[Scribus] The colors from the print are not expected.

[Craig Ringer]

Adolfo Jos? S?ez Folgado wrote:

> I must bear in mind that different inks and papers give different
> results, but how and when?

"it varies". It's rather hard to predict.

The only way you'll find out for sure is to arrange "real" dummy proofs 
with your printer, where they actually make the plates, run up the 
press, etc. This is, unsurprisingly, expensive, though not as costly as 
having to re-print the entire run. It is really the only way, even now, 
to be truly sure that your job will have the colour you wanted it to.

If you can't get dummies, find out if your printer offers any sort of 
soft-proofing. Some printers can do quite reasonable simulations off 
high end laser printers, where they run the job through the same RIP 
they run the press off. Others might offer on-screen proofing in an 
environment with controlled lighting and a very high end monitor. As far 
as I know neither of these options can do a complete job of accounting 
for different inks, media, etc, unless the press has been measured with 
that particular combination. If the press has been measured with the 
media you'll be using, etc, then the results should be very similar to 
what'll print, and with a good soft proofing setup should be pretty 
trustworthy.

If your printer can't offer any sort of soft proof for you, you can try 
doing some yourself. You must have a good ICC profile for your press or 
you are wasting your time with this. Even with a good press profile I 
would not rely on the results personally. This is rarely as easy as 
setting a proof mode on a printer. Maybe there are printers out there 
that do a decent job of simulating a press out of the box - but I've not 
seen one yet. Someone PLEASE yell at me if they've had more success, 
since I'd love to be shown to be wrong. Anyway, as far as I know to get 
results you can rely on you'd probably have to do some proper 
calibration on your printer - requiring somewhat expensive hardware to 
measure its output.* Assuming you *do* have an accurate profile for your 
printer and for your press it's possible to simulate the behaviour of 
the press on the printer (with some limitations if the printer's gamut 
is more limited than the press's). The `tifficc' utility can actually do 
this to an extent, and there are lots of commercial soft proofing 
software options too.

(A soft proof printing option would be an awesome Scribus feature, 
though something it'd be nicer to see the print system handle).

Of course, if your press profile is good, your display profile is good, 
and your software is doing its job, in theory your on screen results 
should look similar to what'll print anyway. This will only be true if 
your software is set up for colour management and you have told it to 
simulate the printer on the screen.

Since right now it seems that you are not actually using a profile that 
is specific to your press, that is probably the first thing you should 
fix. Get an ICC profile from your printer that describes their 
particular press (and preferably media). They might tell you that their 
RIP expects Euroscale Coated (or uncoated, if you're printing uncoated) 
input, in which case you're OK with a generic profile since their RIP 
will compensate. You really do need to check that, though, since some 
printers expect you to use a generic profile and compensate, others want 
you to use their profile, and others again just won't know what you're 
on about with all this "icc colour" stuff.

If your printer is the last type, find a better printer ;-)

> Is the preparation of the PDF? Does the
> color profiles? Is asking the printer to calibrate her printer for my
> work?

I'm not sure I quite understand what you are asking here.

> I can not accept with a red selenium lamp leaves printed in the
> catalog as dark orange

If I'm not mistaken, you will have used a very intense, saturated red. 
What kind of press are you going to be printing on? Many print 
configurations won't handle some really intense reds and some other 
colours. The use of such colours is called using "out of gamut" colours 
- colours that the press cannot reproduce. Since your screen can 
probably handle them just fine, they look OK on screen but won't print 
properly.

The problem you describe would be consistent with colour clipping / 
shifting caused by the use of an out-of-gamut colour.

Have you used the "press simulation" mode with gamut warning enable to 
check if the colour management engine thinks the problem colours are out 
of gamut for your press? It's possible to check for such problem colours 
(within limits) and warn about them. If you're NOT using press 
simulation mode then you can use colours on screen that your press won't 
be able to handle and you won't have any idea you've done it.

Scribus can do on-screen press simulation and gamut warning. See the 
colour management preferences. Photoshop can also do it for particular 
images, which is useful when you're working on adjusting an image to 
print better. I think Krita 2 (coming in KDE 4) will do this too - hooray!

If the colour is out of gamut for your press, you have a couple of options:

- Make the whole image paler / duller / less saturated to bring the out 
of gamut colours into gamut. Playing around with the image in Photoshop 
while press simulation and gamut warning mode is on is quite handy if 
you're trying to do this. This is probably your best option, and it's 
the only one I know of that won't cost you a bunch of extra money 
(except ignoring the problem). At the newspaper I work for we either 
tweak the image to print OK or we find a more suitable image.

- Spend more on a more expensive print run with better stock, fancier 
presses, etc that can handle the difficult colour. For example, if 
you're using cheap uncoated printing with coarse stock, you might find 
that a coated press with better stock gets you an improved gamut that 
can handle the difficult colour. Is it worth it?

- If you're trying to match a very specific colour sometimes your 
printer can add it as an extra ink, you might print with (eg) CMYK + Red 
. This is called spot colour printing. It won't help you if you're 
trying to improve colour in photos etc, and is mostly useful for logos.

- There are very special print processes that use more than four colours 
for all colours - like HexaChrome for example. They're generally really 
rather expensive but I mention them for completeness. If you used a 
printer that used one of these processes you'd probably send them tagged 
RGB PDF, not CMYK.

> or a single PDF file that is so different from
> one machine to another, whether it is also not the fault of my work or
> the printer.

A pdf almost always obviously appear differently on different computers 
with different screens, OS versions, etc. This is because different 
screens, video cards, etc display the "same" colour totally differently. 
The computer cannot see the colour on the screen so it has no way to 
compensate.

The only time there should be any exception to this is when the computer 
you are viewing the PDF on has a properly calibrated display and you've 
enabled colour management in Acrobat. To calibrate a display you really 
must have a hardware calibration device. By eye calibration really 
doesn't do the job. If you have two computers with properly calibrated 
displays the same PDF should look very similar when viewed on both, so 
long as the colour settings in Acrobat are the same.

If you're trying to use Acrobat to preview the PDF, make sure to tell 
Acrobat to use the ICC profile for your press as its CMYK profile. 
Otherwise you'll be getting wrong results since Acrobat will be assuming 
untagged CMYK means something other than you intended.

Complicated, isn't it?

--
Craig Ringer


*  You might get away with a good scanner and a quality IT8 reference to 
calibrate the scanner then build your printer profile from that, but (a) 
there are issues where the printer & scanner gamuts don't overlap, and 
(b) it's less accurate.