably overwrote the old quad file.

>I printed the unlinearized step wedge and, 

Maybe it was unlinearized -- maybe not if my original quad file was still
there and unaltered.


> after hair dryer blowing, had an Lmin of 15.92 (Dmax 1.75)

At least you've got a good dmax now.

>I inserted the scanned data and `relinearized'. 
>Lmin was 16.82 (Dmax 1.73)

Linearization should not affect the end points.  This is probably a random
difference.

If, however, you "reliniearized" but from a print produced by the quad file
that already had my linearized numbers reflected in it, that might be the
problem.


>The curve was not linear. You can see this from the linearized 
>data. 

The second column is clearly not linear.

>The first column takes the Lmin and Lmax values from the 
>second column which are the scanned L values and calculates equal 
>steps between these.

I simply open the spectro's text file output (I now use the
ColorVision/DataColor spectro, in part due to it's output format relative to
my old X-Rite) in Excel and graph it (Insert>Chart>Line).  A linearized QTR
output should be a nice straight line.

I use these Excel graphs a lot.  When I'm making a curve in either QTR or
Photoshop (usually for dropping into QTR's Create ICC-RGB), I work from
these graphs (printed) and a simple ruler line on the paper.  The visual
feedback is very easy to use in changing the curve numbers.  The closer the
curve is to "linear" (where QTR is used for printing) or the gray gamma 2.2
space output (for ICC printing), the more likely that curve will make a good
linearized output -- and for almost all papers with little alteration.  If
the curve's cross-overs are well done for one paper, they are almost always
close enough to linearize well for all papers.  With QTR it may look like
I've made all sorts of different curves -- not so.  They are almost always
the same basic curve re-linearized for the new paper (from the raw, curve,
however, with previous linearization data removed).  With QTR I also look at
the end point for optimizing dmax.  For ICCs I just accept the Epson driver
limit, and it's often very close to optimum (but not always -- one area
where the rip can sometimes show a clear advantage even with an inkset that
is compatible with the Epson driver).

> ... It looks as though the ink limit had been set to a high 
>negative value, ...

I don't think it's an ink limit issue.  You would not be getting those good
dmax (Lmin) numbers if it were too low, and there would be a posterization
if it were too high.

See if your modified qidf file overwrote my quad file numbers or if you
were, in effect, doing a double linearization.

Paul
www.PaulRoark.com