Colour Management – Aaarrgghh

[SteveH]

OK so an extreme reaction but the two phrases seem to go quite well together.

As a new comer with a new 3880 printer and Spyder 3R Print Studio calibration kit (and PC Win 7 64bit) I have a heap of questions I hope someone can answer. I will split them into a couple of posts.
Spyder: Having just done a print calibration I have options to view results under a variety of options:

Saturation
Perception
Relative colorimetric
Absolute calorimetric

All of the post-calibration on-screen example prints look washed out and their colour range highly compressed. I chose “Relative colorimetric” and the print results look better than the screen but I was assuming they should look the same as the screen.
BTW my monitor is also calibrated with Spyder3 elite.
When I choose to print the Spyder test print, having calibrated my printer, should I still choose all the same Epson driver settings, especially “no colour management” or should I be loading my new printer profile somewhere in the Epson print dialogue?

cheers
Steve

You think I’m dumb now – just wait ‘til my next questions :confused:

[Peter De Smidt]

Datacolor recommends "saturation" with the profiles made with your equipment.

"When I choose to print the Spyder test print, having calibrated my printer, should I still choose all the same Epson driver settings, especially “no colour management” or should I be loading my new printer profile somewhere in the Epson print dialogue?"

By "Spyder test print" do you mean the swatch pages printed by the Spyder Print software? If so, then you use the exact print driver settings that you are going to use to print images. You should definitely choose "no color management" in the print driver, and make sure to pick the best media and quality settings.

Here is an overview of color management:
1) Make your input profile, which can be made with an IT8 target, a scanner, and profiling software. You can also make profiles for digital cameras, although many people don't.

2) Make your output profile with the Spyder3Print. Use the exact settings that you want to use to make a print. Make sure to turn off all color management in the print driver.

3) You must now convert the file from the input profile to an "editing" color space for work in photoshop. In the case of a scanned file with an icc attached, when bringing the file into photoshop you convert the file from the input profile to the editing profile, the latter of which is usually Adobe 98, Ektaspace, or Profoto. In the case of a file from a raw converter, these usually tag the image directly in an editing space. As such, as long as you pick the proper color space in your converter, you won't have to convert the file when bringing in to photoshop.

3) Edit the file in photoshop. If you want to see a soft proof of how the image will look when printed, go to View>Proof Setup>Custom. Pick the output profile that you made with your Spyder3Print. This only changes how the image is displayed in photoshop. It doesn't change the file. You can toggle it on and off. You can also play with rendering intents to see if a different gives you a better look.

4) When you go to print, choose "Photoshop manages colors" in the Photoshop print dialogue. Make sure to specify the printer/paper/settings icc file in the appropriate place.

5) Make sure to keep the exact same settings in the print driver that you used to print on the Spyder3Print swatch pages.

6) View the print under some well-thought-out viewing light.

[neil poulsen]

I would add, don't make any selections in the print driver that intentionally adjust color. That is, a lot of drivers enable the user to adjust color right from driver itself. All of this type of capability should be turned off, set to zero, etc., when you're printing the target and when you're printing photographs. Certainly, you don't want any setting in the print driver that "automatically" adjusts color.

What version of Photoshop are you using?

[Lenny Eiger]

OK so an extreme reaction but the two phrases seem to go quite well together.

As a new comer with a new 3880 printer and Spyder 3R Print Studio calibration kit (and PC Win 7 64bit) I have a heap of questions I hope someone can answer. I will split them into a couple of posts.
Spyder: Having just done a print calibration I have options to view results under a variety of options:

Saturation
Perception
Relative colorimetric
Absolute calorimetric

All of the post-calibration on-screen example prints look washed out and their colour range highly compressed. I chose “Relative colorimetric” and the print results look better than the screen but I was assuming they should look the same as the screen.
BTW my monitor is also calibrated with Spyder3 elite.


None of these options mean what you think they mean. Leave it on Relative Colorimetric. This has to do with what the driver does if it finds a pixel outside the color space. In some cases, it maps the whole color space down to compensate (Perceptual). Relative moves only the one color back into the space so it can print it. That's the way to go. Saturation is for graphics and such, not photos...

Just because you did a printer calibration (and a monitor calibration) it doesn't mean your print should - or will - match the screen. Not unless you are using a very limited color space, like cmyk, and a viewing booth. All the rest of us look at the print and make adjustments to get what we want. We slowly tune our eyes to the monitor and printer. (Just like looking at a darkroom print and knowing how much it will dry down - as in experience.) The profiling just makes it more predictable.

Lenny

EigerStudios

[neil poulsen]

I agree.

I use relative colorimetric almost exclusively. It surprises me that your software even suggests saturation, because it's reserved for things like business graphics that benefit from enhanced saturation. If I'm not mistaken, saturation preserves saturation, sometimes at the expense of hue.

Perceptual is often touted for photographs and is used widely. But as Lenny Eiger suggests, it condenses the entire print gammut, versus just those colors that are outside the target gamut. Sometimes, that can be beneficial. So if there's a question between perceptual and relative colorimetric, use the one that looks the better. But for me, it's a rare occasion that I'll choose perceptual over relative colorimetric.

[Brian Ellis]

I don't think there is one single best setting for all papers and all images. Assuming you're soft proofing (View > Proof Set-up > Custom) try relative colorimetric with the black point compensation box checked and then try perceptual with the black point compensation box unchecked (always leave "Preserve RGB Numbers" unchecked and always check "Simulate Paper Color"). Then in the Print window just use whichever of the two looked best in the soft proof (the difference between the two often isn't very great). I've never heard of using Saturation or Absolute Colorimetric for printing photographs.

I don't know that this is the only way way to do things or even the best way. When I was trying to get a handle on color management I read a lot of conflicting or at least confusing information. What I've suggested above is just what I settled on after doing quite a bit of reading and it's worked well for me for a pretty long time.

You might find this paper from Adobe useful http://www.adobepress.com/articles/article.asp?p=1315593&seqNum=7. You might also find Eric Chan's 3800 workflow useful. http://people.csail.mit.edu/ericchan/dp/Epson3800/index.html Although the Adobe paper deals with CS4 rather than CS5 and Eric Chan's workflow deals with the 3800 rather than the 3880 I think most of the information would still be useful for CS5 and the 3880 though I've never used a 3880.

I don't understand the question about the Spyder test sheet, presumably because I don't use Spyder to calibrate.

[Peter De Smidt]

Despite Datacolor's recommendation to use saturation, I usually check out the various rendering intents and use the one looks best to me. Usually that's relative colormetric, but I've been surprised now and again.

[SteveH]

Thanks guys
I did work out that "relative colourmetric" seemed to be the best option.
By test print I meant the Proof Sheet of about a dozen thumbnails that comes up after calibration thatyou can print to view things like skin tones, grey scales, bold primaries etc.

I don't have an IT8 target for the scanner and will be doing mainly B&W scanning.

One issue is that the personal work I am printing is primarily night work with high contrast areas and lots of areas of deep shadow. The test prints I have been making were not holding shadow detail or contrast anything like the image on the monitor, it wasn't even representative.

Having said that I did manage to whip off one print last night that if I turned the monitor brightness way down would probably look pretty close.

So maybe I'm getting closer. I will review all the excellent info here and have another go tonight.
cheers
Steve

[William Whitaker]

OK so an extreme reaction but the two phrases seem to go quite well together.

Not at all extreme if you're a pirate.

[SteveH]

Not at all extreme if you're a pirate.

I used to be a happy pirate.
Now I'm just a confused pirate.:(

All this assistance is starting to show a light on the horizon though.:)

cheers
Steve

[Lenny Eiger]

Thanks guys
I did work out that "relative colourmetric" seemed to be the best option.
By test print I meant the Proof Sheet of about a dozen thumbnails that comes up after calibration thatyou can print to view things like skin tones, grey scales, bold primaries etc.

I don't have an IT8 target for the scanner and will be doing mainly B&W scanning.

One issue is that the personal work I am printing is primarily night work with high contrast areas and lots of areas of deep shadow. The test prints I have been making were not holding shadow detail or contrast anything like the image on the monitor, it wasn't even representative.

Having said that I did manage to whip off one print last night that if I turned the monitor brightness way down would probably look pretty close.

So maybe I'm getting closer. I will review all the excellent info here and have another go tonight.
cheers
Steve

Steve, you are talking about black and white. Calibrating the monitor will get you a clean gray in the background and no more. You can not expect the monitor to look like the print. It simply doesn't work that way. Profiling the printer is a good thing, you can develop a profile that will get you a good gray ramp that is smooth from end to end. I would suggest you look at Cone's b&w ink and QTR for the best results, but for night time contrasty images, you should be fine with anything.

Stop focusing so much on color mgmt and look at the print. If its too light, make it darker... That's how it works...

Lenny
EiberStudios

[John Rodriguez]

As Lenny pointed out, you'll never get a PERFECT match between monitor and print. However, if you're really far off you may need to check your monitor calibration settings. Your software/puck probably walked you through setting the brightness, but you still have testing you can do with luminance, white point and gamma. They'll usually recommend a luminance setting of 120 for LCDs, but I get the best match at 80 on my Apple Cinema LCD. White point is generally best set to D50 for print work, but most software defaults to D65, so try changing it. Also, they'll usually recommend a gamma of 2.2 or possibly 1.8, however this is also something you can modify to get closer to your prints. Just make sure the print you're comparing your screen to is from a profiled device that you trust.

[Peter De Smidt]

The gamma of Pro Photo is 1.8, but aren't you working with a gray scale file? If so, you will be working with Gray Gamma 2.2, right? Why does ProPhoto even enter into the discussion?

If you're using Gray Gamma 2.2, then you want to profile to gamma 2.2 with your Spyder. D65 seems to be the standard choice these days.

[neil poulsen]

I think of color management as an aid, not as an end result. One can waste a lot of time, paper, and ink, if they're not color management. It's really tough to optimize a print by guessing what adjustments might be needed, if there isn't some degree of similarity between the print and the screen. So, it's worthwhile to color manage.

But a screen is e'missive, while a print is reflective. So, differences will always exist. As a result, printing has alwaus been an iterative process for me.

[SteveH]

Thanks Guys for all the helpfull info.
My Mechanics Institutes weekend shoot got rained out so I'm back at the 'pooter again trying to make prints.
Using settings suggested here I have managed to make a print from a section of an image that is 'almost' acceptable (still a bit flat). When I go into View - Proof Setup - Custom, to do a soft proof it looks almost OK. But if I tick the box "simulate paper colour" the screen image looks way too flat.
OK so I could just ignore this option and turn my screen contrast down a bit and work with that. My concern is (as with all of these noobie qu's) that I may not be doing things the best way and I want to start off with the best possible work flow.

Lenny
thanks for the advice. I want to try Cone Inks but it won't be any time soon. I want to go as far as I can with the standard Ink/paper combo first before I branch out. Also I read somewhere that Epson won't honour a warranty claim if you'ved used non-OEM inks. That gives me 12 months to push the limits of the Epson Inks.

thanks for all the help
Steve

[Brian Ellis]

. . . When I go into View - Proof Setup - Custom, to do a soft proof it looks almost OK. But if I tick the box "simulate paper colour" the screen image looks way too flat. . . .

The usual recommendation is to look away from the screen when you hit O.K. after checking the "simulate paper color" box. Then when you look back at the screen things don't look so drastically different as they do if you watch while its happening (Bruce Frazer's recommendation and Adobe's recommendation, not something I came up with on my own but it's something I do and it's much better than looking at the screen while switching to the soft proof).

You also have to understand how to use the soft proof. The idea is to make a duplicate of your original image for comparison, then edit the soft proof and try to make it look as much like the duped original as possible. Usually that involves an increase in contrast and saturation, maybe some dodging and burning. While it's true that as others have said you'll never get an exact match between monitor and print because of the different light sources, you should be able to get very very close. I almost never make have to make more than one print after soft proofing and editing.

[Preston]

Brian,

I took a look at an article about soft proofing from Computer Darkroom (http://www.computer-darkroom.com/softproof/softproof_1.htm). It says that edits should be applied to the original image so that it looks like the duplicate image with 'View Proof Colors' applied. I think we're saying the same thing, but in slightly different ways.

From the article:

"Make sure that you have the original image and NOT the duplicate selected. The duplicate will be used for reference; all edits will be applied to the original.

Choose the New Layer Set option form either the Layer menu or via the Layer Palette.

Name the Layer Set so that it relates to the media profile, e.g. Fuji Hunt SemiGloss "

One creates adjustment layers within this layer group in order to make original image look like the 'proof' image.

I do this with every image that I want to print, and it works nicely.

SteveH: Give the article I linked to a read. It deals with color management in PS and hopefully will be of help to you. (Note: The article is for CS2, but is easily transferable to later versions.)

--P

[Lenny Eiger]

Lenny
thanks for the advice. I want to try Cone Inks but it won't be any time soon. I want to go as far as I can with the standard Ink/paper combo first before I branch out. Also I read somewhere that Epson won't honour a warranty claim if you'ved used non-OEM inks. That gives me 12 months to push the limits of the Epson Inks.
thanks for all the help
Steve

One doesn't need to overwhelm one's self with too many things at once. However, the bit about the warranty is baloney. It's basically illegal to do so, Epson can not void a warranty for putting 3rd party inks in a printer. That's the law. I remember the time when I had an Epson tech out here fixing something, and he said, oh, we'll see tons of grit in this damper when I replace it. Of course, that wasn't true at all when he got it off... it was perfectly clear.

Lenny
EigerStudios

[Tim k]

See if this helps. http://super.nova.org/DPR/ColorManagement/

[SteveH]

One doesn't need to overwhelm one's self with too many things at once. However, the bit about the warranty is baloney. It's basically illegal to do so, Epson can not void a warranty for putting 3rd party inks in a printer. That's the law. I remember the time when I had an Epson tech out here fixing something, and he said, oh, we'll see tons of grit in this damper when I replace it. Of course, that wasn't true at all when he got it off... it was perfectly clear.

Lenny
EigerStudios

Hi Lenny
I'm sure you're right, it's just something I read (maybe here a while ago? I forget).
But generaly I think that I'm expecting a steep learning curve anyway and it may take 12 months to decide wether I want to dedicate the printer solely to B&W.
The problem partially is just going to be getting my 'eye' in. Unfortunately, unlike you guys in the US and parts of Europe, I rarely have a chance to see original prints by master printers.
I have my own silver prints to compare to and a few inkjets I've bought from other local photographers but apart from that the best I can compare to is usually the repros in B&W magazine and the books on my bookshelf.
Anyway, I'll keep working - the proofs I'm generating now are looking better now that my monitor contrast is turned down.
thanks for all your help
cheers
Steve

[Lenny Eiger]

Unfortunately, unlike you guys in the US and parts of Europe, I rarely have a chance to see original prints by master printers.
Steve

Steve, this is a considerable disadvantage. On your next trip to Melbourne or Sydney, make sure you have some extra time to visit a museum. You might even arrange a visit with the Photography department and have them show you the special stuff that's rarely on the walls. They have plenty of resources down there.... and some great history.

I am always amazed with the older prints, whether they be gravure, carbon, albumen, etc., and how much historical photographers were able to do with those mediums. The sense of atmosphere, etc.

Books are often printed too contrasty, IMO, tho' this is not always true. I have a Stieglitz book that is quite beautiful. There is nothing as important in developing one's skills whether it be technologically or aesthetically, than seeing what others have done before you.... in my opinion. Experimentation is great, but experimentation plus understanding is better.

Get thee to a museum! ;-)

Lenny

[SteveH]

Steve, this is a considerable disadvantage. On your next trip to Melbourne or Sydney, make sure you have some extra time to visit a museum. You might even arrange a visit with the Photography department and have them show you the special stuff that's rarely on the walls. They have plenty of resources down there.... and some great history.

I am always amazed with the older prints, whether they be gravure, carbon, albumen, etc., and how much historical photographers were able to do with those mediums. The sense of atmosphere, etc.

Books are often printed too contrasty, IMO, tho' this is not always true. I have a Stieglitz book that is quite beautiful. There is nothing as important in developing one's skills whether it be technologically or aesthetically, than seeing what others have done before you.... in my opinion. Experimentation is great, but experimentation plus understanding is better.

Get thee to a museum! ;-)

Lenny

Thanks Lenny, I fully agree. Although over here Museums are reserved for Dinosaur bones, bug-collections and the patronising of indigenous cultures :confused: . I'd have to go to the Art Gallery instead. :)
I've carried this idea across to a post on the Style etc Thread.
I agree that exposure to the best the world has to offer is important for setting a standard to aspire to.
I do visit the State Gallery almost every time I'm in Melbourne but there is often little in the way of decent Photo art on display.
I'll have to take your advice though and tee-up a viewing of their undisplayed collection. That's an excelent idea; I'd always thought it was just for art history experts or scholars. Sort of shows the Aus attitude to art though I suppose.

And I agree with you that offset printing in general seems to aim for maximum contrast. My only real gripe with B&W magazine.
Thanks for the comments
Steve

[Brian Ellis]

Brian,

I took a look at an article about soft proofing from Computer Darkroom (http://www.computer-darkroom.com/softproof/softproof_1.htm). It says that edits should be applied to the original image so that it looks like the duplicate image with 'View Proof Colors' applied. I think we're saying the same thing, but in slightly different ways.

From the article:

"Make sure that you have the original image and NOT the duplicate selected. The duplicate will be used for reference; all edits will be applied to the original.

Choose the New Layer Set option form either the Layer menu or via the Layer Palette.

Name the Layer Set so that it relates to the media profile, e.g. Fuji Hunt SemiGloss "

One creates adjustment layers within this layer group in order to make original image look like the 'proof' image.

I do this with every image that I want to print, and it works nicely.

SteveH: Give the article I linked to a read. It deals with color management in PS and hopefully will be of help to you. (Note: The article is for CS2, but is easily transferable to later versions.)

--P

Preston - I think we are saying the same thing and the the difference is in which version (original or duped original) is being soft proofed, edited, and printed and which is being used as a reference.

In the article you cite if I understand it correctly the author dupes the original and applies soft proofing to the original, using the dupe as the reference (i.e. the author is editing the soft-proofed version to make it look as much like the soft-proofed dupe as possible). I soft proof the dupe, leaving the original as the reference, and edit the dupe to make it look as much like the original as possible.

Since the dupe and the original start out as identical images I don't think it makes any difference which one you soft proof and edit and which one you keep as a reference. But if I'm mistaken please let me know, I don't consider myself a color management expert by any stretch.