Adobe Photoshop Training with Dan Margulis

Dan Margulis Applied Color Theory

Use of Color Bars in Flat Art Repro

Date: Fri, 04 Jun 2004 09:54:11 -0400

From: Philip Denlinger

Subject: use of color bars in flat art repro

if i have a color transparency, where someone has helpfully included a gray scale and color control patches:

<http: //www.mindbet.com/images/kodakcolorstrip.jpg>

<http: //www.mindbet.com/images/kodakcolorstrip300.tif>

how do i use these for CMYK reproduction?

do i use curves to force the color patches to certain CMYK values..

if so, what values..

do i bend the gray scales to certain tint percentages... if so, what.

after working with this for a bit, i came to to the conclusion that all i could do was neutralize the gray... is that correct?

TIA

philip

___________________________________________________________________________

Date: Fri, 04 Jun 2004 11:05:33 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

Philip,

I’ve always added Gray Scales and Color Bars to repro trannys as a reference only! Using the White and Black patches is sometimes a useful way for the printer to set the CMYK white point and black point but that is about all. It also shows you the accuracy of the transparencies exposure (over or under, with either the whites or black swatches being washed out or blocked up).

Trying to color correct the color bars or gray scale is not what is important here, having color correct artwork is. Does the tranny accurately represent the art when viewed with a 5000º K light source and the tranny on a 5000º K light box? If so color correct the print to the tranny. If not you will need to get the original art and try to color correct to that if available. If the original is not available you need to talk with the client and have them make some judgment calls as to how the print should be “different” than the tranny.

BTW, depending upon the media and how the original was made, the “printed” color bar and gray scale in the tranny may reproduce radically different than the original art. When I show an artist a transparency repro of their art they are not concerned how the scales show, they want the art to be as close as possible. Hence, don’t worry about the scales, color correct to the original.

Hope this helps.

Tom Kaczmarek

___________________________________________________________________________

Date: Fri, 04 Jun 2004 11:03:43 -0400

From: Lanny Lathem

Subject: Re: use of color bars in flat art repro

I think this is a great tool to have in the image.

I would neutralize the highlight, the shadow and the midtone on the grayscale. This will take you toward a match of the artwork assuming you don’t have the origninal art work with which to evalute your image.

I would then check the highlight and shadow (assuming there is a highlight) in the ‘live’ image area itself to make sure those adjustments didn’t do someting undesirable due to uneven lighting when the photo was taken.

Lanny Lathem

Bennett Graphics

___________________________________________________________________________

Date: Fri, 04 Jun 2004 09:06:14 -0700

From: Peter Figen

Subject: Re: use of color bars in flat art repro

Having put these color bars in images next to artwork, and then having the luxury of color correcting to the actual artwork under a Solux light, I can say that the color and gray bars have almost no usable value if you are truly trying to match the art. I think the biggest culprit is that the dyes in the printed color bars don’t have the same spectral characteristics as those in the art, let alone how the film or digital camera used to record the art responds. Anyone who relies on color bars alone is bound to be disappointed.

Peter Figen

___________________________________________________________________________

Date: Fri, 04 Jun 2004 12:12:41 -0400

From: Lanny Lathem

Subject: Re: use of color bars in flat art repro

It all depends on expectation. If you have no other reference point, as I mentioned, a color bar is of some use in removing cast and getting the file to look somewhat more like the original. Of course correcting to the original art is the best way to go if possible. Sometimes there is NO access to the art and the goal would only be to achieve pleasing color which will never be compared to anything else.

Using a color bar is not a perfect scenario, but it’s better than nothing.

Lanny Lathem

Bennett Graphics

___________________________________________________________________________

Date: Fri, 4 Jun 2004 10:20:10 -0600

From: “Les De Moss”

Subject: Re: use of color bars in flat art repro

Forgive me for stating the (what may be) obvious.

It should be noted that the color bar serves to highlight the accuracy of the lighting, exposure, and film professing when the artwork was photographed. This is much more difficult, if not impossible, to determine without the bar as a reference.

As pointed out, one does not print to ‘match’ the bar (unless, of course, the bar IS your artwork!). A cast in the neutral patches of the bar indicates the overall accuracy of the transparency in terms of lighting, exposure and processing and can be reasonably corrected to neutral using this information.

-Les De Moss

DigiGraphics LLC

___________________________________________________________________________

Date: Fri, 04 Jun 2004 12:06:38 -0400

From: “Iliah Borg”

Subject: Re: use of color bars in flat art repro

I think the biggest culprit is

that the dyes in the

printed color bars

actually, Q13 is not printed

Best regards,

___________________________________________________________________________

Date: Fri, 04 Jun 2004 11:34:59 -0400

From: “Iliah Borg”

Subject: Re: use of color bars in flat art repro

Values for Q-13 grey patches are at

http://www.pochtar.com/gamut_view/gamma.xls

I hope my math is accurate

Best regards,

___________________________________________________________________________

Date: Fri, 04 Jun 2004 13:05:15 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

It all depends on expectation.

Having no original to reference you must assume the transparency is the target to reach, not the color bars or gray scale. My expectation would be that the client has given me a color corrected tranny and wants to match that. Unless of course the clients says the tranny is off, then maybe use the color bars and grayscale to correct to.

If you have no other reference point, as I mentioned, a color bar is

of some use in removing cast and getting the file to look somewhat more

like the original.

I think your assumption is wrong, the color bars were not the intended object to be color corrected. Using color compensating filters to color correct the original transparency to match the artwork will adversely effect the color bars and gray scale so if you color correct them, the artwork will become un-corrected.

Of course correcting to the original art is the best way to go if possible.

Right!

Sometimes there is NO access to the art and the goal would only be to

achieve pleasing color which will never be compared to anything else.

Wrong. Having worked with hundreds of artists, they are trying to reproduce their artwork as closely as possible. Assuming we’re talking about Limited Editions Prints or the like here. Sometimes the original will hang close to or be available to compare to latter. You’re idea of pleasing color will differ from the artists in most cases.

Using a color bar is not a perfect scenario, but it’s better than nothing.

True, but I think the basic assumption, unless told otherwise, is that the art as it appears on the transparency is correct. As in my original post: using the gray scale to gage the highlight and shadow values is correct. The only other use the color bars and gray scale serve is to determine if the transparency has been exposed properly. If the printer matches the art on the tranny he’s doing his job. If the artist comes back and complains that the print doesn’t match the original explain you went by the information provided, the transparency. If you match the color bars you won’t have a leg to stand on. More importantly is to talk with the client, we do not need to make decisions like this in a vacuum.

Respectfully,

Tom Kaczmarek

___________________________________________________________________________

Date: Fri, 04 Jun 2004 13:49:00 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

Lanny Lathem wrote:

I think this is a great tool to have in the image.

I would agree, it is a great tool, but only as a rough reference for limited information.

I would neutralize the highlight, the shadow and the midtone on the

grayscale. This will take you toward a match of the artwork assuming you

don’t have the origninal art work with which to evalute your image.

I would disagree. The color bar and gray scale are a different medium as the artwork in most cases and in many cases reproduce differently on the transparency. Artists paints and mediums often have whiteners and brighteners that may fluoresce in the UV and IR ranges, wavelengths that are not visible to us but are to the film. These inconsistensies make using the included references invalid. Oils reproduce differently than acrylics, watercolors or any other medium. Watercolors can be very problematic as well due to the layering of the translucent colors, Ektachrome and Fujichrome, or “chrome” film in general is usually more sensitive to blue and the layering of a yellow over blue may be result in a different color on film than blue over yellow. Also, the photographer shooting the artwork is not color correcting toward the color bars or grayscale, he should be trying to match the art as closely as possible.

I would then check the highlight and shadow (assuming there is a highlight)

in the ‘live’ image area itself to make sure those adjustments didn’t do

someting undesirable due to uneven lighting when the photo was taken.

This is the goal, a correctly colored artwork. “Uneven” lighting during the photographic reproduction might not be evident unless you have the original artwork to compare to. If you do, correct to that in the first place. If the transparency has a noticeable density shift caused by unbalanced lighting the photographer didn’t do his job, have it re shot if possible. If this isn’t possible, warn your client that the reproduction you’re doing won’t match as closely as it could because the transparency isn’t optimal. A chain is only as strong as it’s weakest link and the chain in reproduction work is not very strong to begin with, everyone needs to do the best job possible, especially the photographer if the original will not be available when doing color critical reproduction.

My 2 cents worth anyway. Humbly,

Tom Kaczmarek

___________________________________________________________________________

Date: Fri, 4 Jun 2004 13:19:41 -0700

From: “R. S. Davidson”

Subject: Re: use of color bars in flat art repro

Many photographers have been using the Q-13 and Q-14 Color Patches and Grayscale as an “absolute” requirement when photographing (flat) art. Yet many of the respondents seem to have some “issues” with the Kodak Q-13 and Q-14. How about the Macbeth Color Checker and Mini Color Checker?

1 - Is it “better” than the Kodak Q-13/Q-14?

2 - What are the values for the patches ... does anyone have any definitive values?

Kodak Q-13 URL

http://av.rds.yahoo.com/_ylt= AgelJskVvBw68hd_RfGEKreFBqMX;_ylu= X3oDMTBwMjNqdWVsBHBndANhdHdfd2ViX3Jlc3VsdARzZWMDc3I-/SIG= 127d1rkm1/**http%3a//www.tiffen.com/KPA%2520Quality%2520Control%2520Listl802.PDF

Macbeth Color Checker URL

http: //www.gretagmacbeth.com/index/products/products_color-standards.htm

http: //www.gretagmacbeth.com/index/products/products_color-standards/products_colorchecker-charts/products_colorchecker.htm

Also ... How about using something like the reflective IT8 Target in the image frame?

___________________________________________________________________________

Date: Fri, 4 Jun 2004 20:51:20 -0400

From: John Castronovo

Subject: Re: use of color bars in flat art repro

From: “R. S. Davidson”

Also ... How about using something like the reflective IT8 Target in the

image frame?

I’ve actually done this with transparencies of flat art. We scanned the chrome and made a profile from the IT8 that was in the shot. When I assigned that profile to the scan rather than the usual scanner profile, the result was a highly accurate representation of the art with the only errors coming from certain pigments in the art that reflect a great deal of infrared light. These are paints that would be a problem anyway and the reason why we can’t simply rely on matching scales or targets alone.

This approach isn’t much different than profiling a digital camera when you think about it - the profiling software doesn’t know that film and a drum scanner were involved. To be useful, however, all future shots must be shot and scanned in the same way.

john c.

___________________________________________________________________________

Date: Fri, 04 Jun 2004 21:23:39 -0400

From: “Iliah Borg”

Subject: Re: use of color bars in flat art repro

2 - What are the values for the patches ... does anyone

have any definitive values?

Target values I use are here:

http://www.pochtar.com/s/GMB_CC.tif

(in Lab, 44kbytes)

Best regards,

___________________________________________________________________________

Date: Fri, 04 Jun 2004 17:05:50 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

Many photographers have been using the Q-13 and Q-14 Color Patches

and Gray scale as an “absolute” requirement when photographing (flat)

art. Yet many of the respondents seem to have some “issues” with the

Kodak Q-13 and Q-14.

The issue is not with the actual targets themselves. The fact is that the targets are altered by the process of correcting the color of the image of the original artwork. It is also problematic that the art reproduces differently than the targets due to many reasons. With the target and artwork reproducing differently, the targets are rendered useless for doing any color correction during the prepress cycle. BTW, it has not been pushed by the photographers, these use of these targets has been pushed by the printing industry who wants the photogs to include these.

How about the Macbeth Color Checker and Mini Color Checker?

They will not be any better. It’s not the target but the artwork that throws in the monkey wrench.

Printers need to realize this and base color corrections on the artwork itself.

Tom Kaczmarek

___________________________________________________________________________

Date: Sat, 5 Jun 2004 07:33:20 -0400

From: John Castronovo

Subject: Re: use of color bars in flat art repro

People should realize that the value of the target isn’t that it represents the painting, but that it represents the LIGHTING. As such, when the gray scale is correct, the best AVERAGE color correction has been achieved.

Of course, there are going to be colorants in the original that will be off color, sometimes dramatically so, and those have to be handled on a case by case basis. However, we have to begin with the best overall correction, and that happens when the gray scale is correct.

john c.

___________________________________________________________________________

Date: Sat, 5 Jun 2004 09:05:36 -0400

From: Dragonfly Imaging & Printing

Subject: Re: use of color bars in flat art repro

Agreed jc

I find that once I have corrected for the lighting, I am much closer to the ‘original’.

This makes color adjusting easier for the artwork.

Only the grayscale seems to be of much use in this context.

John Toles

http://www.dragonflyprinting.com/

http://www.dragonflygallery.ca/

___________________________________________________________________________

Date: Sat, 05 Jun 2004 09:22:09 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

jc castronovo wrote:

People should realize that the value of the target isn’t that it represents

the painting, but that it represents the LIGHTING. As such, when the gray

scale is correct, the best AVERAGE color correction has been achieved.

John, I would have to disagree. Even on average, neutralizing the gray scale will not reproduce the best average color correction of the art. It is not only specific colors that reproduce differently but different mediums on the whole that reproduce differently from the color bars and gray scales. If we just had to color balance the film, as you indirectly state, the job of the photographer would be easy. Color balance film, then shoot any original since only the gray scale will be corrected and used to pull the first proof.

However, we have to begin with the best overall correction, and

that happens when the gray scale is correct.

The gray scale hasn’t been corrected to to begin with. What actually happens, during photography, is that each original artwork will be color corrected individually, sometimes with a radically different combination of CC filters. Some other methods used to reproduce art is using cross polarized lighting and using a separate bump flash to control contrast. In fact, by adding different CC filters to the bump flash a subtle shift in color can be introduced from the shadows to highlights, what does that do to the gray scale?

All these corrections are solely aimed at reproducing the art, and will create reference materials that do not represent the lighting or any other element of shooting the original. All I’m trying to get across is that the aim of the printer should be to correct toward the image of the original. I’ve seen too many proofs that a printer made assuming the reference materials should be correct, thus causing unnecessary rounds of corrections. If the printer makes the assumption the image is correct some of those corrections can be eliminated.

Does it make sense to even include the references in the transparency? Probably not, but everyone wants to see them even though they’re worthless in color correction. They do show the accuracy of the exposure in reproducing the delineations of the gray scale from black to white, not the intended neutrality of the lighting conditions. The Color bars are completely inaccurate for color corrections. With this said will you get a customer that likes the proof if you correct toward the gray scale, pretty sure you would, but not many. At least not the ones who want accurate reproductions and artists can be a very picky bunch. Unless told otherwise by your client, a printer must assume the intended color of the original is accurately represented by the image on the transparency, not a color bar or gray scale. Sometimes you have to place your trust in the person making the repro trans, they should be color correcting toward the original, not the reference materials.

Respectfully,

Tom

___________________________________________________________________________

Date: Sat, 05 Jun 2004 15:50:30 -0400

From: “Iliah Borg”

Subject: Re: use of color bars in flat art repro

one should take into account that grey scale is not perfect grey to begin with :)

I measured many of them w/spectrophotometer.

Best regards,

___________________________________________________________________________

Date: Sat, 5 Jun 2004 18:37:57 -0400

From: John Castronovo

Subject: Re: use of color bars in flat art repro

I think we’re in total agreement that the appearance of the art takes precedence over the gray scale. A “correct” gray scale isn’t necessarily a neutral one. The scale is only meant to help judge the overall exposure, and as such, I feel that it’s useful rather than irrelevant. How the scale reproduces tells me something about the exposure, but the photo is ALWAYS balanced for the art and not the scale. I shoot for one artist who’s work routinely requires 2/3 of an f stop more exposure than usual, and another who’s work requires 1/2 stop less. One requires flashing and the other doesn’t. One uses oils and lots of cobalt blue, and the other is acrylic of unknown origin. Therefore, the scales run all over the board, but they’re always useful, if only to tell me what’s happening in the paint vs. the process.

Also, if the photo should turn up at some point in the future and there’s nothing else to go on, the person who discovers it will have to determine what is real. If we’re talking about a faded transparency, then the scale may help reveal the nature of the fade. If we’re talking about a file capture, a scale can help confirm a profile. It’s the only known quantity, so it can help if we know how to treat it.

The gray scale is a tool that tells us something about the shot and NOT necessarily to be neutralized. In fact, I find that the scale will render slightly greenish compared to most corrected originals, and someone who has worked with them will know this and not try to balance it to neutral unless we’re copying B&W photos (which is what the scale is made of). It’s also good to see the shadows somewhat compressed, but the highlight end should never be lost. To achieve this when I shoot chromes, I cross polarize, flash the film and pull the processing 1/3 stop while over-exposing to compensate for the pull and the use of cc filters. Digital scans are similar in their adjustments.

Extremely dark old masters are the hardest things to copy and a real case where the gray scale will be far off from the best corrected image. Again, the person correcting the scan has to intuit a little about the photography to know when and how to ignore the scale.

___________________________________________________________________________

Date: Sun, 6 Jun 2004 15:08:29 -0300

From: “Cicero Rodrigues”

Subject: Re: use of color bars in flat art repro

Finally I’ve read something about the use of reference scales that is according my own experience as a photographer. Especially Mr. Tom Kaczmarek posts.

For years I’ve been trying to explain the limits (if not the useless) of the reference scales to artists, art dealers and press house people. Everybody is so convinced of the importance of using reference scales that I was about to give it up.

Thanks Colortheory List to give my confidence back :)

Best regards,

Cícero Rodrigues

www.cicerorodrigues.com.br

___________________________________________________________________________

Date: Sun, 6 Jun 2004 14:32:50 EDT

From: Dan Margulis

Subject: Re: use of color bars in flat art repro

John Castronovo writes,

I think we’re in total agreement that the appearance of the art takes

precedence over the gray scale. A “correct” gray scale isn’t necessarily a

neutral one. The scale is only meant to help judge the overall exposure, and

as such, I feel that it’s useful rather than irrelevant. How the scale

reproduces tells me something about the exposure, but the photo is ALWAYS

balanced for the art and not the scale.

The responses to this question among the most sophisticated I’ve seen on the list. John is right in what he says above and also right in pointing out the consensus. I believe that even three years ago we would have seen more controversy.

Just to clarify a couple of matters: first, granted that one tries to find neutrality in the image and work with that rather than the grayscale, sometimes that information isn’t available and we have to use the grayscale as a crutch. In doing that, we *don’t* pay attention to how light or dark the gray is. There’s no inference at all that the darkness values should be close—changing luminosity values is the main goal of color correction.

Also, while the color patches aren’t nearly as valuable as the gray, they aren’t worthless, especially the red patch. All these swatches have a patch of a bright red, something that in principle should convert to 0c100m100y. Whether it actually does that is almost entirely irrelevant, because that color rarely exists in photographs and in any event it might be better treated as an out of gamut color for purposes of correction. However, it can give an important hint about the rest of the picture.

Regardless of what colorspace you work in, you can set up the right side of the info palette to read CMYK values. In principle, that red patch should read equal values of magenta and yellow (it wouldn’t necessarily read equal blue and green in RGB, which is why we need to look at the CMYK). If the yellow is significantly higher than the magenta, it can indicate a cast that would affect all dark reds.

Dan Margulis

___________________________________________________________________________

Date: Sun, 06 Jun 2004 19:57:53 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

John Writes:

I think we’re in total agreement that the appearance of the art takes

precedence over the gray scale. A “correct” gray scale isn’t necessarily a

neutral one. The scale is only meant to help judge the overall exposure, and

as such, I feel that it’s useful rather than irrelevant. How the scale

reproduces tells me something about the exposure, but the photo is ALWAYS

balanced for the art and not the scale.

Agreed! Have had many instances where someone tried to correct toward the gray scales, hence my “strong” feelings on this subject. Hope this thread makes it’s way into the right hands.

Respectfully,

Tom Kaczmarek

___________________________________________________________________________

Date: Sun, 06 Jun 2004 20:32:35 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

Dan writes:

Also, while the color patches aren’t nearly as valuable as the gray, they

aren’t worthless, especially the red patch. All these swatches have a

patch of a bright red, something that in principle should convert to 0c100m100y.

Whether it actually does that is almost entirely irrelevant, because that color

rarely exists in photographs and in any event it might be better treated as an out

of gamut color for purposes of correction. However, it can give an important

hint about the rest of the picture…

If the yellow is significantly higher than the magenta, it can

indicate a cast that would affect all dark reds.

Why key on the red, wouldn’t the Green patch provide the same information (100c0m100y)? Does this information truly represent an unintentional color cast? Why would it be better treated as an out of gamut color, isn’t the intent to color correct toward the original where that red may actually exist? I’m probably not understanding a basic concept here, could you explain further?

Cheers,

Tom Kaczmarek

___________________________________________________________________________

Date: Mon, 7 Jun 2004 03:28:19 -0400

From: John Castronovo

Subject: Re: use of color bars in flat art repro

Right. Now for another thought, what about replacing the original gray scale

with a digital one once the scan and color adjustments are finished? Since

the scale next to the painting is used mainly by the photographer to help

him see what’s going on in the paints, it really serves little purpose after

the scan is done and it might be misleading, but a digital scale may have

great use down the line and it satisfies those who’d like to see one next to

the art. We’ve been doing this for artists who have us make LVT

transparencies from original digital art, but why not from real art as well?

I imagine that it would also be a good thing to have in case the profile tag

got stripped away too.

What should such a scale ideally look like? How many steps? Do we need the

tints of the solid colors? How about size information and profile

information right on the scale?

john c.

___________________________________________________________________________

Date: Mon, 7 Jun 2004 08:12:56 -0400

From: “John”

Subject: Re: use of color bars in flat art repro

Shooting conditions change, and lights change color with age. Color bars on

the original only assure consistency of the input capture, as the color bars

on the press sheet are there to assure consistency of the output of the ink

to paper. Although the final appearance of the reproduction does take

prefference, the color bars make everything we do to get to that point a

whole lot easier when you know you are starting from the same place

everytime. Yes certain pigments do not behave well at times, and yes certain

whiteners in the canvas can tend to mis-behave at times, but as a whole, the

color bars puts most of the image where it belongs right off the bat,

requiring very little work.

There seems to be a trend in the industry these days to let the input

controls of image capture, be it digital camera, scan backs, or scanners, go

by the wayside, using the theory of lets just grab the data however we can,

and as long as it’s in the ballpark, and we’ll fix it later. Sometimes you

can, sometimes you can’t.

Any thoughts from the group? How do you judge how far off, is too far off,

before rescanning or reshooting something?

John Rawlins

___________________________________________________________________________

Date: Mon, 07 Jun 2004 07:50:58 -0400

From: Lanny Lathem

Subject: Re: use of color bars in flat art repro

on 6/5/04 7:33 AM, jc castronovo wrote:

People should realize that the value of the target isn’t that it represents

the painting, but that it represents the LIGHTING. As such, when the gray

scale is correct, the best AVERAGE color correction has been achieved.

Thank you. This is what I was getting at.

It’s a flawed tool at best, but sometimes the only one available. Trusting

large format transparencies to be ‘color correct’ has gotten me in more

color trouble than not.

Lanny Lathem

Bennett Graphics

___________________________________________________________________________

Date: Mon, 7 Jun 2004 08:25:38 EDT

From: Dan Margulis

Subject: Re: use of color bars in flat art repro

Tom writes,

Why key on the red, wouldn’t the Green patch provide the same

information (100c0m100y)? Does this information truly represent an

unintentional color cast? Why would it be better treated as an out of

gamut color, isn’t the intent to color correct toward the original where

that red may actually exist? I’m probably not understanding a basic

concept here, could you explain further?

Things we think of as “green” are almost invariably heavily biased toward yellow. Nothing that’s particularly common has equal amounts of cyan and yellow, therefore this patch has little relevance.

By contrast, any deep reddish color is likely to have approximately equal amounts of magenta and yellow. The deeper the color, the more likely it is to favor magenta rather than yellow, but certainly it’s possible that the yellow is higher.

If the picture features a person wearing a shirt that averages 25c85m95y, and we don’t have anything else to go by, I’m suspicious, because that’s an uncommon color. Yet it might be right, so I won’t change it. If there’s a red swatch in the image that’s got 10 points more yellow than magenta in it, it’s a different story—now I’ll make the change.

In greens, we never get to that point. If it’s an avocado that measures 95c25m85y, I don’t have to be suspicious or to refer to any swatch, because this is a known wrong color.

As for the out of gamut issue, 0c100m100y isn’t a color that occurs in natural photographs with any regularity. But, there may be things that are that red in the RGB file, but lighter or darker. Since 0c100m100y is by definition the reddest possible thing in CMYK, those other reds are out of gamut. Setting them to 0c100m100y may be the best compromise. Or, whenever the picture is loaded with deep, rich reds, it’s conventional to set some part of it to be 0c100m100y as otherwise there may not be enough variation in the reds to satisfy. In either of these cases we are basically treating the hypothetical 0c100m100y equivalent in the RGB file as an out of gamut color, because after conversion, we’re assigning 0c100m100y in CMYK to something else.

Dan Margulis

___________________________________________________________________________

Date: Mon, 07 Jun 2004 09:32:58 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

jc castronovo wrote:

Right. Now for another thought, what about replacing the original gray scale

with a digital one once the scan and color adjustments are finished?

Interesting idea. Could include an IT8 target and include the target data along with the file. But this would just essentially duplicate the embedded profile. On the other hand, developing a reference bar with the target info included beneath each swatch might be a more useful way to go. My first thought would be to duplicate existing references as not to confuse things. Actually scanning the materials and forcing the numbers of each swatch should do. The actual numbers were supplied as a link in this thread:

For a 14 step scale maybe use this:

Q-55 KODAK Control Scale, T-14 - 14 step Kodak Gray Scale with a density range of 0.04 to 2.05 in density

increments of 0.15 (a two step shift is equal to a change of 1 f-stop or a doubling/halving of the exposure).

We’ve been doing this for artists who have us make LVT

transparencies from original digital art, but why not from real art as well?

I imagine that it would also be a good thing to have in case the profile tag

got stripped away too.

“Real” art? That should start a real flame war. ;-o

What should such a scale ideally look like? How many steps? Do we need the

tints of the solid colors? How about size information and profile

information right on the scale?

All good ideas! Might as well add the tints and other color targets, I can’t think of a good reason not to. Just recreate digital Q13/14 targets, less confusing in the long run. Adding originals size, medium, copyright, date created, artist and contact info would all be useful info to include as well. Including embedded profile might or might not be a good idea, could become confusing during production or if the file ever is reassigned a different profile.

Cheers!

Tom

___________________________________________________________________________

Date: Mon, 07 Jun 2004 10:28:07 -0400

From: Tom Kaczmarek

Subject: Re: Re: use of color bars in flat art repro

John Rawlins wrote:

Shooting conditions change, and lights change color with age. Color bars on

the original only assure consistency of the input capture, as the color bars

on the press sheet are there to assure consistency of the output of the ink

to paper. Although the final appearance of the reproduction does take

prefference, the color bars make everything we do to get to that point a

whole lot easier when you know you are starting from the same place

everytime…

The point here is that you will not be starting from the same place every time. The color corrections are unique to each original and not solely dependent upon the lighting conditions and the film’s emulsion variances.

Yes certain pigments do not behave well at times, and yes certain

whiteners in the canvas can tend to mis-behave at times, but as a whole, the

color bars puts most of the image where it belongs right off the bat,

requiring very little work.

Since the reference materials and the original are of two different mediums they inherently reproduce differently, not just the mis-behaved colors. Since a color correction is global to the original, not the color bars, the targets may be well off the mark of their intended colors and in possibly a different direction than that of the artwork itself. Hence, correcting initially to the reference materials may not be correct and will cause unnecessary work later.

There seems to be a trend in the industry these days to let the input

controls of image capture, be it digital camera, scan backs, or scanners, go

by the wayside, using the theory of lets just grab the data however we can,

and as long as it’s in the ballpark, and we’ll fix it later. Sometimes you

can, sometimes you can’t.

That may well be, but you have to keep in mind the intent of the capture. I will tell a service provider doing a scan for me not to color correct or apply USM. I would much rather have a scan with unadulterated color information so I can be the controlling factor in how my image will look. Looking at the histogram is the first thing I do when I get the scan, preferably when I pick it up and before I pay. As long as the scan has the proper density and the tonal range accurately depicts the original I’m happy. This is my personal work flow as a “creative type” and may, will, be radically different than say a printer’s.

Any thoughts from the group? How do you judge how far off, is too far off,

before rescanning or reshooting something?

How far is too far? Experience will need to be your guide here I think. Never thought of quantifying or measuring this, if it’s even possible. Customer expectations will also govern this I’m afraid. Can anyone else answer this?

Cheers,

Tom

___________________________________________________________________________

Date: Mon, 7 Jun 2004 21:29:51 -0400

From: John Castronovo

Subject: Re: Re: use of color bars in flat art repro

From: “John”

Yes certain pigments do not behave well at times, and yes certain

whiteners in the canvas can tend to mis-behave at times, but as a whole,

the

color bars puts most of the image where it belongs right off the bat,

requiring very little work.

The “safest” work I’ve ever shot is done by an artist who mixes all of his colors out of only four tubes of paint - that’s three primaries plus white. In this way, there is only one way to mix any color, so his paintings are always well behaved under the lights. This guy understands that the reproduction becomes the art, so he’s learned to paint with the reproduction process in mind.

Most artists aren’t aware of these concerns. For example, we see situations where two browns were mixed with different pigments - and so they reproduce as two different colors even though they appear the same to the eye. Entire paintings can take on color casts that have nothing to do with the way they appear to the eye because of the palette that the artist used to mix the colors.

For reasons such as this, gray scales can’t be trusted to tell us the color of the painting, but they’re very usuful to the photographer to tell him what’s going on in the painting vs. the lighting and the exposure. To be sure, John, I always begin with a good gray scale when I shoot, and then make adjustments from there to suit the art. I think we’re all in agreement on that. The real question is whether one can trust the scale later on in the reproduction process after the best match to the original has been made, and that depends on the original. Oftentimes you can’t.

Any thoughts from the group? How do you judge how far off, is too far off,

before rescanning or reshooting something?

Ah, that’s entirely up to the customer - the budget and his sensitivity notwithstanding.

john c.

___________________________________________________________________________

Date: Mon, 07 Jun 2004 15:12:19 -0400

From: Tom Kaczmarek

Subject: Re: use of color bars in flat art repro

Dan writes:

In greens, we never get to that point. If it’s an avocado that measures

95c25m85y, I don’t have to be suspicious or to refer to any swatch,

because this is a known wrong color.

Makes sense that greens, having a normally higher amount of yellow, would be a poor reference for a possible color cast. Thank you!

As for the out of gamut issue, 0c100m100y isn’t a color that occurs in

natural photographs with any regularity. But, there may be things that are that red

in the RGB file, but lighter or darker. Since 0c100m100y is by definition the

reddest possible thing in CMYK, those other reds are out of gamut. Setting

them to 0c100m100y may be the best compromise. Or, whenever the picture is loaded

with deep, rich reds, it’s conventional to set some part of it to be

0c100m100y as otherwise there may not be enough variation in the reds to satisfy. In

either of these cases we are basically treating the hypothetical 0c100m100y

equivalent in the RGB file as an out of gamut color, because after conversion,

we’re assigning 0c100m100y in CMYK to something else.

Interesting, I understand this as it would apply to a photographic image but how about fine art reproduction? I guess since we’re dealing with colors that may not be possible to be reproduced faithfully in CMYK, this might be a way to achieve a more pleasing compromise to at least help define colors that fall out of gamut. Wouldn’t this also be handled by the rendering intent set to Relative Colorimetric as opposed to Perceptual or Absolute Colorimetric? Getting a great education here, also makes me reevaluate things I have taken for granted or how to do things better. Thanks for the good feedback!

Respectfully,

Tom

___________________________________________________________________________

Date: 8 Jun 2004 07:55 +0200

From: Jeff Suckow

Subject: Re: Re: use of color bars in flat art repro

What do you mean when you all say to match the art ? Match the art to what ? the monitor ? to the final print ?

Jeff Suckow

___________________________________________________________________________

Date: Tue, 8 Jun 2004 08:01:01 -0400

From: "John"

Subject: Re: use of color bars in flat art repro

On Mon, 07 Jun 2004, "Thomas P. Kaczmarek" wrote:

The point here is that you will not be starting from the same place

every time. The color corrections are unique to each original and not

solely dependent upon the lighting conditions and the film's emulsion

variances.

Ummmm... not really. If you do repetitive artwork reproduction for the same artist, that is using the same subtrates and pigments everytime... then the color correction is close to be the exact same moves, each time, everytime, assuming your baseline starting point is identical each time, ie: the color chart. Digital capture does not have to contend with film or processing variances, so that part is out of the equation.

Since the reference materials and the original are of two different

mediums they inherently reproduce differently, not just the mis-behaved

colors. Since a color correction is global to the original, not the

color bars, the targets may be well off the mark of their intended

colors and in possibly a different direction than that of the artwork

itself. Hence, correcting initially to the reference materials may not

be correct and will cause unnecessary work later.

We do repetitive art reproduction for about a dozen different artists. Yes they all use varied mediums and materials, but shooting something in balance means for the whole, its in balance. If we have to later color correct for the problem colors that themselves happen to reproduce out of balance, that's fine. But its' fewer of the colors to deal with than more of them. If we really didn't want to use the targets, I guess that's fine too. We could balance on the artwork itself doing manual highlight/ midtone/ shadow/ neutal selections. At the time of capture, we're going for the basics that won't change, and can't change, if we are going to be putting ink on paper. Most of the color matches fall in line, and the "problem children" color will require special attention. In no way are you going to get everything in one shot. You're always going to have to do something. It is just our opinion that shooting by using the targets gets us so much closer, and saves time later.

John Rawlins

___________________________________________________________________________

Date: Tue, 08 Jun 2004 08:05:52 -0400

From: Lanny Lathem

Subject: Re: Re: use of color bars in flat art repro

on 6/8/04 1:00 AM, Jeff Suckow wrote:

What do you mean when you all say to match the art ? Match the art to

what ? the monitor ? to the final print ?

I never presumed one could match the art using a color bar for gray balance. I said you could move toward the art if the original art was not available. I'm not talkin about fine art reproduction here as that would be folly without the original as a reference. I was referring to a situation where a transparency of an art piece is provided and it is obvious from the color bar that there are issues. In an attempt to provided something 'somewhat' more like the unseen original the color bar can be used as a blind reference to 'improve' the image.

Sorry for not being more clear.

Lanny Lathem

Bennett Graphics

___________________________________________________________________________

Date: Tue, 08 Jun 2004 08:16:13 -0400

From: Tom Kaczmarek

Subject: Re: Re: use of color bars in flat art repro

Jeff,

We've been referring to the repro transparency mainly as what to match proof/print to if the original is not available at the time of printing. Hope that clears thing up.

Cheers,

Tom

___________________________________________________________________________

Date: Tue, 08 Jun 2004 07:35:56 -0400

From: Jim Rich

Subject: Re: Re: use of color bars in flat art repro

On 6/8/04 1:00 AM, Jeff Suckow wrote:

What do you mean when you all say to match the art ? Match the art to

what ? the monitor ? to the final print ?

Jeff,

Match the art usually means make the reproduction look like the original piece of art.

Jim Rich

___________________________________________________________________________

Date: Tue, 8 Jun 2004 10:48:15 EDT

From: Dan Margulis

Subject: Re: use of color bars in flat art repro

Tom writes,

Interesting, I understand this as it would apply to a photographic image

but how about fine art reproduction? I guess since we're dealing with

colors that may not be possible to be reproduced faithfully in CMYK,

this might be a way to achieve a more pleasing compromise to at least

help define colors that fall out of gamut. Wouldn't this also be handled

by the rendering intent set to Relative Colorimetric as opposed to

Perceptual or Absolute Colorimetric? Getting a great education here,

also makes me reevaluate things I have taken for granted or how to do

things better. Thanks for the good feedback!

There are two kinds of "fine art" reproduction and I'm not sure which one you're talking about. Type 1 is, you're trying to output something at large size to hang on somebody's wall. Type 2 is, you're trying to make a much smaller reproduction of some piece of fine art for appearance in a catalog or book or magazine or whatever. Both types are difficult to handle but the considerations for each are entirely different.

With Type 1 you have to spend a lot of time analyzing the original to see what colors, if any, are going to be impossible to reproduce in CMYK. That examination will govern how the conversion into CMYK will be done. There won't be one method that works for all such images.

With Type 2 it's much more free-form. A large piece of artwork hanging on a wall invites careful scrutiny and the artist can get away with some extremely subtle effects that are entirely lost when reproduced at a smaller size. So, we try to imagine what subtleties the viewer is seeing in the larger piece, and then exaggerate them in the smaller reproduction. Minor variations in color are not the factor that they are in Type 1; the key will be retaining perceived detail.

Dan Margulis

Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.