Re: ink laydown order - more discussion please
Re: ink laydown order - more discussion please
- Subject: Re: ink laydown order - more discussion please
- From: Graeme Gill <email@hidden>
- Date: Tue, 10 Dec 2002 12:52:35 +1100
email@hidden wrote:
>
In November Jeff Crump of the Leaf-Chronicle brought up the subject of ink
>
laydown order.
>
This subject fell out of discussion rather quickly, and like Neil I'm
>
wondering if I've missed out this detail.
There are many ways of going about profiling a printing device.
One extreme tries to make as few assumptions about the device under test
as possible. With such an approach the only assumption that you make is
that the behaviour of the device between the points you measure is not
going to be too different from those points. All you do is use some
sort of general purpose interpolation algorithm. To get accuracy, you
need to measure a lot of points. You don't care how the colors were
created, and things like laydown order, screens etc. are ireleavant
to the making of the profile. If you vary any aspect of how the device
numbers got turned into colors, then the profile is no longer valid of course,
but this is always the case.
The other extreme is to model everything about the printing process. You model
the press action. You model the screening. You model the paper. You model how
the ink interacts with the paper. You model how the ink interacts with the ink.
You spend a lot of time tuning you model up for a particular printing process,
and you allow for a certain range of parameters to vary.
For a particular profile, a list of things need to be specified and measured.
Probably not many measurements are needed. One of the parameters you might well
have to specify is the laydown order.
In practice a particular profiling approach may fall somewhere between these
two extremes.
The second approach can be very flexible in accomodating different parameters
without needing to do more than specify them (laydown order, screening lpi),
and may need few measurements to acomodate other variations (a new paper type
might require taking 1 to a dozen readings). If the model is a good one,
the resulting profile may be more accurate than a profile using the
first approach.
The problem with the second approach is that if the model isn't appropriate
for the process, it won't work very well, and will give innacurate or just
wrong results. What laydown order does an inkjet printer have ? - Answer,
it can't be answered in a simple way. An inkjet inteleaves the laydown of
different colors, and the order will alternate when running in bidirectional
mode. Where do I specify the light cyan and light magenta inks that the RIP/device
generate from the CMYK values ? Answer, you can't, it doesn't work that way,
the model isn't appropriate.
It's not stated anywhere, but I think it's fair to say that most popular profiling
packages tend to use the first approach. Such profilers don't care what the laydown
order is. But having made a profile with a given laydown order, you can't just
flick a switch on the profiler and have it generate a profile for that different
order, you have to print and measure all the patches again for that new order.
For a profiling package that uses the second approach, you have to tell it
what the laydown order is. Once you've made a profile, you _can_ flick a switch
and generate a new profile for that laydown order, without taking any new
measurements.
Graeme Gill.
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.