Re: [Lprof-users] Unable to select 22-column Pick Template

[Hal V. E. <hv...@as...>]

On Friday 25 January 2008 06:56:00 Paul Waldo wrote:
> Hi all,
>
> I'm using lprof-1.11.4.1.  I select my IT8 target, which is a 22-column
> target from Wolf Faust.  I select "IT8.7/2 - 22 column picker" for my Pick
> template. I select OK, then Done.  When I got to select the corners on my
> image file, I only get 19 columns of green squares, not 22.
>
> Here is another interesting data point.  If I select the 22-col pick
> template first, then point to the Target directory, the OK button is greyed
> out.  I have to select another Template, before the OK button is clickable.
>
> I'm getting very poor results at 19 columns:
>
> I have chosen CIE XYZ as PCS
> Unknown white point (X:1e-314, Y:7e-312, Z:7e-312)
> Primaries (x-y): [Red: 0.625415, 0.277302] [Green: 0, 1] [Blue: 0.0833993,
> 0] Estimated gamma: [Red: 1.19788] [Green: 1.22472] [Blue: 1.24135]
> Loading sheets...
> Reference sheet: /home/paul/.lprof/target_refs/R060101.txt
> Measurement sheet: /home/paul/.lprof/temp/meaurement.cgt
> Gamut hull: 195 inside, 0 outside, 57 on boundaries
> Fitting error (delta E CIELAB): mean=1.97997, RMS=3.01769, 95%%=6.1929,
> median=1.09214, MAD=0.666138
> Fitting error (delta E CIE94): mean=1.10034, RMS=1.5087, 95%%=3.06707,
> median=0.777355, MAD=0.432199
> Performing 2-fold cross validation ...
> 2-fold CV: estimated RMS dE=4.04654, dE94=2.04384
> 2-fold adj. CV: estimated RMS dE=3.70133, dE94=1.86485
>
> Here are my profile parameters:
>
> Resolution (CLUT points):  33
> Profile verbosity: Store anything
> Regression:
>     Smoothness: 0.500
>     Deviation: 1.00e-05
> Linear Bradford
>
> I am hoping that getting the additional 3 columns will improve the dE,
> although I somehow doubt it...
>
> Any help would be appreciated!
>
> Paul

There is a bug in 1.11.4.x that in some cases results in the reference file 
being incorrectly installed.   Predicting when this might happen is difficult 
and it happens more often on Windows machines than on unix like machines.  
But it does occasionally happen on *nix machines.  You are the third person 
to report this problem on a Linux box.  This is fixed in CVS and this work 
was done as part of the Google Summer of Code project.  For 1.11.4.x there is 
a simple work around that only takes a few minutes.

To fix this you need to modify the lprof configuration file.  For *nix users 
this file is:

/home/<user id>/.lprof/config/lprofrc

It is a text file and to fix the problem you will need to edit the file and 
make a small change.  The file is divided into sections.  The section you are 
interested in is the [Files] section.  In that section you will find some 
lines that looks something like this:

TARGET0=/home/heng/.lprof/target_refs/R041110.txt
TEMPLATE0=/home/heng/.lprof/templates/IT8_22.ITX

These are the configuration items that pair up a target and the targets 
template.  The likely issue for you is that you either have extra /'s or a 
missing a / at the path/file name location in one of these entries (I forget 
off hand if it is an extra / or a missing /).  Just correct that so that the 
path/file names for both are correct and it should start working.

If you do not find the / issue please post back here and I will have a closer 
look.  But I am 95% sure that this is what you are seeing.  And if it works 
please let us know.

Hal