Hi Davide,
That's a good question and worth explaining a bit.
The idea of multiple gray inks is using the principle of "partitioning"
where a light gray ink
is used in the light parts of an image and then a darker ink in the
darker areas.
The transition between these areas is the crucial part. Since there's
a limit of how much
ink a paper can absorb the transition areas are a switch over from the
light ink to the
dark ink. It's important that this switch over is smooth as far as
density on the paper is
concerned. These weightings or relative densities are used to make
this smooth transition.
As you've noticed, ink on paper in a "raw" mode is very non-linear.
But to get a smooth transition
what actually happens on the paper is the most important consideration.
So matching a
C100% with a K45% is looking at exactly the range where the transition
will happen.
Conceptually, when the lighter C ink gets to max 100% coverage we must
switch to the
darker K ink and what we're measuring is how much K ink should be used
right after the
transition. Note that the matching of the next ink, say M, is done
against the previous ink
not with K because the transition will be done on paper with C not with
K.
This also has the benefit as you say in (6) that the units make no
difference.
Whether you measure in Lab, density or a scanner in RGB you always get
the same result.
Matching could probably be done just by eye if you were careful.
Anyway the idea is to match what really happens on the paper with the
percents of ink.
The transitions are actually a lot more gradual and coverup any
variation of density
measurements, ink color, printer variations etc.
Roy
On Wednesday, November 23, 2005, at 05:01 AM, Davide Barranca wrote:
> Hi to you all,
>
> I'm kind of a newbie with quadtone rip; if I'll ask you questions
> someone already answered, please redirect me to the thread (I've made
> a little search but with no great results).
>
> Here is my setup: K6 quadblack inks from Piezography with EP7000
> printer, EyeOne Pro spectrophotometer; I'm trying to create a custom
> curve and linearization, following the pdf instructions.
> My doubts start as it suggest how to measure the relative ink
> densities. I did it his way, here is a little example:
>
> 1. I've measured the K100% (L*17,4)
> 2. I've measured the C100%, M100% and so on (actually they're not
> Cyan, Magenta etc. but gray shadows, I'll call them C,M,Y,lC,lM to be
> clearer). Let's say that C100% is L*23,9
> 3. I've found that the nearest patches in the K 21 steps are the 40%
> (L*26,0) and the 45% L*23,3)
> 4. Doing little math, I've used the Y=aX+b equation to find the
> interpolate % between 40% and 45% that gives me L*=C100%=23,9 and
> eventually I've found C100%=K44,78%.
> 5. I've done it for all the 6 inks.
> 6. I could have been using Density instead of L*, but numbers
> shouldn't change: D=((L+16)/116)^3
>
> And that's OK.
> But...
>
> Does it really make sense to look for the nearest printed K patch? I
> mean: in order to find the relative ink density, wouldn't be better to
> do not rely on the 0%,5%,10%... 95%,100% K scale (printed with no
> linearisation, no known gamma), and use only the 100% patches?
> I can do this:
> 1. Measure the K100% (L*17,4) and calculate Density= 1,63
> 2. Measure the C100% (L*23,9) and calculate Density= 1,39
> 3. Write the proportion 100:1,63=X:1,39 so X=85,3%
> That seems to me a much logical way to find relative density. And the
> value is so different.
>
> Maybe I'm wrong, I'll be glad to hear from you experienced users your
> comments and suggestions.
> Best regards,
>
> Davide
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
-
Roy Harrington
roy@...
Black & White Photo Gallery
http://www.harrington.com