[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

re: How do you refine Quadtone curves? long

> I made a bunch of assumptions about how each shade of ink should cover the
> image...

Under the circumstances you have to make those assumptions. But in
reality, any quad set I've worked with has not actually been 25%, 50%,
and 75% on paper, usually much darker. You are lacking some very useful information-
Actual density of each ink at 100% and it's dot gain characteristics on
a given media.
Actual density of the inks in combinations -MY, CY, CM, CMY (through the
Epson driver this becomes K).
The limits of each ink and in combination on a given media before
bleeding or mottling occurs.
All of the above will change with any driver setting change.

The density of two inks together at 100% each isn't necessarily the sum
of their individual totals. Some quad inks actually have negative dot
gain, so when you think you are bringing one in linearly, actually very
little may be happening on paper! For example, with Piezo ink, if you've
brought the lightest ink in up to 100% and begin bringing in the next
ink very little happens because of it's dot gain characteristics. A very
steep curve on that ink is required, and if you're backing the lightest
ink out underneath, really steep!

, for example the 25% ink should reach max density at around 25% of the
> light levels and it should start right at the 0% tone level; other inks are
> excluded from the lighter parts of the scale and build differently. So, I
> pushed the curve around until I got something like that with a bit of
> tapering where it terminated in the next tone.

You're on the right track, but you need the real number instead of 25%,
as above.
> The problems that turned up were initially fairly easy to correct, but the
> closer I got to a decent scale the more subtle the errors became and the
> more difficult they were to correct.

All of the problems you mention are fixable, but it's not easy. You're
stuck with trial and error without more info. Even armed with this
information some final trial and error is required.

> Several posters and respondents have said that it is important to keep the
> lighter inks from building up under the darker ones, but every time I try to
> bring my curves down in 25 and 50% inks the scale gets screwed up really
> badly, so I'm currently leaving the 25% ink, for example, at full value
> after it reaches maximum at around 1/4 of the way from the light end of the
> scale. If you bring the ink levels down, how do you do it, an overall
> adjustment curve? Some method not available in PS4, like "remove
> undercolors", whatever that is?

I assume you're working with the Epson driver, which will deal with
these problems for you. You just have to find curves that work with what
it does. Stay in RGB, bring in your lightest ink, and the next lightest
ink to your satisfaction leaving them at 100% each for now. Leave the
3/4 ink and K out of the picture for now by having the RGB channel that
affects that ink at 255 all the way across the curve. Get this right
first, print with RGB color in the space box and no color adjustment in
the driver. This way there is no further messing with your inks by color
management or the driver beyond whatever limits the media setting is doing.
Now you can start bringing in the 3/4 ink. All will be fine for a bit,
but at a certain point the driver will start to replace CMY with K, so
the darker parts of your 3/4 tone curve will really be affecting all
four inks, this will be the trickiest part of the curves to perfect. The
driver is now removing the lighter inks all on it's own, and it might be
best to start with letting it do the work for you.
You can certainly back the lighter inks out yourself by pulling the
endpoint back down (or up, depending on how you do your curves), there
is a lot to be said for this but it is very tricky. If you pull your
lightest ink out, the other inks have to be changed to make up for that,
these tend to be the finickiest corrections. And you still have to bring
them all back to full to make sure you have 100%K at the end.
The driver tends to limit total ink and take care of UCR for you that
works fine for many inks and papers. So it might be best to let it take
care of that for you rather than making things far more difficult by
working with reverse curves. Also, there is something to be said for
having a lot of dots packed in for good smooth tones, a "ceiling" of
100% light ink for the next ink to start piling on.
> Do you find it easier to start in the lightest areas and work to dark or
> vice versa? How about starting in the middle tones? Is one way better
> beacuse problems may tend to get pushed to the opposite end of the scale to
> be corrected more easily? Are the dark tones more of a problem than the
> light ones with excessive ink deposited?

See above, I think it's best to start with the lightest and work down.
Unless you see mottling on your paper, don't even worry about excessive
ink. If you see mottling with only your first 2 inks down, you have to
limit them to less than 100% each before even worrying about the next
inks. In darker parts of the curves the driver will start to limit for
you. Unless you have mottling, don't even worry about it, dumping as
much ink down as possible with closely packed dots helps make beautiful prints.
I'd look over Dan's info on RGB quad printing on Lincoln's site here-
If there is any way you can get your hands on a decent preview profile,
you'd be miles ahead.
There are many approaches to all this, and I'm talking out of my butt as
much as the next guy, but if none of this works for you perhaps it will
bring some other ideas to mind. I hope others comment with what works
well for them too. If Dan reads this he can correct me, he's been
working with the RGB driver more than I.
Ain't quads grand?
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.

[Photo]     [Photo Printers]    [Yosemite Photos]     [Scanner Discussion]     [Gimp]     [Gimp Users]

Powered by Linux