Re: Qualifying a CMYK Press Profile
Re: Qualifying a CMYK Press Profile
- Subject: Re: Qualifying a CMYK Press Profile
- From: Graeme Gill <email@hidden>
- Date: Mon, 25 Nov 2002 12:26:22 +1100
Darrian Young wrote:
>
Ok. I believe we are talking about the same thing which wasn't clear in your
>
last post. I do not use the term RIP interchangeably with proofing system.
>
In fact a proofing system many times does not need a RIP component - for
>
proofing raster data, whether TIFF 1Bit, Tiff-IT, etc. the only thing that
>
is needed are the linearizazion tools (which once again are always to be
>
separated from the simulation, or profile to profile conversion, or whatever
>
one prefers to call it) the printing module or program which handles the
>
jobs, and a spot color database. The RIP element of course is needed for
>
PS, PDF and EPS. If you make the distinction between "simulated device
>
space to proofing device colorspace conversion" and "calibration", then no
>
complaints here.
Depends what is meant by "RIP". Technically, filtering, scaling, color
conversion, linearisation and screening are all Image Processing operations
that are being applied to Raster data. "RIP" has also come to mean a subset of
this to some people, implying just page description language to
raster conversion, or even more narrowly, Postscript to raster
conversion. Our particular product handles both page description
input and raster format input, for flexibility in using it
for proofing. At the editorial stage, most documents are CMYK PDF
or PS documents (in fact we have a special feature to highlight
any page elements in these documents that are not CMYK).
At the end of the workflow the documents tend to be in Raster
formats, but the major steps in the proofing system processing
aren't a whole lot different, apart from skipping the rendering step.
>
Rock solid??? With an HP5000 or an Epson 7/9600 print a full calibratio
>
wedge and measure density. You will find that the density starts to go down
>
in cyan and magenta in the 80% up region (with the 5000 it is just in cyan).
>
Then, when yo look at the wedge, you will see that the density is not in
>
fact less in this range but really higher.
I don't know what systems you've been looking at, but this not
something we see under normal operating conditions. We've seen
some weirdness on pigment based printers operating in their
"fake high resolution" modes, where they increase the pixel
addressability without a corresponding reduction in dot size,
leading to 200%+ ink coverage from a single channel, but it
generally isn't useful to keep pouring ink on once you've
got 100% coverage.
Since the purpose of calibration is to put a printer back in a known
operating mode, it doesn't matter whether the density reading
corresponds to visual results or not.
>
What has happened is that there
>
is a color shift with this amount of density so the density meter does not
>
read the values correctly.
I doubt such an explanation. The whole point about density measurements
is that they are tuned for process control, and are deliberately chosen
to be narrow band to reduce sensitivity to the behaviour of the
colorant at the edge of its block band. Wideband measurements
(such as used for colorimetry) are going to be more sensitive to
color shifts (as they should be).
>
Aside from this interesting problem there is
>
also the matter of simple experience. Having two printers with the same
>
densities has not given me the same colors. If you think about it
>
logically, this is not surprising. That is why when you profile a printer
>
(which is a different process) you do not do so with 31 patches but rather
>
with a few hundred on up.
Well our experience is quite the opposite. Using a solid calibration system
has been very successful in practice in correcting for variations in individual
printer responses, and the results have withstood extremely critical
evaluation.
>
GMG Colorproof and the new ORIS Colortuner about to be released use
>
colorimetric calibration to name a couple. The issue of more expensive
>
instruments is a non-issue for two reasons. The first is that you need them
>
anyway to creat your profiles, and the second is that if you are talking
>
about having to have numerous devices at remote sites, etc. the i1 is not
>
expensive. I also don't think linearizazion is a misleading term.
As a manufacturer, we need a wider view of what is expensive or not.
Many of our customers aren't prepared to purchase a densitometer,
let alone a spectral instrument. Few really want to make their own
profiles (profiling is time consuming, and currently requires some
skill to get right), although more feel that somehow they should be
making their own profiles, even though they're often not sure why.
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.