Colour Calibration Nightmare. Nurse required.

[jaapv]

Advertisement (gone after registration)

[ATTACH]269759[/ATTACH]

 

[ATTACH]269758[/ATTACH]

 

Umm.. You chose 16-bit output in LR, yet did not tick the 16-bit box on the printer. Are you sure that is correct?

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[jaapv]

Jaap - please don't be sorry - it's me whose making the mistakes here, and you're trying to help, so I should be saying sorry.

 

But i think you're assuming I have greater understanding of the issues than I actually do. I don't understand, for example, why you're drawing a connection between the Adobe RGB profile, which seems to work for me better than my own profiles (another thing I fail to understand!), and sRGB, which I don't use at all for these purposes.

 

It suggests to me that there's a connection between sRGB and AdobeRGB that I'm missing. I thought they were simply different colour spaces.

 

You misunderstand-I meant as Adobe RGB was OK, you had your problems in another colourspace eg. sRGB. Which not, I understand now.

[jaapv]

Jaap - please don't be sorry - it's me whose making the mistakes here, and you're trying to help, so I should be saying sorry.

 

But i think you're assuming I have greater understanding of the issues than I actually do. I don't understand, for example, why you're drawing a connection between the Adobe RGB profile, which seems to work for me better than my own profiles (another thing I fail to understand!), and sRGB, which I don't use at all for these purposes.

 

It suggests to me that there's a connection between sRGB and AdobeRGB that I'm missing. I thought they were simply different colour spaces.

 

You misunderstand-I meant as Adobe RGB was OK, you had your problems in another colourspace eg. sRGB. Which not, I understand now.

[Guest WPalank]

Umm.. You chose 16-bit output in LR, yet did not tick the 16-bit box on the printer. Are you sure that is correct?

 

Not to mention, "Highspeed" is ticked in PS as well. As another poster stated, I would not suggest it as well.

[Guest WPalank]

Umm.. You chose 16-bit output in LR, yet did not tick the 16-bit box on the printer. Are you sure that is correct?

 

Not to mention, "Highspeed" is ticked in PS as well. As another poster stated, I would not suggest it as well.

[archi4]

Jaap

You are correct! In my hurry to show an example, I didn't check the box in both.

maurice

 

William,

After many comparisons, at sizes up to A2, I have been unable to detect any difference in the prints using Highspeed or not, either with my 3800 or now my 3880.

This concurs with the original test of the 3800 in the Luminous Landscape site if I remember correctly.

[archi4]

Advertisement (gone after registration)

Jaap

You are correct! In my hurry to show an example, I didn't check the box in both.

maurice

 

William,

After many comparisons, at sizes up to A2, I have been unable to detect any difference in the prints using Highspeed or not, either with my 3800 or now my 3880.

This concurs with the original test of the 3800 in the Luminous Landscape site if I remember correctly.

[t024484]

Maybe I can contribute in some way to the discussion, because there are still some important details that I think are missing so far .

Printing is all about matching three steps.

 

Step 1 is the color space of the image to be printed, let's say Adobe RGB is (1) and sRGB is (2).

 

Then comes the printing program like PS, LR or whatever. Choises are between: let the printer do the color management (A) or select an ICC profile (

 

Then select between possible settings for your (Epson) printer driver: No Colormanagement (I), AdobeRGB (II) or sRGB (III)

 

What I have read is that if your image is in Adobe RGB and your printer is set for Adobe RGB everything seems fine, this would be the combination (1) + (A) + (II).

 

When making your own ICC profile, it is important to know the color space of the testimage to be printed. This could be either Adobe RGB or sRGB and can be converted between both spaces with the help of Photoshop.

In the first case use (1) + (A) + (II), in the second case (2) + (A) + (III).

If you do this wrong, the result is useless.

 

Once you have created your Adobe RGB or sRGB ICC profile, printing goes like:

(1) + ( being your Adobe RGB ICC profile + (II), or

(2) + ( being your sRGB ICC profile +(III)

 

Hans

[t024484]

Maybe I can contribute in some way to the discussion, because there are still some important details that I think are missing so far .

Printing is all about matching three steps.

 

Step 1 is the color space of the image to be printed, let's say Adobe RGB is (1) and sRGB is (2).

 

Then comes the printing program like PS, LR or whatever. Choises are between: let the printer do the color management (A) or select an ICC profile (

 

Then select between possible settings for your (Epson) printer driver: No Colormanagement (I), AdobeRGB (II) or sRGB (III)

 

What I have read is that if your image is in Adobe RGB and your printer is set for Adobe RGB everything seems fine, this would be the combination (1) + (A) + (II).

 

When making your own ICC profile, it is important to know the color space of the testimage to be printed. This could be either Adobe RGB or sRGB and can be converted between both spaces with the help of Photoshop.

In the first case use (1) + (A) + (II), in the second case (2) + (A) + (III).

If you do this wrong, the result is useless.

 

Once you have created your Adobe RGB or sRGB ICC profile, printing goes like:

(1) + ( being your Adobe RGB ICC profile + (II), or

(2) + ( being your sRGB ICC profile +(III)

 

Hans

[Guest WPalank]

William,

After many comparisons, at sizes up to A2, I have been unable to detect any difference in the prints using Highspeed or not, either with my 3800 or now my 3880.

This concurs with the original test of the 3800 in the Luminous Landscape site if I remember correctly.

 

Agreed maurice,

But when someone is trying to find the source of a problem, best to eliminate some of the variables. I think we can both agree that turning Highspeed on will not result in a superior print.

The printers are so darn fast these days, at least my 7900 is, I don't see the value. Also, when using the Image Print RIP (I know the OP is not) it is recommended not to use Highspeed.

[Guest WPalank]

William,

After many comparisons, at sizes up to A2, I have been unable to detect any difference in the prints using Highspeed or not, either with my 3800 or now my 3880.

This concurs with the original test of the 3800 in the Luminous Landscape site if I remember correctly.

 

Agreed maurice,

But when someone is trying to find the source of a problem, best to eliminate some of the variables. I think we can both agree that turning Highspeed on will not result in a superior print.

The printers are so darn fast these days, at least my 7900 is, I don't see the value. Also, when using the Image Print RIP (I know the OP is not) it is recommended not to use Highspeed.

[george +]

I am very late in this discussion and use Win7 and a Canon 9000 printer. But there is something parallel here with the problems I had.

 

Printing from Photoshop did not work for me no matter what settings I used.

I wrote it off as something special, a problem just for me.

 

I print these days direct from file views and my ColorMunki profiles work well.

 

My apologies for interjecting these somewhat irrelevant comments but our problems may be similar.

[george +]

I am very late in this discussion and use Win7 and a Canon 9000 printer. But there is something parallel here with the problems I had.

 

Printing from Photoshop did not work for me no matter what settings I used.

I wrote it off as something special, a problem just for me.

 

I print these days direct from file views and my ColorMunki profiles work well.

 

My apologies for interjecting these somewhat irrelevant comments but our problems may be similar.

[ho_co]

FWIW--about six hours ago, the OP started another thread in which he thanks you for your help and says his problem is solved thanks to your input: http://www.l-camera-forum.com/leica-forum/digital-post-processing-forum/193074-colour-calibration-update.html.

[Peter H]

FWIW--about six hours ago, the OP started another thread in which he thanks you for your help and says his problem is solved thanks to your input: http://www.l-camera-forum.com/leica-forum/digital-post-processing-forum/193074-colour-calibration-update.html.

 

Howard, thanks for putting up this link - I truly am very grateful indeed for all the help and good suggestions I've recieved.

 

I do have a number of questions about some of the terminology and apparent repetitions/duplications/contradictions that the various pieces of software contain, but I'll save these for another time - at the moment I have a lot of experimenting to do.

 

Thank you all again.

[adan]

I need to jump in here - off-topic slightly from Peter-H's problem - to clear up some misconceptions.

 

Don't mix up "color space," "profile" and "gamut." They are all different things.

 

Color spaces define the color range of an image file or document - they do NOT apply to printers, monitors, scanners, or other devices directly. There is no such thing as an "sRGB" or "Adobe 1998" printer (although marketers may say otherwise). Color spaces are very clearly defined by parameters such as white point, black point, and coordinates for the three primary colors RGB, as a subset of the master color space called CIE-L*a*b.

 

Here's the master CIE 1931 LAB color space (all the colors the eye can perceive), with the smaller Adobe 1998 color space mapped into it as a triangle: Adobe RGB color space - Wikipedia, the free encyclopedia

 

Devices such as printers or monitors have "gamuts" - the range of colors they can reproduce. Generally gamuts are smaller than most color spaces, and the gamut of illuminated monitors are different from the gamut of inks on opaque paper.

 

Monitor gamuts are usually pretty close to the common color spaces (or more accurately, the common color spaces (1998 and sRGB) were specifically designed to approximate the gamut of an "average" monitor). "Approximate" being the key word - they are not perfect matches, which why we still have to do our own monitor calibrations and create our own monitor profiles.

 

Printers have much smaller gamuts than monitors or color spaces. That particular lime green that is 0R/255G/0B on-screen simply cannot be reproduced by any combination of inks, which can at best print a dull forest or "Kelly" green.

 

A "profile" is essentially a translation, from the "color space" of the electronic image to the "gamut" of the printer or monitor. It translates the colors of your image into amounts of ink the printer can actually print, and color brightnesses your monitor can actually display.

 

The process of creating a profile is basically sending "known" colors to your printer, or to your screen; measuring how "wrong" they come out, and then calculating the corections needed to make the wrong look right.

 

To summarize: images are in defined color spaces (Adobe 1998, sRGB, ProPhoto, Bruce RGB, etc.), devices (screens, scanners, printers) have color gamuts (almost never the same as any color space), profiles translate between the two.

[delander †]

Off topic but a question for Andy. A CMS (color mangement system) uses profiles to map colours between devices. In Photoshop on a mac you have a choice of Adobe ACE or colorsync, does it matter which you choose? I have always left it at the default which is ACE.

 

Jeff

[Peter H]

Thanks yet again Andy, and if you're not yet exhausted by all this here's another question which I've been avoiding so far:

 

In PS CS4, under the edit tab, there's a "Color Settings" control, which gives options such as North American Prepress 2. This selection determines a number of factors such as whether to use embedded profiles or not, and the use of ACE or Apple CMM, amongst other things.

 

I have mine set for North American Prepress 2, which is what Epson suggest in their Guide to Color Management.

 

Then, immediately beneath that in Edit, there' the option to assign a profile, allowing "Don't Manage This Document", Working RGB: Adobe RGB (1998), or a full choice of profiles.

 

I've never used these selections, so they're set to the default "working RGB: Adobe RGB" setting, but I can't work out whether this is correct, or whether the prepress selection is correct, and I'm reluctant to run yet another set of test prints at the moment, changing all the variables one at a time .

 

There seems to be a multiplicity of ways of getting it wrong. This is either more complicated than I ever imagined, or its simple and I'm making it complicated. Either way, it ain't easy!

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[archi4]

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

This is my usual setup.

I use Prophoto RGB profile because that is close to what LR uses - I believe they actually use Melissa RGB.

For Black and white I sometimes use 2.2 Gamma color space and I sometimes use relative colorimetric instead of Perceptual.

By hovering your cursor over choices a short explanation is given in the space at the bottom.

 

I always check the boxes under color management policies

[adan]

For Jeff - I've always just used ACE as well.

 

For Peter H - don't worry about the CMYK color space. It has no effect on desktop printing. It is strictly for making half-tone color separation plates for 4-color printing presses (newspapers and magazines and books).

 

While desktop printers (and their bigger free-standing inkjet cousins) do use CMYK inks (actually, most of them use CcYMmKkk inks, sometimes with orange, green, red or blue thrown in (Canon, e.g.)) - they are not converted to CMYK by Photoshop, but by the printer driver on the fly.

 

You send the printer driver RGB color data, and it calculates how to use CMYK inks to reproduce those colors (again that is why printer gamuts are so different from monitor gamuts - monitors produce pure green or red light; the printer must make those colors with yellow + cyan ink, or yellow + magenta ink).

_________________________

 

However - if you want some fun that will REALLY mess with your head - try choosing the "Custom CMYK" item from that "Working Spaces: CMYK" menu. You will get a wild little dialogue box - which is what the pro color separators use - that allows you to control HOW Photoshop will create color separations. It gets into things like UCR (UnderColor Removal) and GCR (Gray Component Removal) - which are ways of NOT using colored inks except in the most brightly colored areas.

 

I.E. if you are printing a low-saturation or dark color (and most colors are not fully saturated), there is no point to piling up cyan, magenta and yellow inks on top of one another if you can just print more gray. To the extent that C, Y, and M add up to "gray" - you might just as well use black ink at 1/3rd the price.

 

Take a look at these separations, and you'll see that the dark purple fruits left and right actually have less cyan and magenta ink than the lighter, more saturated areas (the images look reversed, like a negative) - the "gray component" of C+Y+M has been "removed" and replaced with additional ink in the black plate: http://www.sirspeedystpete.com/images/4ColorSeparation.jpg

 

You can even see it with your own image by converting a picture to CMYK and then looking at the separate channels - with a high GCR or UCR setting, the CMY channels will look really weird, pale and solarized.

 

Desktop printer drivers have built-in GCR (as you may have noticed if you've ever run out of black ink in the middle of a print) - but you usually never see it since it is done internally during that brief pause while the driver thinks about things before the printer starts to print.

_______________

 

archi4 mentions "Relative Colorimetric" vs. "Perceptual" intent (which is a choice that shows up in several places: printing setups, and conversion of color files from one color space to another).

 

Here's the difference (and usually it is a small one):

 

Imagine you have a picture of a fire engine, with a whole range of reds in it. And you are going to change it to a smaller color space (say, sRGB for the Web) or print it with a smaller gamut - thereby compressing the most saturated reds until they fit in the new gamut or color space.

 

Consider 3 of the reds: A is out of gamut for the new color space/gamut, B is just barely within the new color space/gamut, C is a dull red well within both color spaces/gamuts.

 

Relative Colorimetric Rendering will compress only the out-of-gamut colors (red A), and not change B or C. The upside is that most of the colors stay as saturated as possible - the downside is that A ends up looking just like B, and you lose color/tonal gradations in the brightest colors.

 

Perceptual Rendering will compress/desaturate ALL the colors proportionally: A is compressed to fit, B is shifted towards C, C is shifted towards gray. The upside is that red A still looks different from red B (a difference can still be "perceived") - the downside is that the whole image has duller colors.

 

I've printed pictures using both rendering intents, and usually the difference, if any, is in the blues, cyans, and greens. Adobe 1998 has "more" blues and greens in it than printers or sRGB - so those are the colors that change most "differently" when compressed down to a printer gamut or the sRGB color space using different rendering intents. Usually it is more apparent as a hue shift than saturation per se. RC prints skies more cyan, and Perceptual prints skies more purplish.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!