Computer Graphics with Photoshop - Calibration Advice --- Hardware/software

Author

Calibration Advice --- Hardware/software

2005-12-13, 6:19 pm

I have been struggling with monitor and printer
calibration.

Could someone recommend a unified solution ---
something that will aid me in calibrating my
monitor, printer and hopefully, scanner?

I do not mind spending a reasonable amount
of money (I think the good ones cost around
$500).

Good documentation is critical, as I have not
fully grasped all the complications of color
management. If the documentation that comes
with the product is not all that one might ask
for, some good web based tutorials would do.

I have PS CS2, a Canon S9000 printer, and an Epson
4780 scanner. I run windows XP.

Thanks in advance,

--
Andrew Hall
(Now reading Usenet in comp.graphics.apps.photoshop...)

Jim

2005-12-13, 6:19 pm

<ahall@no-spam-panix.com> wrote in message
news:kpgk6e8hi2j.fsf@panix3.panix.com...
>
> I have been struggling with monitor and printer
> calibration.
>
> Could someone recommend a unified solution ---
> something that will aid me in calibrating my
> monitor, printer and hopefully, scanner?
Monaco or Colorvision do all of those.
>
> I do not mind spending a reasonable amount
> of money (I think the good ones cost around
> $500).
>
> Good documentation is critical, as I have not
> fully grasped all the complications of color
> management. If the documentation that comes
> with the product is not all that one might ask
> for, some good web based tutorials would do.
Join the croud. Color management can be extremely frustrating.
Usually, however, the documentation is not a tutorial. For that, I suggest
"Real World Color Management" by Bruce Fraser.
>
> I have PS CS2, a Canon S9000 printer, and an Epson
> 4780 scanner. I run windows XP.
>
> Thanks in advance,
>
>
> --
> Andrew Hall
> (Now reading Usenet in comp.graphics.apps.photoshop...)
Jim

peevee_hermann

2005-12-13, 6:19 pm

On Tue, 13 Dec 2005 20:37:21 GMT, "Jim" <j.n@nospam.com> wrote:

>
><ahall@no-spam-panix.com> wrote in message
>news:kpgk6e8hi2j.fsf@panix3.panix.com...
>Monaco or Colorvision do all of those.

Monaco aint too hot for printer calibration, at least, the version I
have is not wonderful. I have EZColor Optix. It calibrates the monitor
nicely to where I get a nice neutral flat grey where it is supposed to
be grey. However, to profile a printer, you have to print their
calibration tiff and then scan it. BBBZZZZTTTTTTT. Not good. Then you
are supposed to edit the profile created with their not very intuitive
profile editor.

>Join the croud. Color management can be extremely frustrating.
>Usually, however, the documentation is not a tutorial. For that, I suggest
>"Real World Color Management" by Bruce Fraser.

thats a good book but very tedious in spots to read. I did learn some
stuff however.

gretag, thats what I would get if i had the money to spend. You get a
densitometer or calibrator hardware of some type. Print their patch
and then read the patch right into the program, no scanner required.
Not cheap tho.

http://world.i1color.com/

Bill Hilton

2005-12-13, 6:19 pm

> Andrew Hall writes ...
>
>I have been struggling with monitor and printer
>calibration.

Monitor should be easy, printer is harder (been there, done that).

>Could someone recommend a unified solution ---
>something that will aid me in calibrating my
>monitor, printer and hopefully, scanner?

My advice is to spend the $$ necessary to accurately profile your
monitor, but buy a printer with excellent built-in profiles or have
custom ones made and save your money because it's tougher to get
excellent printer profiles on your own at a reasonable cost.

>I do not mind spending a reasonable amount
>of money (I think the good ones cost around
>$500).

I think the three best monitor profile solutions are around $200 - 230
each, but I personally feel you cannot get a decent printer profile
solution for less than whatever the Eye-One spectrophotometer package
is going for, which is probably $1,200 - $1,400. The cheaper printer
profiler packages use flatbed scanners for the critical measurements
and for most all of us this isn't good enough. The ColorVision
PrintFix (or something like that) is in your price range but most
reviews of this have been luke-warm to negative.

Here are a couple of reviews of good monitor profile packages ... I've
used the Spyder (original) and recently the Gretag Eye-One, which does
much better on LCDs, and I've used the Monaco software for scanners and
printers but found the documentation obtuse and the support staff
lacking.
http://www.outbackphoto.com/color_m...m_08/essay.html
http://www.macworld.com/2005/03/rev...rator/index.php

>Good documentation is critical, as I have not
>fully grasped all the complications of color
>management.

To really learn this stuff get "Real World Color Management" by Fraser
et al.

>I have PS CS2, a Canon S9000 printer, and an Epson
>4780 scanner

>From what I hear the S9000 doesn't ship with very accurate profiles
(typical of many consumer printers, unfortunately). If you use just a
couple of papers it might be worth it to pay someone to build custom
profiles for just those papers rather than spend another grand for the
Eye-One spectrophotometer.

I personally think Epson has done a pretty good job of shipping
accurate profiles with their better Photo pigment ink printers the past
couple of years. They came out with the 9600/7600 wide format pro
printers about 3 years ago and were very embarassed when a genius
amateur (Bill Atkinson, who was one of the early software whizzes at
Apple) created very accurate profiles that were far superior to what
Epson created. Bill gave these profiles away to anyone who wanted to
download them and Epson came to him to learn how he did it and he
showed them, and subsequently the profiles they've provided are a lot
better these days. I have an Epson 4000 Pro for example with excellent
built-in profiles from Epson, and Epson also created profiles for the
2200 that were better than what were shipped earlier and better than
cusom profiles a friend was able to make for my 2200 with an Eye-One.
Of course if you use non-Epson papers you'll need to make a custom
profile or rely on what the paper vendor supplies, and these
vendor-supplied profiles are very much hit or miss.

At any rate, spend the bucks to get the monitor done right, but that's
the easy part. Just my opinion, based on several miles of rocky road
on the way to completely accurate color management :)

Bill

Jim

2005-12-13, 6:19 pm

"peevee_hermann" <peevee_hermann@yahoo.com> wrote in message
news:2heup154bo515m3pq4ld2q5565cu1ilep0@4ax.com...
> On Tue, 13 Dec 2005 20:37:21 GMT, "Jim" <j.n@nospam.com> wrote:
>
>
>
> Monaco aint too hot for printer calibration, at least, the version I
> have is not wonderful. I have EZColor Optix. It calibrates the monitor
> nicely to where I get a nice neutral flat grey where it is supposed to
> be grey. However, to profile a printer, you have to print their
> calibration tiff and then scan it. BBBZZZZTTTTTTT. Not good. Then you
> are supposed to edit the profile created with their not very intuitive
> profile editor.
Monaco works for me, and that is all I care about.
Jim

Jim

2005-12-13, 6:19 pm

"Bill Hilton" <bhilton665@aol.com> wrote in message
news:1134509281.110584.173000@g47g2000cwa.googlegroups.com...
>
> Monitor should be easy, printer is harder (been there, done that).
>
>
> My advice is to spend the $$ necessary to accurately profile your
> monitor, but buy a printer with excellent built-in profiles or have
> custom ones made and save your money because it's tougher to get
> excellent printer profiles on your own at a reasonable cost.
>
>
> I think the three best monitor profile solutions are around $200 - 230
> each, but I personally feel you cannot get a decent printer profile
> solution for less than whatever the Eye-One spectrophotometer package
> is going for, which is probably $1,200 - $1,400. The cheaper printer
> profiler packages use flatbed scanners for the critical measurements
> and for most all of us this isn't good enough. The ColorVision
> PrintFix (or something like that) is in your price range but most
> reviews of this have been luke-warm to negative.
>
> Here are a couple of reviews of good monitor profile packages ... I've
> used the Spyder (original) and recently the Gretag Eye-One, which does
> much better on LCDs, and I've used the Monaco software for scanners and
> printers but found the documentation obtuse and the support staff
> lacking.
> http://www.outbackphoto.com/color_m...m_08/essay.html
> http://www.macworld.com/2005/03/rev...rator/index.php
>
>
> To really learn this stuff get "Real World Color Management" by Fraser
> et al.
>
>
> (typical of many consumer printers, unfortunately). If you use just a
> couple of papers it might be worth it to pay someone to build custom
> profiles for just those papers rather than spend another grand for the
> Eye-One spectrophotometer.
>
> I personally think Epson has done a pretty good job of shipping
> accurate profiles with their better Photo pigment ink printers the past
> couple of years. They came out with the 9600/7600 wide format pro
> printers about 3 years ago and were very embarassed when a genius
> amateur (Bill Atkinson, who was one of the early software whizzes at
> Apple) created very accurate profiles that were far superior to what
> Epson created. Bill gave these profiles away to anyone who wanted to
> download them and Epson came to him to learn how he did it and he
> showed them, and subsequently the profiles they've provided are a lot
> better these days. I have an Epson 4000 Pro for example with excellent
> built-in profiles from Epson, and Epson also created profiles for the
> 2200 that were better than what were shipped earlier and better than
> cusom profiles a friend was able to make for my 2200 with an Eye-One.
> Of course if you use non-Epson papers you'll need to make a custom
> profile or rely on what the paper vendor supplies, and these
> vendor-supplied profiles are very much hit or miss.
>
> At any rate, spend the bucks to get the monitor done right, but that's
> the easy part. Just my opinion, based on several miles of rocky road
> on the way to completely accurate color management :)
>
> Bill
>
Just an observation - Bill is discussing higher end products whereas I
mentioned the low end ones.
And, I do agree that the current profiles from Epson are quite good.
Jim

2005-12-13, 6:22 pm

>>>>> Jim writes:

Jim> "Bill Hilton" <bhilton665@aol.com> wrote in message
news> 1134509281.110584.173000@g47g2000cwa.googlegroups.com...[color=darkred]
http> www.outbackphoto.com/color_management/cm_08/essay.html
http> www.macworld.com/2005/03/reviews/mo...rator/index.php[color=darkred]
Jim> Just an observation - Bill is discussing higher end products whereas I
Jim> mentioned the low end ones.
Jim> And, I do agree that the current profiles from Epson are quite good.
Jim> Jim

Thank you all. I actually have a version of the Monaco that
came with the scanner, but I did not think it worth exploring
until I got the monitor right.

I will take Bill's advice and get the monitor right, then try
out the Monaco for the scanner and printer. If that does not
work, perhaps I will give the printer to a relative and buy an
Epson :)

Thanks again,

--
Andrew Hall
(Now reading Usenet in comp.graphics.apps.photoshop...)

Bill Hilton

2005-12-14, 6:14 pm

> Andrew Hall writes ...
>
> I will take Bill's advice and get the monitor right, then
> try out the Monaco for the scanner and printer.

After you create the printer profile with Monaco I'd recommend printing
a test pattern with a black-white gradient to evaluate the profile.
The one I use was created by Bill Atkinson and he has made it available
as a free download (let me know if you want a link) ... here's a
version of it (which you can't use because it's a small jpeg, you
should get the original) to show what I mean ... I use it for profile
evaluation and look carefully at the skin tones and colors I care most
about for landscape photography, like the aspens and the red rocks and
sky, but first I check the gradient at the bottom (left on this rotated
image) ... bad profiles will stay black and/or white for a while and
then make an abrupt jump to the gradient (indicating limited tonal
range), or more commonly there will be a green or sometimes magenta
cast in say 25% of the gradient, a sign the profile isn't properly gray
balanced thru the whole tonal range. The dye ink printers like the
Epson 1280 are especially hard to gray balance, I've found, but even
some custom vendor-supplied profiles for the 4000 Pro and pricey custom
art papers often have this problem, since they skimped on paying for
good accurate profiles ...
http://members.aol.com/bhilton665/a...n_test_file.jpg

> If that does not work, perhaps I will give the printer to a
> relative and buy an Epson :)

Look around for a site that offers custom test prints from various
printers/papers and maybe buy a couple of prints and print the same
pattern on your S9000 and compare side-by-side before shelling out for
a new printer ... the S9000 is a good dye ink printer, similar to the
Epson 1280, which I still own (though it doesn't get much use). When
the Epson pigment printers first came out I ordered test prints from a
2000p and felt I was better off keeping the 1280 because of the colors,
but when I got 2200 test prints I felt it was time to change. There is
no substitute for having prints of the same file in hand and doing a
side-by-side comparison. The S9000 you have is a pretty good printer
so make sure you have good reason to switch to Epson :)

Bill

ChuckT

2005-12-15, 6:14 pm

One of the questions we should all be asking is "What if ..."
When things go wrong, or your results aren't what you been led to
expect - who answers your questions?

I've been to $everal seminar$ that are supposed to be the be-all &
end-all for color management and am not at all convinced that anyone
has the whole answer.

Epson, especially, has a lot to explain on what the switches in their
printer drivers do and how the printer profiles work. I have several
of the "Real World Photshop" books as well as Real World Color
Management" and the Monaco Eye-One Photo suite. I remain
under-whelmed. We have 3 2200s and a 7600 (I use the Atkinson profiles
and their labeling is obtuse).

cvt

ToniG552@att.net

2005-12-15, 6:15 pm

Bill Hilton wrote:
> If you use just a
> couple of papers it might be worth it to pay someone to build custom
> profiles for just those papers rather than spend another grand for the
> Eye-One spectrophotometer.

Can you recommend a good printer profile vendor? Drycreek and Cathy's
got mentioned a lot, and are lower cost. But are they any good?

> I personally think Epson has done a pretty good job of shipping
> accurate profiles with their better Photo pigment ink printers the past
> couple of years. They came out with the 9600/7600 wide format pro
> printers about 3 years ago and were very embarassed when a genius
> amateur (Bill Atkinson, who was one of the early software whizzes at
> Apple) created very accurate profiles that were far superior to what
> Epson created. Bill gave these profiles away to anyone who wanted to
> download them and Epson came to him to learn how he did it and he
> showed them, and subsequently the profiles they've provided are a lot
> better these days.

At one time Atkinson made profiles for the 2200. Are these still
available?

ToniG552@att.net

2005-12-15, 6:15 pm

Bill Hilton wrote:

> After you create the printer profile with Monaco I'd recommend printing
> a test pattern with a black-white gradient to evaluate the profile.
> The one I use was created by Bill Atkinson and he has made it available
> as a free download (let me know if you want a link)

Please provide the link. How should this test pattern be printed in PS?
For example, when opened in PS, should it be converted to a working
color space first before printing? After opening the file, it probably
should be printed like you would with a normal image, i.e. convert to
printer profile and No Color Management at Epson??

Bill Hilton

2005-12-15, 6:15 pm

>> The one I use was created by Bill Atkinson and he has made it
[color=darkred]
> ToniG...@att.net writes ...
>
>Please provide the link.

You can get it from http://www.inkjetart.com/custom/ ... scroll down to
"You can download the original file we are using from Bill Atkinson's
mac.com page (it's the LAB test image)" or use the cut down version
that InkjetArt uses since the original is a very large file.

> How should this test pattern be printed in PS?

Exactly like you would normally print an image ... it's in LAB mode so
you may want to convert to RGB, but that shouldn't make a difference.

>For example, when opened in PS, should it be converted to a working
>color space first before printing?

It's LAB mode so if you just change the mode to RGB it will
automatically pick up whatever 'working space' you are using as
default, ie, sRGB or AdobeRGB or whatever.

>After opening the file, it probably should be printed like you
>would with a normal image, i.e. convert to
>printer profile and No Color Management at Epson??

You can do it that way or you can use the 'soft-proof' method, which is
probably more popular (and what I'm used to).

Bill

Bill Hilton

2005-12-15, 6:15 pm

>Chuck T writes ...
>
> We have 3 2200s and a 7600 (I use the Atkinson profiles
> and their labeling is obtuse).

I assume you mean the "PLU1" in a profile name like "9600 PremLuster
PLU1 Std V3.icc" ... Bill had a PDF that explains what "PLU1" and the
other cryptic headers meant, basically describing the exact Epson
settings used for ten parameters in the printer driver software. This
was a free download on the Epson Pro Graphics site with the profiles, I
think (I got mine on a CD when I took a class with him). So you just
set all these parameters to the same settings and save this off as a
preset and it's easy after that.

If you don't have the PDF email me and I can send it to you, it's a 171
KB file, the one I have is titled "About Bill's Profiles Dec4.pdf" (Dec
4, 2002 so if you have a later version you're already golden).

Bill

Bill Hilton

2005-12-15, 6:15 pm

> ToniG...@att.net writes ...
>
>Can you recommend a good printer profile vendor?

I don't have any personal recommendations but when I took the class
with Atkinson he recommended "Profile City" and also Andrew Rodney's
company, do a Google search on "andrew rodney custom profiles" for
plenty of info. This was probably 3 years ago so dunno how current
these recommendations are.

> At one time Atkinson made profiles for the 2200. Are these
> still available?

At the workshop he said he felt there was too much variation between
units and too much drift over time on any single unit for him to make
them available as a professional standard (the consumer grade printers
are made to different standards than the Professional models). He
specifically said the 9600 profiles wouldn't work right on a 2200. So
as far as I know he never released 2200 profiles. The ones Epson put
out in spring 2004 are a lot better than the earlier Epson-supplied
ones (after Epson learned from Bill how to do it :) and better than a
custom profile I had made by a friend with a Gretag Eye-One, in case
you haven't downloaded those.

Bill

C Wright

2005-12-15, 6:15 pm

On 12/15/05 10:59 AM, in article
1134665948.347224.183050@g43g2000cwa.googlegroups.com, "Bill Hilton"
<bhilton665@aol.com> wrote:

>
>
> You can get it from http://www.inkjetart.com/custom/ ... scroll down to
> "You can download the original file we are using from Bill Atkinson's
> mac.com page (it's the LAB test image)" or use the cut down version
> that InkjetArt uses since the original is a very large file.
>
>
> Exactly like you would normally print an image ... it's in LAB mode so
> you may want to convert to RGB, but that shouldn't make a difference.
>
>
> It's LAB mode so if you just change the mode to RGB it will
> automatically pick up whatever 'working space' you are using as
> default, ie, sRGB or AdobeRGB or whatever.
>
>
> You can do it that way or you can use the 'soft-proof' method, which is
> probably more popular (and what I'm used to).
>
> Bill
>

A comment and a question -
I have used this image (down loaded from a source Bill Hilton provided
previously) and have found it to be an outstanding way to test paper profile
quality.
The question - Bill, does it really make no difference if the image is
converted to RGB from lab color? I have always operated under the theory
that the less conversion done to a critical test image like this one the
better. Does it really make no difference once the image is down loaded and
on your computer?
Chuck

Mike Russell

2005-12-15, 6:15 pm

"ChuckT" <ctintera@digital.net> wrote in message

> One of the questions we should all be asking is "What if ..."
> When things go wrong, or your results aren't what you been led to
> expect - who answers your questions?

There are always questions, because in most situations calibration in and of
itself cannot produce good color, let alone excellent color. For that you
must pay attention to the characteristics of each image, and the printing
process.
---
Mike Russell
www.curvemeister.com

Bill Hilton

2005-12-15, 6:15 pm

> Chuck Wright asks ...
>
>I have used this image (down loaded from a source Bill Hilton provided
>previously) and have found it to be an outstanding way to test paper
>profile quality.
>The question - Bill, does it really make no difference if the image is
>converted to RGB from lab color?

The way I understand it, LAB mode is what's used every time there's a
conversion as the intermediate space/mode ... that is, if you convert
to profile from AdobeRGB to sRGB for example the AdobeRGB data gets
translated to LAB and then from LAB to sRGB. So if you keep Bill's
image in LAB mode and print it you are saving yourself two
translations.

I've printed Bill's image directly from LAB mode, from AdobeRGB and
from Ektaspace (wide gamut working space that matches the color range
of most E-6 films) and didn't see any differences in the final prints
though (just looking at them, not making measurements).

> I have always operated under the theory that the less conversion
> done to a critical test image like this one the better.

I think all the experts would agree with this as a general theory, but
in practice the changes are apparently small for typical conversions.
I seem to remember reading Dan Margulis' book where he said he
converted back and forth 20 times between CMYK and RGB (if I remember
it correctly) to check this and didn't see any difference in the final
output ... for sharpening I switch to LAB mode and sharpen via an edge
mask on the L channel (something I learned from Atkinson) and so was
concerned about converting between modes as well ... I ran some tests
and found that if you make a gradient in say AdobeRGB and look at the
RGB histogram you'll see some "fuzz" along the top of the histogram,
then if you convert to LAB and back to RGB the "fuzz" is smoothed out
of the histogram, which indicates *something* was changed. But as far
as actually seeing a difference in prints, I'm with Margulis on this as
I just didn't see it ...

>Does it really make no difference once the image is down loaded and
>on your computer?

Have you tried it both ways (just LAB vs LAB converted to RGB) and if
so did you see a difference in the prints? Just curious ...

Bill

C Wright

2005-12-15, 6:17 pm

On 12/15/05 1:13 PM, in article
1134674019.860591.266060@f14g2000cwb.googlegroups.com, "Bill Hilton"
<bhilton665@aol.com> wrote:

snip
>
>
> I think all the experts would agree with this as a general theory, but
> in practice the changes are apparently small for typical conversions.
> I seem to remember reading Dan Margulis' book where he said he
> converted back and forth 20 times between CMYK and RGB (if I remember
> it correctly) to check this and didn't see any difference in the final
> output ... for sharpening I switch to LAB mode and sharpen via an edge
> mask on the L channel (something I learned from Atkinson) and so was
> concerned about converting between modes as well ... I ran some tests
> and found that if you make a gradient in say AdobeRGB and look at the
> RGB histogram you'll see some "fuzz" along the top of the histogram,
> then if you convert to LAB and back to RGB the "fuzz" is smoothed out
> of the histogram, which indicates *something* was changed. But as far
> as actually seeing a difference in prints, I'm with Margulis on this as
> I just didn't see it ...
>
>
> Have you tried it both ways (just LAB vs LAB converted to RGB) and if
> so did you see a difference in the prints? Just curious ...
>
> Bill
>
Yes, I have tried it both ways; and no, I have not seen any difference in
prints! I just can sometimes get kind of anal when it comes to test prints.
Based on the examples/tests that you site I think that I won't let it bother
me any further!
Chuck