ICC profiles

[Mark Herring]

I am getting set up with MIS archival inks on my 1280. **Excellent**
results so far. I will be moving to their "perpetual archival" ink,
for which they recommend profiling. I also will be trying some B&W
inks.

What is the most cost-effective way of getting ICC profiles? Some
Googling** has yielded a wide range of solutions:

Monaco---very expensive

Colorvision---expensive but within reach

Profile Prism (DDI Software)---cheap!!

Color valet (Chromix): $99 per profile

thanks

-Mark

"Google" (verb, transitive): 1) To search thru hundreds of search
results that come very close to your criteria, but never quite get
there. 2) To get an instant short cut to a known site, because you
are too lazy to type it or look it up in Favorites. 3) To search for
a specific word or phrase---as in: Why don't you Google <<word>>?
**************************
Mark Herring, Pasadena, Calif.
Private e-mail: Just say no to "No".

 

[KenP]

I am getting set up with MIS archival inks on my 1280. **Excellent**
results so far. I will be moving to their "perpetual archival" ink,
for which they recommend profiling. I also will be trying some B&W
inks.

What is the most cost-effective way of getting ICC profiles? Some
Googling** has yielded a wide range of solutions:

Monaco---very expensive

Colorvision---expensive but within reach

Profile Prism (DDI Software)---cheap!!

Color valet (Chromix): $99 per profile

thanks

-Mark

"Google" (verb, transitive): 1) To search thru hundreds of search
results that come very close to your criteria, but never quite get
there. 2) To get an instant short cut to a known site, because you
are too lazy to type it or look it up in Favorites. 3) To search for
a specific word or phrase---as in: Why don't you Google <<word>>?
**************************
Mark Herring, Pasadena, Calif.
Private e-mail: Just say no to "No".

I use the ColorVision ProfilerPLUS program. It seems to work as
claimed, but don't be surprised if you have to do some minor tweaking
to get it right. It's easy and straightforward to use, and for about
$100 (no hardware) it's an easy way to manage the printer/paper
problem. ColorVision may be disontinuing it, since it doesn't show up
on some of the price lists now. The Profiler program depends on the
accuracy of your scanner, and this is likely to be the weak link.
They do offer a more expensive solution that includes a small
dedicated scanner.

Maybe a scanner could be checked for color accuracy by scanning in a
reference color sample, such as a Macbeth chart, then comparing the
scanned RGB values to those that are known for the 24 color samples of
the chart. Does anyone have a listing of the reflected color density
values for the Macbeth chart?

 

[JSK]

You might want and try inkjetsusa.com CFS systems.
I just recived mine - Make's MIS systems looks like
a Kmart special. They even use medical filters on their
bottles.

J

 

[Roland Karlsson]

Profile Prism (DDI Software)---cheap!!

I have not got Prism to work correctly.
For me it produces very strange profiles.
They are not smooth and I get bad results.

I think that this software makes by far
to aggressive profiles, assuming that the
printer can be arbitrarily bad. It would be
better if the profiler assumed that the
printer does a decent job and then just
tried to smoothly adjust the color rendition.


/Roland

 

[David Jacobson]

You might want and try inkjetsusa.com CFS systems.
I just recived mine - Make's MIS systems looks like
a Kmart special. They even use medical filters on their
bottles.

J

This discussion is addressing two topics. 1. Monitor profiles, and
2. printer
profiles.

For monitor profiles I use the ColorVision PhotoCal system. It is
simple, and works perfectly. I love it except for their highly
restrictive licensing. The
license agreement says you can only calibrate one computer. That
would be like
me buying a wrench and the maker telling me I can only use it to work
on one car.

Printer profile making is much more troublesome. I first used
ColorCalRGB
from the same company. Total junk. It claims that it can build a
good profile
with an uncalibrate scanner. Common sense would tell you that it
can't work.
But their FAQ addresses that very issue, but the answer basically says
"We've
got a secret algorithm." Don't believe it. They had a free upgrade
to the next
product, ProfilerPLUS. Also junk.

I finally gave up and bought MonacoEZColor. (Not terrbily expensive.)
That seems to be pretty good. It has a calibrated IT8 test target
that you
scan along with the printed target. But I'm still not happy with the
way it
works on my Epson 1270 using premium glossy or competing papers such
as
Olmec Brilliant White Gloss. The prints seem to have posterizing in
areas of deep reddish brown, like leather and stained wood. The
problem might be with the
printer. If I just print out gradient patterns in raw mode, you can
see definite
non-monotonic regions, and regions where there are step changes. If
my
conjecture is correct, there is some bad thing with the 1270 and
glossy papers, and ordinary profilers that assume that the funcction
mapping between code and color is monotonic in each dimension are
going to have trouble.

One more word of advice to people using MonacoEZColor. They act like
you
just scan the IT8 target and test print in and all will be well. Not
true.
Scan into Photoshop and carefull look for clipped colors (one of the
channels reading 255). If you process the scan with clipped colors,
the
results will be truly terrible. Repeat with lower the gain or
saturation
or something on your scanner to make sure all colors are in range, and
then
make the profile.

-- David Jacobson

 

[KBob]

snip<
Printer profile making is much more troublesome. I first used
ColorCalRGB
from the same company. Total junk. It claims that it can build a
good profile
with an uncalibrate scanner. Common sense would tell you that it
can't work.
But their FAQ addresses that very issue, but the answer basically says
"We've
got a secret algorithm." Don't believe it. They had a free upgrade
to the next
product, ProfilerPLUS. Also junk.

I got ProfilerPlus to work well with a couple glossy papers, but it
failed terribly with Epson Matte, giving way oversaturated colors. Be
sure to go over the scanned target and fix bad spots prior to
generating a profile with it. I'm still undecided as to whether the
program really works, or whether it's operator error on this end. Is
there any way to profile/correct a scanner?

 

[Michael J Davis]

I got ProfilerPlus to work well with a couple glossy papers, but it
failed terribly with Epson Matte, giving way oversaturated colors. Be
sure to go over the scanned target and fix bad spots prior to
generating a profile with it. I'm still undecided as to whether the
program really works, or whether it's operator error on this end. Is
there any way to profile/correct a scanner?

I have not been successful in my bungling attempts to get profiling to
work, but that's because I haven't followed the route through - I've
tweaked monitor and printer profiles, (the latter by importing profiles
for different paper types) but never been fully satisfied.

However, I have a problem with the whole process, that seems to me to be
'dumbing down' the colour gamut to the weakest link. If the process is
consistency in the scanner -> screen -> printer -> scanner circuit,
doesn't that mean that, often, each component restricts the colour
range, and we end up with the worst of all worlds *but* with more
control?

Or am I just misunderstanding something?

Mike
[The reply-to address is valid for 30 days from this posting]
--
Michael J Davis
<><
Some newsgroup contributors appear to have confused
the meaning of "discussion" with "digression".
<><

 

[Mark Herring]

I got ProfilerPlus to work well with a couple glossy papers, but it
failed terribly with Epson Matte, giving way oversaturated colors. Be
sure to go over the scanned target and fix bad spots prior to
generating a profile with it. I'm still undecided as to whether the
program really works, or whether it's operator error on this end. Is
there any way to profile/correct a scanner?

I have not been successful in my bungling attempts to get profiling to
work, but that's because I haven't followed the route through - I've
tweaked monitor and printer profiles, (the latter by importing profiles
for different paper types) but never been fully satisfied.

However, I have a problem with the whole process, that seems to me to be
'dumbing down' the colour gamut to the weakest link. If the process is
consistency in the scanner -> screen -> printer -> scanner circuit,
doesn't that mean that, often, each component restricts the colour
range, and we end up with the worst of all worlds *but* with more
control?

Or am I just misunderstanding something?

Mike
[The reply-to address is valid for 30 days from this posting]

That does not sound right to me. Profiling is about calibrating the
transfer function of each component in the system. When you combine
system elements, you do not somehow combine there calibration factors
in such a way that the whole system degrades.

What is true is that the residual **errors** are combined---ie the
more things in the loop, the more uncertainty there will be.

I am just getting ready to use MIS "perpetual archival" ink in my
1280. MIS states that one needs profiling to get good results from
this ink set. This is true of many 3rd party ink / paper combos.

-Mark
**************************
Mark Herring, Pasadena, Calif.
Private e-mail: Just say no to "No".

 

[Graeme Cogger]

I have not been successful in my bungling attempts to get profiling to
work, but that's because I haven't followed the route through - I've
tweaked monitor and printer profiles, (the latter by importing profiles
for different paper types) but never been fully satisfied.

However, I have a problem with the whole process, that seems to me to be
'dumbing down' the colour gamut to the weakest link. If the process is
consistency in the scanner -> screen -> printer -> scanner circuit,
doesn't that mean that, often, each component restricts the colour
range, and we end up with the worst of all worlds *but* with more
control?

Or am I just misunderstanding something?

There is some truth in what you say - each profile limits the
gamut (range of colours) in the image to what the profile can
handle. But the only place where you can lose data is in your
working space profile.

The scanner profile describes what the scanner is capable of
outputting - there are no losses here, since the purpose of the
profile is to describe what the RGB values from the scanner
represent.

After scanning, the image is normally converted to a more 'well
behaved' profile such as Adobe RGB. This has nice features
such as neutral tones having equal R, G and B values. This is
the only point where you can lose data. If some of the colours
in the scanner profile are out of gamut in the working space,
then colour information is lost. This is the main reason to
use a nice wide gamut profile like Adobe RGB instead of a
narrower one such as sRGB.

The monitor profile does not cause problems since the image is
never converted to it - the profile is only used for display
purposes.

The printer profile, similarly to the scanner profile, simply
describes what the printer is capable of printing. Colour data
is often lost here, but it would be lost anyway since the
printer cannot reproduce all the colours of the working space
or scanner. By using the printer profile, we are managing
those losses rather than letting them occur in an uncontrolled
way. By soft-proofing (and making sure we display out of gamut
colours) we can decide exactly how to deal with the problem.

So in summary: as long as the working space has a wider gamut
than the input devic,e we do not suffer any more losses than
are inevitable due to the hardware we are using.

 

[Michael J Davis]

There is some truth in what you say - each profile limits the
gamut (range of colours) in the image to what the profile can
handle. But the only place where you can lose data is in your
working space profile.

The scanner profile describes what the scanner is capable of
outputting - there are no losses here, since the purpose of the
profile is to describe what the RGB values from the scanner
represent.

Yes. My interest is in the photo input so my scanner was a Canon film
scanner, but I have now gone digital - my Fuji camera seems to have a
wider gamut.

After scanning, the image is normally converted to a more 'well
behaved' profile such as Adobe RGB. This has nice features
such as neutral tones having equal R, G and B values. This is
the only point where you can lose data. If some of the colours
in the scanner profile are out of gamut in the working space,
then colour information is lost. This is the main reason to
use a nice wide gamut profile like Adobe RGB instead of a
narrower one such as sRGB.

Yes, the 'working space' has to be larger than any part. But of course,
in theory the working space can be infinite. (so to speak)

The monitor profile does not cause problems since the image is
never converted to it - the profile is only used for display
purposes.

That's a good point. The limit of the monitor is merely that we may not
be able to 'see' what is happening in the working space.

The printer profile, similarly to the scanner profile, simply
describes what the printer is capable of printing. Colour data
is often lost here, but it would be lost anyway since the
printer cannot reproduce all the colours of the working space
or scanner. By using the printer profile, we are managing
those losses rather than letting them occur in an uncontrolled
way. By soft-proofing (and making sure we display out of gamut
colours) we can decide exactly how to deal with the problem.

Thanks. I have found that various ICC profiles available for the printer
and papers that I use, tend to introduce colour tints (I'm not sure why)
so have found a printer profile that seems manageable, and then I
process the image to 'fill' the colour space (I usually slightly
desaturate somewhat before saving and printing). But then I'm after
dramatic images rather than accuracy! ;-)

So in summary: as long as the working space has a wider gamut
than the input devic,e we do not suffer any more losses than
are inevitable due to the hardware we are using.

Thanks that's a very helpful response.

Mike
[The reply-to address is valid for 30 days from this posting]
--
Michael J Davis
<><
Some newsgroup contributors appear to have confused
the meaning of "discussion" with "digression".
<><

 

[Paul]

I have not got Prism to work correctly.
For me it produces very strange profiles.
They are not smooth and I get bad results.

I think that this software makes by far
to aggressive profiles, assuming that the
printer can be arbitrarily bad. It would be
better if the profiler assumed that the
printer does a decent job and then just
tried to smoothly adjust the color rendition.


/Roland

On the other hand

I use profile prism and get excellent results with my Epson 1200 & 1290.

How good your profiles are is often dependent on how good your scanner is
and how good you are at following instructions.

I use an Agfa Snapscan e20 with Vuescan in RAW mode and get vary accurate
profiles, much better than I got with the ColorVision ProfilerPLUS program.

Paul