sRGB or Adobe RGB?

Mark - I would suggest one correction: RGB and CMYK are the two ways to build color that people work in. sRGB and aRGB are variants of RGB.

 

I'd never tell anyone to change something that works, but I'd also not recommend to others to set their working space as their monitor profile.

p.s. It is possible that some previous problems came from color gamut mismatches with the wider aRGB space, and your current set-up has a smaller gamut that is entirely or nearly-entirely contained within the gamuts of the output devices you work with, which would simply your workflow.

Marc, I'd appreciate some step by step painting by numbers if you don't mind.

 

I open my nef's, psd's and such and it gives you three choices, ("How do you want to precede?") 1,"Use the embedded profile" 2,"Convert documents colours to the working space" 3,"Discard the embedded profile"

 

Embedded, for me, is adobe rgb (1998) and working is srgb iec61966-2.1

 

Which of the three do you choose? Assuming you are printing from a frontier with a profile.

Then after I have made a specific duplicate to send to the printers and have adjusted and tweaked, I go image/mode/assign profile and click off the icc profile for one of the labs that I am using. Sometimes the image goes a little green or red, for which I then go up to curves and adjust it in one of the rgb channels until it looks good again. I then save it and send it off, and everytime the print looks identical to what I proofed on my monitor. Am I doing this "right"?

Eric - You asked Mark, but I'll be nervy and jump in.

 

I think you're doing it close to right. You can edit in aRGB or sRGB, but since you've captured in aRGB, I'd keep hold of those extra colors and edit there, too. [You could even set that as your working space if you want to avoid the profile mismatch dialog.]

 

The only thing I'd say to change is that your profile switch should be a Convert to Profile instead of an Assign Profile. Assigning profile doesn't change the numbers; it changes the way they're interpreted. What you'd like PS to do is preserve the appearance of your file, but put it into the gamut of your output device, and that's where Conversion comes in.

Cheers Marshall, i didn't mean to exclude you. I've asked Marc so many questions, i guess it was habbit...

 

the next question is, what is the difference between this method and soft proofing?

Well, my interpretation goes like this: soft-proofing is basically viewing at the image as if you'd Converted to that profile, but without executing the change. Any edits you make while soft-proofing are still applied in the original color space. Thus, you can prepare your file for output and keep those changes, all without putting the image through a potentially destructive mode change, and you even get to toggle back and forth to see what happened to your colors.

 

That said, if you've got a master file saved and any edits you make are solely for the purpose of output to that one device, you can make the Conversion and get the same results. Just don't overwrite the masterfile.

-->No I didn't make an error in saying that I calibrate my monitor and then select that as

my working space ... then also select that as my PS proofing profile. Right or wrong, it's a

decision based on consistent results.

 

If you understand the implications of using a display profile (which is changing everytime

you make a new one) as a working space by all means. IF all the documents are tagged,

the working space here plays no role. Any new document you make (and leave this way) is

based on a highly device dependent color space that's always changing, based on the

gamut of your one device and will not be gray balanced where R=G=B (something ALL

synthetic RGB working space do). It's also a really bad idea for users to work this way so

while you can certainly do this, don't advise other less knowledgeable users to do this. It

took me a year to get good ol' Will Crockett to stop suggesting this sillyness on his site (I

suspect many poor users are still doing this).

 

There's also no reason to load this profile in a soft proof (FWIW, the feature is hard wired

there anyway in the Proof Setup and that's why you have those options; Macintosh RGB,

Windows RGB and Monitor RGB, the later using the display profile). But that's a big huge

difference compared to using the profile as a working space. That's only being used for

previews. The working space is the foundation of the colorspace you're editing in.

 

The very reason we have RGB working space that are synthetic and Quasi-Device

Independent and have had them since version 5 is so users would NOT edit their

documents based on their display profiles. If I create a document in my display profile

space and you create one in yours, if the numbers in our files are the same, they will

appear differently. If the were to appear the same, the numbers would be different and all

conversions would be different. IF you have 89/12/139 in Adobe RGB (1998) and I have

the same numbers in Adobe RGB (1998) we see identical previews from the same numbers

and if we convert to the same output space, they resulting numbers are identical. That's

NOT what's going to happen if we both use a highly device dependent RGB space to edit

our files.

 

There's NO advantage to using an display profile as a working space and many reasons

why it's not good. But again, if you like it, understand the limitations of doing this, feel

free. For those out there who are not sure, don't do it, it's a really bad idea.

Okay Andrew, I'm up to learn anything that'll improve the stability and consistency of the

photos. I'm a photographer not a computer expert. However, I've been doing this now for

over a year and have re-calibrated my monitor quite a few times during that time span.

DVDs burned a year ago still print just as displayed when printed now. So, prints from

then look the same when printed today.

 

One question I do have is why there is a "work space" option at all?

 

Now I admit that I may have been doing something wrong back before this, so I'm open to

any suggestions if the images improve. Otherwise if it ain't broke, why fix it?

-->One question I do have is why there is a "work space" option at all?

 

A working space is just an editing space.

 

In Photoshop 4 and earlier, that space was based on your display. Photoshop simply sent

the numbers in the file directly to the display. So if your display was 3 units too yellow,

you were editing based on this bias. This was before we built and used ICC profiles for

handling our displays, our output devices and to describe the numbers in our documents.

 

Photoshop 5 began the process of divorcing our individual displays from the editing of our

files. To substitute for this, synthetic color spaces like sRGB, Adobe RGB (1998) and so

forth were created. These spaces are not based on any single real world device. Best of all,

someone on a Mac who might calibrate his display to a 1.8 gamma 5000K aim point could

load a digital image full of numbers (that's all computers really understand) while someone

on a PC with a 2.2 gamma at 6500K could view the SAME numbers the same way. That's

simply not possible when a million users are all viewing the same numbers based on the

Idiosyncrasies of their displays which are all over the map. The RGB working space is

totally independent of your display, That's where the display profile comes into play. It's

the great equalizer. So while the numbers in a one document in one color space is the

same, on a dozen different users machines, each display is different. But the ICC profile

that describes that display is used to ensure all the numbers (which are the same) preview

the same.

 

What Adobe did was quite brilliant.

 

Unfortunately there are some so called experts (and color management nay sayers) who

just don't get it and post articles on web sites that fly against everything the engineers at

Adobe have implemented.

Greg, catch all that?

 

cheers Andrew, very good.

Yes, I want to understand this in more depth. Well ... deep enough to keep the consistency

I now have while stabilizing viewing by any other system other than my own. And you are

right Andrew, when I couldn't achieve any kind of consistency with my prints, others

pointed me in this direction.

 

So, to make sure I get it right...

 

when opening an image and the color management dialog box asks what to do, I select

Adobe RGB 1998 (if it isn't already that). Then before correcting the open image I make

sure monitor RGB is selected in PS proof set up and be sure proof colors is checked

(selected). Right?

 

Well I did exactly that, and opened an old image that used my current monitor calibration

as the color space, switched it to Adobe RGB 1998 which threw the color off quite a bit ...

but when I selected Monitor RGB in PS it went right back to the original color. With proof

colors selected I then sent it to the 2200... and yep the print matched the screen.

 

Now I need to see if it will do the same on a completely different computer/printer.

in the second link provided Marc, way at the top, Ethan said this;

 

Ethan Hansen , sep 20, 2004; 12:51 p.m.

Assigning a profile and converting to a profile are two entirely different things. Assignment simply attaches a label to the image. You use this if the image either was untagged (no color space associated with it) or was incorrectly tagged with the wrong color space. The image data remain unchanged after assignment - only the label is different. Conversion alters the actual image data to move it from one color space to another.

 

It helps me to think in terms of languages. Languages give meaning to our squeaks, grunts, and scribbles on paper. Similarly, a color profile gives meaning to RGB or CMYK numbers in a digital image. Without the associated color space, the numbers are nothing more than numbers. They do not correlate with any physical color.

 

If you are handed a text in an unknown language, for the sake of argument let's assume it is a tablet from ancient Crete inscribed in Linear A, it makes no sense if you try to interpret as being English text. This is the equivalent of feeding a Fuji Frontier an Adobe RGB image. If you learn that the text is in Linear A (we've just "assigned" the Linear A profile) you can findo someone to translate (convert) the text into a language you understand. This is the same process followed when you convert an image from your working color space into the printer space. If all you do is assign the printer profile, it is the same as loudly proclaiming our ancient Cretan tablet is written in English. That's nice, but it still makes no sense.

 

Soft proofing to a printer profile is used to see how the image will appear in print. Photoshop does several conversions on-the-fly. The first is from the document color space to the printer profile. The second is from the printer profile to your monitor profile. This is what is shown on-screen. The advantage here is that you can work in a color space designed for editing while seeing how the final print will appear. If the printer's limited shadow range will clip your shadow details, you can apply selective dodging. Any colors that exceed the printer gamut can be flagged and edited if you dso not like the color profile's default conversion.

 

Soft proofing to your monitor profile only makes sense if you are preparing images to be viewed on your monitor outside of a color managed application. A PowerPoint slide show might qualify here.

and then he provided this <a href="http://www.drycreekphoto.com/Frontier/using_printer_profiles.htm">link</a>

So Eric, is what I did in the post above correct or incorrect? It also worked at the print

level just as well as the way I've been working for the last year.

 

My head is spinning just like it was a year ago.

 

I think I'll go make some more photographs ; -)

 

BTW, I picked up an EPSON Picture mate to make quick 4X6s, and the prints are

surprisingly good ... and can be handled without fingerprints despite being a semi-gloss

finish. Prints are 29 cents each and everything needed comes in a box to make 100

prints. Inks for a 100 prints are self contained in a single cartridge. It's really small.

To all of you, thank you. This thread has been both educational and delightful, maybe the best thread I've read in the past six months. I'll bookmark it...

"So Eric, is what I did in the post above correct or incorrect?"

 

I'm really not sure Marc. i have such a hard time getting my head around this stuff from the printed word. I need it spoken into my ear while someone has their hand on my mouse. But what i do is working perfect. except that i'm going to try the above recomendation for 'converting to profile' instead of 'assigning profile'. I'm mostly using this for actors headshots and model testing, which i need to be bang on, and it is suprisingly great. wisiwig, and no more teenage c-41 button pushers guessing. i just check to see if the lab has an updated icc, if so they email me, i drop it into the labrynth of windows folders and voila! i just want a better monitor now :P

"But even than you are still left with the same question what is better sRGB or Adobe for the amount of colors in the final print?"

 

The paper and printer determine the amount of colors ("color gamut") of the final print. Therefore use your lab's recommended profile for the files you submit to them. Why? Because most labs ignore profiles (Fuji Frontier equipment doesn't seem to notice them at all) and therefore your large-gamut RGB values will be scaled to fit into their small-gamut system. What this means is that your brightest possible red will become their brightest possible red, and your bright, but not fully saturated, red will be desaturated.

 

My recommendation: work on the image in Adobe RGB (which is indeed large, and includes the entire gamut of the common CMYK space--a big plus for prepress), then convert it to the lab's recommended profile for printing. You can use Photoshop's soft-proofing feature to see if your image overruns their equipment's color gamut and make adjustments accordingly.

Wow, Eric and Marc.

 

I was able to follow your comments. Thank you very much, I did go back and played with some settings here and as Eric said converting to a profile makes it a whole lot of difference as compared to Assigning a profile.

 

I played with my Epson 2200 and soft proofing and I absolutely loved the results I got.

 

I thought I will also share with you the paper that I have started to use and I am very excited by the quality and the price.

 

I started buying from www.lexjet.com the paper is semimatte

 

I hope some of you will also find it increadible.

 

Best,

Greg

<I>when opening an image and the color management dialog box asks what to do, I select Adobe RGB 1998 (if it isn't already that). Then before correcting the open image I make sure monitor RGB is selected in PS proof set up and be sure proof colors is checked (selected). Right?</I>

<P>

You don't have to convert to aRGB if the image is already in sRGB. It gives a little more breathing room in some colors, but isn't necessary. [The camera can capture in a larger space than sRGB, but once the image is in the smaller space, any additional color information is already gone. For some subjects (e.g., most portraits), it won't make any difference). For other subjects, you may be making adjustments that would push it into the larger gamut.] But yes, if you want to edit in aRGB, Convert (not Assign) on opening.

<P>

For the second point quoted, you'd want the profile for your output as the proof option, not the monitor profile. If you're sending to a 2200 on one of Epson's Premium papers, say, you'd select the profile for that paper/ink combo and review that way. If you have a Frontier profile, you'd set the proof setup to that profile.

<P>

Onward.

... and upward !