Clarification on Color Management (Working Spaces and Rendering Intents)

[Abitconfused]

Concerning the question of working color spaces... I use Adobe RGB because I do not believe my printer (Epson 3800) can print much beyond sRGB but I may be wrong here. I read that ProPhoto may not be an acceptable space because it is so big that colors may lose accuracy during the conversion to a smaller color space by printer drivers. Just because a person is able to send a printer ProPhoto images doesn't mean the printer can actually make use of that much information...as has been said.

[Colin Southern]

Concerning the question of working color spaces... I use Adobe RGB because I do not believe my printer (Epson 3800) can print much beyond sRGB but I may be wrong here.

I think you might mean "much beyond Adobe RGB"? - any modern CMYK printer should be able to exceed sRGB quite easily.

[Abitconfused]

You are quite correct! I did some research and was somewhat stunned to discover the capabilities of modern inkjet printers. I will have to revisit ProPhoto as an optimal working color space.

[Colin Southern]

You are quite correct! I did some research and was somewhat stunned to discover the capabilities of modern inkjet printers. I will have to revisit ProPhoto as an optimal working color space.

It's not so much "the capabilities of modern Inkjet printers" (although having just said that they're definitely getting better, not worse) as it is the fact that the technologies between monitors and printers is as different as chalk and cheese.

With monitors we start with a black screen and then add red, green, and blue - so there isn't a primary illuminant for colours like cyan, magenta, and yellow. On the other hand, with a printer, we start with a white page (reflecting all colours) and use inks cyan, magenta, and yellow (and black) inks - so printers struggle to produce certain hues of red and blues.

So in real-world terms, sRGB represents more or less the lowest common denominator between the two technologies, but when you start getting outside of that - especially with printers - you get into the realm of Adobe RGB for starters (which is still a pretty good choice because many monitors are getting close to it's gamut). Prophoto is of course bigger again; in my opinion it's advantages are more theoretical, but no problem using it if people know what they're doing -- the downside is that if people don't know what they're doing then they could well find themselves adjusting something to taste on the monitor not realising that the monitor by virtue of it's limited gamut is actually displaying a completely different colour to the one they've adjusted to -- and then of course the fun really starts when they print the image and then wonder what on earth is wrong with the printer ...

My suggestion is that if people are getting their work printed elsewhere - and aren't strong on colour management - then stick to sRGB ("Safe RGB"). If they are doing their own printing - have a monitor that can handle most of the Adobe RGB gamut or they're reasonably comfortable with colour management then use Adobe RGB; they might clip a few colours in some scenes but it's very unlikely that they'll ever be able to see the difference. At the other end, my suggestion is for people to avoid the likes of ProPhoto unless they're either uber familiar with colour management and softproofing (unless it's a package like LR which doesn't give you a choice).

[fenix]

What colors are the camera able to capture, in general? Does it matter if the camera gamut is larger than Adobe RGB?

A camera sensor can 'capture' colors both inside and outside of the human visual range, but most sensors have filters on them that restrict the range somewhat. Even so, there are some cameras that can capture infrared without modification if you expose long enough. Purists might object, saying that if you can't see it, it's not a color.

BTW, in full we should say "Adobe RGB (1998)" so as not confuse it with Adobe Wide Gamut RGB.

If they haven't already, someone will be along shortly to tell you that cameras don't have gamuts as such.

Eric

[Manfred M]

Concerning the question of working color spaces... I use Adobe RGB because I do not believe my printer (Epson 3800) can print much beyond sRGB but I may be wrong here.

As you have noted, yes it can (as can my Epson 3880)

A bit of a history lesson.

sRGB - was a HP / Microsoft collabaration back in 1996. It represented the technology of computer screens and printers of the day. Being the lowest common denominator, it remains the standard used by the internet, even today

AdobeRGB - put out by Adobe in 1998 and was a colour space that was developed for CMYK printers that could be produced by RGB colours. I believe this is where the view that inkjet printers are RGB devices comes from.

Neither sRGB nor Adobe RGB were specifically developed with photography in mind.

Hence; ProPhoto, which was developed by Kodak as a colour space with photographic output in mind. If I recall correctly, it covers 100% of "real world" colours and has a range that is imaginary and not visible. It has a wider range than the (Adobe) Wide Gamut RGB colour space.

So pick your poison; my default is ProPhoto and I will convert to the most relevent colour space as the last step of post.

[Rex]

I think you might mean "much beyond Adobe RGB"? - any modern CMYK printer should be able to exceed sRGB quite easily.

They don't, well certainly the 3800 doesn't. If you create an sRGB Granger Chart and do Gamut warning for the best Epson Paper and Epson Inks you will see a lot of the Granger Chart go grey. If you like textured art papers then you will see even more of the Granger Chart go grey in the Gamut Warning!

The two Gamuts are a different three dimensional shape so there are colours in the aRGB gamut the Epson will reproduce even though it doesn't completely encompass sRGB.

In reality you won't see the difference in one print taken through an aRGB workflow if it is on one side of the room and one taken through sRGB is on the other side of the room. You need to carefully pick your images and probably saturate colours in the aRGB in order to show a difference between aRGB and sRGB prints (on my 3800 and on some other printers too).

As photographers who produce a single image for inspection by a judge the absence of colours between aRGB and the gamut of your printer will not be noticed by a judge. What they will notice is incorrect skin tones and over saturation in colours especially in grass.

Very few images rely on having a large gamut for success, they are more about composition, shape, texture, tone, pattern, juxtaposition, humour, sadness, joy, beauty, etc

There is no beauty or art in a gamut.

[Rex]

and probably saturate colours in the aRGB in order to show

should be

and probably saturate colours in order to show (Cyans are a weak point for sRGB and better in aRGB)

[ajohnw]

Pity Colin has left as he is probably the only one that would answer that i full. A Granger chart is 16bit colour so some colours will be out of gamut what ever it is displayed on. If it was pure sRGB and confined to that then an aRGB display would show out of gamut colours as would the converse. Other factors come into it as well when viewed on a display such as the minimum brightness levels that can show details due to back lighting More generally the bit depth of what ever the output is in and that areas actual available dynamic range as well. When printing the gamut conversion used specifically for that should indicate yet more information on what is in and out of gamut just as a paper profile will.

I'd hope that previous posts have cleared up what a working space actually is and how it's manipulated to various output gamuts.

In terms of output intents in some specific output gamut the best place to look is probably the CinC tutorial on that specific area.

John
-

[xpatUSA]

[CMYK printers] don't, well certainly the 3800 doesn't.

Eric Chan of Adobe says otherwise, sorry.

It is sometimes thought that printer gamuts are smaller than monitor gamuts. This may be true, but we can see from the first row that this idea is misleading. My LCD (which is an average performer and roughly matches the sRGB color space) certainly contains some colors that my 3800 can't reproduce on Premium Luster (see the light gray areas that poke out of the colored volumes), but similarly PL can hold many colors that my LCD can't show (see the light yellow, mid green, and dark reds and magentas that are poking out of the gray blob). Which gamut is bigger? The answer doesn't matter. The important point is that neither gamut contains the other: each one has colors the other can't reproduce.

He knows his stuff, IMHO.

Most importantly, he concludes:

If you use sRGB or Adobe RGB as your working space (e.g., in Photoshop), then you won't be able to access all the colors the 3800 is capable of printing. These missing colors are usually the light yellows and the mid-to-dark greens and red/magentas. Use a wider space such as ProPhoto RGB to unlock these colors and make full use of your printer and paper's gamut!

See more here:

http://people.csail.mit.edu/ericchan...00/gamuts.html
.

[ajohnw]

I don't think even the Adobe comments are entirely clear Ted really. If some one has an epson printer and can't generate their own profile I assume they should look here

http://www.epson.com/cgi-bin/Store/j...seBVCookie=yes

and some where like this for paper

http://www.redrivercatalog.com/profi...rprofiles.html

The comments are sort of clear in one area - ProPhoto can contain the colours that the printer can produce. Only problem with that is that no printer can actually print the full ProPhoto gamut.

So information flow can be

Rawfile --> Workspace Profile --> Printer profile --> Paper profile -> Monitor
................................................--> Printer
............................................... --> sRGB profile --> Monitor / Web
............................................... --> aRGB profile --> What ever some one uses it for including 10bit aRGB on a monitor.
............................................... --> Monitor / Web ... some one on here does do this complete with a prophoto profile !

Using the word profile as this is what causes the transformation. An aRGB or an sRGB profile is only of any concern when one is used for some reason or the other. I for instance would use an sRGB profile for printing as that is what my printer is intended to cover. The catch with extended gamut printers is that some of the colours that they can cover can't be seen via any monitor profile. Methods of coping with that are mentioned on the web. Colin's method is to keep an eye on saturation of the ones he can't see. He did admit once if I remember correctly that things were still doubtful until the image was actually printed. Other clever clogs on the web pretend otherwise.

Why gamuts with more range for printing - because the dynamic range is rather limited. Often countered by an increase in saturation as maximum contrast is rather limited.

Some people insist on posting images with other than an sRGB profile. The effects are mixed. Sometimes hardly anything changes at all. At other times it does. The web doesn't help at all in this respect as the same images are shown with different profiles and a comment is added that all should look the same. As different gamuts cover different ranges and coarseness of colour variations this can't be the case all of the time. Here again a profile is used to convert from what ever it is to what ever it has to become with a caveat that colours may change.

Doh I missed out one of the profiles - the one that corrects the response of the monitor.

John
-

[Downrigger]

Eric Chan of Adobe says otherwise, sorry.
Most importantly, he concludes:

Thanks for this, Ted. Chan's "these missing colors are usually the light yellows and..." is part of the problem I'm struggling with (in addition to my own incompetence with exposure, WB, and PP) in my post today about Aspens, I think.

[xpatUSA]

I don't think even the Adobe comments are entirely clear Ted really.

I'm a bit confused, John. firstly, I should have made it clear that the comments are not Adobe's -they are from Eric Chan's personal site.

Having said that my response was only to the following two posts:

[QUOTE = Colin Southern] - any modern CMYK printer should be able to exceed sRGB quite easily.[/QUOTE]

[quote = Rex] - They don't, well certainly the 3800 doesn't.[/QUOTE]

Then I said "Eric disagrees" and posted a link to where he said that - just as a tie breaker for "Colin said this and Rex said that."

Eric is a proper Scientist and works at, or consults with, Adobe on ACR, LR, PS and more. Given the choice of believing his word and Rex's, I incline to the former.

May I ask what you find unclear about Eric's specific comments about the Epson 3800?

[ajohnw]

I'm a bit confused, John. firstly, I should have made it clear that the comments are not Adobe's -they are from Eric Chan's personal site.

Having said that my response to the the following chain of posts:

[QUOTE = Colin Southern] - any modern CMYK printer should be able to exceed sRGB quite easily.

[quote = Rex] - They don't, well certainly the 3800 doesn't.[/QUOTE]

Then I said "Eric disagrees" and posted a link to where he said that - just as a tie breaker for "Colin said this and Rex said that."

Eric is a proper Scientist and works at, or consults with, Adobe on ACR, LR, PS and more. Given the choice of believing his word and Rex's, I incline to the former.

May I ask what you find unclear about Eric's specific comments about the Epson 3800?[/QUOTE]

Just crossed wires Ted - as he seems to be an Adobe person I assumed it was the company talking. I do feel though that partial statements in this area do confuse people. Another one that crops up with wider gamuts is more saturated colours - probably true but only if they are used.

John
-

[Rex]

My word there is quite a bit of misunderstanding here.
I've read the quote that has been repeated from Eric

certainly contains some colors that my 3800 can't reproduce on Premium Luster (see the light gray areas that poke out of the colored volumes), but similarly PL can hold many colors that my LCD can't show

and as far as I can see he and I agree within that quote.
The 3800 has a gamut which is different from the sRGB gamut and in some place it is bigger and in some it is smaller, you need to see it in 3D. I'm pretty sure that's what I said.
Including:

The two Gamuts are a different three dimensional shape so there are colours in the aRGB gamut the Epson will reproduce even though it doesn't completely encompass sRGB.

A Granger chart is 16bit colour so some colours will be out of gamut what ever it is displayed on.

This is a misunderstanding, you create the Granger chart you want (sRGB, aRGB ProPhoto or any other gamut) , and the the fact that it is or isn't 8/16 bit is immaterial, that is to do with the number of individual colours it contains not the gamut.
The Granger chart is a two dimensional representation of a 3D space and using Gamut warning within PS it is a tool you can use to explore comments about gamut.

You can believe Eric or you can believe me (we seem to agree) or much much better you can do the experiments at home with your own software , your own printer/paper icc profile and find out for yourself. Then you don't need to quote anyone, you can tell us what your own experience is, which in my book is much more useful.

The fact that things are often repeated on the net doesn't make them true.

[Rex]

The comments are sort of clear in one area - ProPhoto can contain the colours that the printer can produce. Only problem with that is that no printer can actually print the full ProPhoto gamut.

ProPhoto is an all encompassing gamut and as such it can mathematically represent colours that do not exist. It has to be as big as it is as in 3D space it is fairly simple and therefore it needs to be large such that it is all encompassing the complex shapes of 'real' gamuts and real colours.

Elsewhere on this forum I read about Pointer's Gamut, I went to a lecture which included this Gamut and since then I've been trying to find out if it is available as an icc file.

Pointer's gamut is contained within ProPhoto. :-)

So your comment is perfectly correct. :-)

[xpatUSA]

My word there is quite a bit of misunderstanding here.
I've read the quote that has been repeated from Eric

certainly contains some colors that my 3800 can't reproduce on Premium Luster (see the light gray areas that poke out of the colored volumes), but similarly PL can hold many colors that my LCD can't show

and as far as I can see he and I agree within that quote.
The 3800 has a gamut which is different from the sRGB gamut and in some place it is bigger and in some it is smaller, you need to see it in 3D. I'm pretty sure that's what I said.
Including:

The two Gamuts are a different three dimensional shape so there are colours in the aRGB gamut the Epson will reproduce even though it doesn't completely encompass sRGB.

A Granger chart is 16bit colour so some colours will be out of gamut what ever it is displayed on.

This is a misunderstanding, you create the Granger chart you want (sRGB, aRGB ProPhoto or any other gamut) , and the the fact that it is or isn't 8/16 bit is immaterial, that is to do with the number of individual colours it contains not the gamut.
The Granger chart is a two dimensional representation of a 3D space and using Gamut warning within PS it is a tool you can use to explore comments about gamut.

You can believe Eric or you can believe me (we seem to agree) or much much better you can do the experiments at home with your own software , your own printer/paper icc profile and find out for yourself. Then you don't need to quote anyone, you can tell us what your own experience is, which in my book is much more useful.

The fact that things are often repeated on the net doesn't make them true.

Getting a bit provocative there, Rex . .

I took your statement to mean that any modern CMYK printer is not able to exceed sRGB - in direct contradiction to Colin Southern. If that was not your meaning, as apparently it is not, then Eric Chan is right according to the both of us. Case closed?

Forgive me for thinking that the statement "A Granger chart is 16bit colour so some colours will be out of gamut what ever it is displayed on." meant that some colors will be out of gamut because it is 16bit. Your later explanation does clarify your meaning somewhat.

I don't use PhotoShop, never have, so didn't know what a Granger Chart is. I study gamuts with ColorThink which does rotatable 3D views in the Lab, Luv or xyY color models.

Since I don't print, don't have printer profiles, and have little real interest in this thread, I will not be following your suggestions as to testing. Might as well also say that, if I feel it appropriate to quote credible sources such as Eric Chan, I will continue do so in spite of your barbed comment.

Happy New Year . . .

[Rex]

Eric is a proper Scientist and works at, or consults with, Adobe on ACR, LR, PS and more. Given the choice of believing his word and Rex's, I incline to the former.

I read the quote you made and found that I did not have a difference of opinion with the quote you made.

then Eric Chan is right according to the both of us. Case closed?

Yes.

If I came over barbed, it wasn't intended, probably the same as your quote didn't intend to dismiss me as less of a scientist.

The ColorThink software is good for viewing Gamuts in 3D and making comparisons. The reason I wanted Pointer's Gamut was to load into that software. The thing that the software lacks is the human. What we are looking for is a "Convincing Lie" in our colours. (Courtesy of Dr Tony Kaye FRPS)

[Rex]

[Richard Lundberg]

It's Poynter.