colprof: Error - gamut: vector_isect failed! after several "nan"'s

[S.]

Hi,

I'm trying to create an icc profile for my Canon printer using a DTP20 device.
In order to do that I followed to instruction in the "Typical usage Scenarios
and Examples" section on the Argyll website.

The targets were obtained by issuing the following commands
* targen -v -G -d4 -l400 -f2160 -B16 -e16 -g128 Canon
* printtarg -v -i20 -pA4 Canon

and printing the created .ps file with all color correction options turned off.

The prints were left to dry for a couple of hours and were read in using
* chartread -v -c1 -h Canon
This took quite some effort because very often communication problems occured
with the DTP20 (seems to be some initialisation issue: either it works
perfectly or it will only give communication errors after connecting the
device) but finally I did succeed.

If I then try to create an icc profile from the results using
* colprof -v -qh -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp Canon

I get the following error:
colprof: Error - gamut: vector_isect failed!

Changing several options does not make any difference.
Taking a closer look at the output of colprof, at several locations "-nan"
occur. This starts (I think) when the black point is determined because in the
output the first occurence is in the line

Black point XYZ = -nan -nan -nan, Lab = -nan -nan -nan

I have no clue what I am doing wrong. Should I use different options?
Any help is greatly appreciated!

Sander

[Nikolay Pokhilchenko]

Post by S.
I'm trying to create an icc profile for my Canon printer using a DTP20 device.
In order to do that I followed to instruction in the "Typical usage Scenarios
and Examples" section on the Argyll website.
The targets were obtained by issuing the following commands
* targen -v -G -d4 -l400 -f2160 -B16 -e16 -g128 Canon Which type of the printer do you have? Are you using a RIP? Which one? Is your printer capable to print separate C, M,Y and K channels? Or this is just an "RGB-driven" printer?
If I then try to create an icc profile from the results using
* colprof -v -qh -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp Canon

colprof -v Canon
?

Post by S.
Taking a closer look at the output of colprof, at several locations "-nan"
occur. This starts (I think) when the black point is determined because in the
output the first occurence is in the line
Black point XYZ = -nan -nan -nan, Lab = -nan -nan -nan It seems like you've tried to profile "RGB"-printer as CMYK printer.

[S.]

Post by Nikolay Pokhilchenko

Post by S.
I'm trying to create an icc profile for my Canon printer using a DTP20
device. In order to do that I followed to instruction in the "Typical
usage Scenarios and Examples" section on the Argyll website.
The targets were obtained by issuing the following commands
* targen -v -G -d4 -l400 -f2160 -B16 -e16 -g128 Canon Which type of the
printer do you have? Are you using a RIP? Which one? Is your printer
capable to print separate C, M,Y and K channels? Or this is just an
"RGB-driven" printer?
If I then try to create an icc profile from the results using
* colprof -v -qh -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp Canon
colprof: Error - gamut: vector_isect failed! Have you tried simpler command

colprof -v Canon
?

Post by S.
Taking a closer look at the output of colprof, at several locations "-nan"
occur. This starts (I think) when the black point is determined because in
the output the first occurence is in the line
Black point XYZ = -nan -nan -nan, Lab = -nan -nan -nan It seems like you've
tried to profile "RGB"-printer as CMYK printer.

Thanks for the response!
I'll first describe my systema bit more/better (sorry I did not do that in my
first mail)
My system consists of a PC running both Windows XP and Linux, a Canon Pixma
IP4300 printer, a Scanjet 8250 and a DSLR camera. Although the equipment for
sure is not the best available I'm attempting to get the best match between
colors on the various devices. To do that I'm using an i1Display2 for the
screen, an IT8.7/2 chart for the scanner and I'm trying to use an DTP20UV for
the printer.

I'm pretty new to the "color-stuff" and reading a lot to get to know more, but
even though I thought I read everything thoroughly your remarks made me review
and re-read everything again.
I think now your last remark was spot on, and I should have used "printerRGB"
in the targen line.
After doing that the errors disappeared. Thanks a lot!

I do have an extra question however: using Linux with CUPS it is possible to
set the color model to either RGB or CMYK. Is either RGB of CMYK to be
prefered for better results (maybe fewer conversions) or is this kind of trial
and error?

Regards,
Sander de Waal

[Graeme Gill]

Post by S.
This took quite some effort because very often communication problems occured
with the DTP20 (seems to be some initialisation issue: either it works
perfectly or it will only give communication errors after connecting the
device) but finally I did succeed.

Hi,

What sort of system are you using ? What sort of reports do you get
at the point of failure using "-D6" ?

Post by S.
If I then try to create an icc profile from the results using
* colprof -v -qh -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp Canon
colprof: Error - gamut: vector_isect failed!

The typical explanation in the past for this sort of error is
that the gamut of the device has collapsed in some way, and doesn't
have a reasonable volume. But without your .ti3 file I can't
really tell.

Create a profile without -S and view it using iccgamut. Does
it look normal ?

Graeme Gill.

[S.]

Post by Graeme Gill

Post by S.
This took quite some effort because very often communication problems
occured with the DTP20 (seems to be some initialisation issue: either it
works perfectly or it will only give communication errors after
connecting the device) but finally I did succeed.

Hi,
What sort of system are you using ? What sort of reports do you get
at the point of failure using "-D6" ?

Post by S.
If I then try to create an icc profile from the results using
* colprof -v -qh -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp Canon
colprof: Error - gamut: vector_isect failed!

The typical explanation in the past for this sort of error is
that the gamut of the device has collapsed in some way, and doesn't
have a reasonable volume. But without your .ti3 file I can't
really tell.
Create a profile without -S and view it using iccgamut. Does
it look normal ?
Graeme Gill.

Thanks for the response!
My system consists of a PC running both Windows XP and Linux, a Canon Pixma
IP4300 printer, a Scanjet 8250 and a DSLR camera. Although the equipment for
sure is not the best available I'm attempting to get the best match between
colors on the various devices. To do that I'm using an i1Display2 for the
screen, an IT8.7/2 chart for the scanner and I'm trying to use an DTP20UV for
the printer.

The main problem regarding the NaN's disappeared when I revisited all my
commands due to the remark of Nikolay Pokhilchenko (needed to use printerRGB
instead of CMYK).
Regarding the read-errors: for some unknown reasons the DTP20 started to
behave consistently correct after a while and I was not able to get any debug
output. Two things changed in the mean time: I connected an extra power
adapter to the device and I only connected the device when the system was
already up and running (so it was not connected while booting). Whether this
made a big difference I do not know, but at least things seem to work correctly
now.

Regards,
Sander de Waal