Printing - sRGB and Adobe questions

[Mike Buckley]

Christina,

Is it possible that you're referring to the choices Spyder4Express software provides for displaying the graphs of the various color spaces after your monitor is calibrated and the profile has been automatically saved? That is an informational tool such as the graphs that have been posted in this thread and does not affect the performance of your monitor or the Spyder software. If I'm wrong about that, I would love to be corrected. EDIT: I just now reviewed the software (because, ironically, it just now reminded me that it was time to recalibrate) and I'm as positive as I can be that I'm right about that.

The reason I mention this is that you probably remember that I also use the same software and that's the only place that the subject of color space comes up on my system.

[xpatUSA]

I use LR 4 and Adobe Elements 9, but perhaps in a few years I might upgrade to Photoshop in which case it would be good to have my photos set to Adobe?

Hi Christina, back on your case . .

By "set to" I assume you mean "saved with an ICC profile embedded in the file" (that checkbox in the 'save as' pop up tells you how you're saving it). However, your question reads like in your mind you're associating Adobe RGB with the application (PhotoShop) . . as if Adobe RGB is the 'correct' color space for PhotoShop, in other words. I want to be sure you understand that a good post-processing application will open a file in any color space you want it to and will save it in any other color space you need, depending of course where the output file will be used or displayed. For example, RawTherapee is good in that regard. FastStone Viewer, on the other hand, happily strips the profile out of any file you edit - leaving any application that opens your Viewer-crafted file free to do whatever it likes. Had you saved a ProPhoto file out of FastStone Viewer, I assure you it could look pretty bland in, e.g., Internet Explorer 6.

Before anyone tells me how bad those apps are, I'm just using them as examples

[ajohnw]

It can be hard to find a web site that covers sRGB and Adobe RGB as most concern themselves with coverage and neglect practicalities. There was one that did this well and showed that on a well corrected monitor, that's the rub, an Adobe RGB image will look very similar to an sRGB one. When pictures are shown there is no way of knowing how they have been processed or what reality looked like so try this site. 100% sat bars are hard to argue with. Assuming that they are viewed in sRGB.

http://www.kenrockwell.com/tech/adobe-rgb.htm

That shows some of the claims made at times aren't correct and effects will be worse if the monitor is out and many are

The printers comment about looking flat are down to gamma. A fudge HP and Microsoft came up with to make images look better needed really because of the limitations of displays and the colour space available on them. While an Adobe RGB monitor looks an attractive option many obtain the extra gamut by dithering colours. Many people find that unacceptable. The mid grey setting image on most sites uses that to simulate mid grey and it has to viewed through squinted eyes to be effective as the dithering can be seen easily. Tftcentral reviews usually look at the actual display panel that is used.

What a calibrator really needs to produce are these as there are well established standards for what is acceptable, good or bad - apart from gamma. Some deliberately depart from the standard. Where some states that they like Adobe when printed I'm inclined to feel that the eyecandy should be added during processing in sRGB so that all stand a fair chance of seeing it and if it's printed in sRGB it will look the same anyway.

Doh. I can't post pictures. That aspect of the forum wont work for me at the moment, However I posted what I would have posted here some time ago. As mentioned there full calibration takes 50min and it seems to be important to run the calibrator for 10 to 15min first to ensure stable results. The same was true of another calibrator i used.

Any thoughts on LCD screen calibration and gamut coverage against colour temperature

The 1st monitor I calibrated was a poor one. As soon as I moved to the one above I realised that I was over sharpening with bells on. The calibration shown is rather good and since that I have wondered just how others see anything I might post. I have also used as supplied with colour calibrators software but then look at it with Argll and see things aren't correct at all as they just don't look thoroughly enough.

John
-

[Brownbear]

Hi Ted,

Thank you for the graph, it's another good visual... And thank you for staying on my case.

Yes, that is exactly what I thought and I now see that this is not the case at all. My understanding of the complications of printing, colour spaces, different file types and what my camera, computer and printer (and commercial printers) and even my spyder is capable of has improved immensely due to this thread. Thank you for taking the time to clarify.

All far more complicated than I thought.

Hi Christina, back on your case . .

By "set to" I assume you mean "saved with an ICC profile embedded in the file" (that checkbox in the 'save as' pop up tells you how you're saving it). However, your question reads like in your mind you're associating Adobe RGB with the application (PhotoShop) . . as if Adobe RGB is the 'correct' color space for PhotoShop, in other words. I want to be sure you understand that a good post-processing application will open a file in any color space you want it to and will save it in any other color space you need, depending of course where the output file will be used or displayed. For example, RawTherapee is good in that regard. FastStone Viewer, on the other hand, happily strips the profile out of any file you edit - leaving any application that opens your Viewer-crafted file free to do whatever it likes. Had you saved a ProPhoto file out of FastStone Viewer, I assure you it could look pretty bland in, e.g., Internet Explorer 6.

Before anyone tells me how bad those apps are, I'm just using them as examples

[Brownbear]

Thank you to everyone for your replies and the helpful links and information.

Tim, that is a wonderful easy to understand link that you included. I have book marked it for ease of reference.

Mike, I suppose it is a good thing that you have the same spyder as me. Yes, for some reason I though that when I did the calibration I was able to choose the sRGB or Adobe, and this caused some confusion. I calibrated my monitor again, and the display choice is at the end. It sounds like my monitor is not capable of displaying Adobe, and the only chance I get to see the true adobe colours is when I print it myself and only if one stays true to the colours of the scene as photograph in post processing. Very interesting.

John, Thank you. You are a wealth of information. The link by Ken Rockwell is helpful, as surely the link to the thread you provided will be, after I read it a couple of more times.


Thank you to all. I didn't realize that printing was so complicated and my understanding of the subject has improved.

[Mike Buckley]

Christina,

It would be a good idea for you to get in the habit of referring to "Adobe RGB" rather than just "Adobe." The first term is widely known as a color space and the second term is widely known only as the name of the company. Using the latter term in the context of a discussion of color spaces can lead to confusion.

[Brownbear]

Consider it done. Thank you for helping me along with my technical vocabulary.

Christina,

It would be a good idea for you to get in the habit of referring to "Adobe RGB" rather than just "Adobe." The first term is widely known as a color space and the second term is widely known only as the name of the company. Using the latter term in the context of a discussion of color spaces can lead to confusion.

[casper]

hello Christina,
Certainly a thought provoking question, which has generated a very informative discussion. To add to the references already suggested, there is a not-too-technical presentation that I found very helpful at www.Lynda.com: Colour Management Fundamentals (Joe Brady), Chapter 1, The Basics of Digital Colour, Understanding Colour Spaces (sorry, I haven't worked out how to insert a link here, yet). This episode is free to view for non-members, and uses an app. that enables all of the colour spaces to be represented in 3-D. Regards, Noel.

[Brownbear]

Thank you Noel. You did in fact, insert a link

hello Christina,
Certainly a thought provoking question, which has generated a very informative discussion. To add to the references already suggested, there is a not-too-technical presentation that I found very helpful at www.Lynda.com: Colour Management Fundamentals (Joe Brady), Chapter 1, The Basics of Digital Colour, Understanding Colour Spaces (sorry, I haven't worked out how to insert a link here, yet). This episode is free to view for non-members, and uses an app. that enables all of the colour spaces to be represented in 3-D. Regards, Noel.

[Richard Lundberg]

Christina
When Microsoft believed it would become the next world government, they averaged all the tvs and monitors they could find and came up the sRGB color space as being something everyone could see . That was a long time ago. Monitors and many tvs are much better now. sRGB is still a safe choice for the web, and many commercial printers have been forced to acknowledge its existence, but their pressmen only know their press and their ink and will proof what you send them, for you, for approval. Monitor calibration will give you all the gamut the monitor has, but you can restrict it to sRGB and use that. Since the web has become so pervasive, many "experts" will tell you to calibrate to 2.2 gamma and 6500 color temperature because this is the basis for sRGB and people will generally be viewing web images with monitors that are somewhere near sRGB. Printing to ink jets and dye subs is not the same. If you want to match your screen and your inkjet, your monitor should be calibrated at 1.8 gamma and 9300K. You also have to make sure your printing app, Photoshop for example, recognizes your printer AND YOUR PAPER. The color space of choice for RAW conversion, ProPhoto, is 1.8 gamma for good reason. In Photoshop you go to Print Settings, choose Photoshop manages colors, and then you see a list of space choices which includes your printer and a whole bunch of others, but the printer list includes it with several paper types. Choose the correct one. You will also see , over to the left a clickable box called "Print Settings", What? I'm already in Print Settings!. Yes, but there's one more step (So true, always). For Epson, at least, you have to click there and select the paper type AGAIN, or you will hate the results.
People whose opinion I respect think that Ken Rockwell should be ignored for the most part. He lives in an alternate universe.
Only Photoshop "Save For Web" converts to sRGB automatically whether you want it to or not. I have never printed from Elements or Lightroom but they should follow the method above in some manner.
If you start with converting to ProPhoto for RAW images out of your camera (others have told you that the camera doesn't know from color space, it just records pixel amplitudes and the converter demosiacs them into an image) and choose perceptual, then do your corrections, then convert to aRGB (perceptual)and do more editing, then print properly, you will get all there is to get. When you use JPEG as a camera output it does its own conversion to a color space and I think many cameras allow you to choose one.
Another note about gamma. Gamma refers to the exponent of the transfer function of a pixel's amplitude in either current or voltage to luminance of the monitor screen. Gamma as used in calibration actually means the gamma correction, the amount of gamma that is eliminated. 1.8 "gamma" results in prints better matched to paper reproduction. Most monitors have a native real gamma that varies quite a bit from the 2.2 assumed for sRGB and each color has its own. Modern LCD monitors are tailored internally to 2.2 using software or firmware. You can still cal them to 1.8 correction. Correction is done so that the luminance is linearized to provide the eye with the signal it expects, since the eye has its own gamma that varies with light level.

[Colin Southern]

When Microsoft believed it would become the next world government, they averaged all the tvs and monitors they could find and came up the sRGB color space as being something everyone could see.

Um - close. It was actually a proposed standard put forward by Microsoft and Hewlett Packard.

sRGB is still a safe choice for the web

Unfortunately, it's pretty much the only real-world choice at the moment, for quite a few reasons;

- if someone publishes an image on the net in Adobe RGB then it'll be seen as dull and under-saturated by anyone not running a colour-managed browser (still many people)

- Most people who see the under-saturated and dull looking image won't realise that it's because they're not running a colour-managed browser - so they'll just think it's a sub-par photo

- Even if someone has a colour-managed browser then unless they have a profile created for their monitor then the colour management system will in essence just convert the image back to the standard profile which is sRGB anyway.

- Even if they create a profile for their monitor - but the monitor only supports an sRGB gamut, then again, all they get to see is an sRGB image (because anything more would exceed the physical capabilities of the device)

So for it to be of any benefit, the majority of the target audience needs to have a monitor capable of something close to the Adobe RGB gamut - use a colour-managed browser - and have AT LEAST a default manufacturer supplied profile (preferably a custom one) - (and have the monitor calibration set approximately right) - which all in all I suspect is a VERY small percentage of the population, unfortunately.

So for now anyway, anyone who posts in Adobe RGB is probably going to be on the back foot

If you want to match your screen and your inkjet, your monitor should be calibrated at 1.8 gamma and 9300K.

The industry standard is D50 (5000 Kelvin lighting)

If you start with converting to ProPhoto for RAW images out of your camera (others have told you that the camera doesn't know from color space, it just records pixel amplitudes and the converter demosiacs them into an image) and choose perceptual, then do your corrections, then convert to aRGB (perceptual)and do more editing, then print properly, you will get all there is to get.

Maybe yes, maybe no. Personally I use Relative Colorimetric 99% of the time. The problem with Perceptual is that it in essence compresses the source gamut into the target gamut so on one hand you get "all she's got", but on the other hand, this is done at the expense of colour accuracy. RC rendering on the other hand preserves colour accuracy for in-gamut colours, and only clips OOG colours to the nearest reproducible tone (which usually isn't an issue).

[Andrew76]

RC rendering on the other hand preserves colour accuracy for in-gamut colours, and only clips OOG colours to the nearest reproducible tone (which usually isn't an issue).

Is that a program-specific algorithm? Like will LR calculate the same out of gamut colours, as any other programs?

[ajohnw]

There is a good general run down on this subject here

http://www.marcelpatek.com/monitor.html

I've found that if the monitor is calibrated and has an icc file that browsers are less important if the icc file is installed system wide. Also for instance if a colour managed application is using it's own general sRGB calibration profile this arrangement still works out. I had problems with a viewer and one editing application in particular before setting things up like this. In fact it made me realise that I had to calibrate the display. The editing application used it's own in built colour management. Viewer and browsers were we matched.

John
-

[Colin Southern]

Is that a program-specific algorithm? Like will LR calculate the same out of gamut colours, as any other programs?

It depends on the rendering engine. Adobe products possibly all use the same one (although I haven't checked), but I doubt others would, unless they're using a built in one like Microsoft ICM.

[Colin Southern]

I've found that if the monitor is calibrated and has an icc file that browsers are less important if the icc file is installed system wide.

If the browser isn't colour management aware then any image that uses any profile other than sRGB (eg Adobe RGB or ProPhoto RGB) will display incorrectly, as it will be assumed to be sRGB -- regardless of whether or not an accurate .icc profile is installed as the default system display profile.

[Richard Lundberg]

I guess you know exactly what to do now, Christina. I kid.
Don't worry about the web, dear, no matter what you do, it won't look right to a whole lot of people. Get your prints right and relax.
There's no clipping in perceptual, as Colin says, starting in Pro Photo converting down to aRGB and printing with the correct printer and paper profiles will bring in yellows and oranges that nothing else will, as well as getting most of the blue the printer can do.
I sure hope everybody's apps do rendering the same. Never heard any criticism and I've been listening for 18 years.

[ajohnw]

It depends on the rendering engine. Adobe products possibly all use the same one (although I haven't checked), but I doubt others would, unless they're using a built in one like Microsoft ICM.

I only work in sRGB Colin and can not see me ever buying a ProColour printer and if anyone sends me an Adobe RGB image I wont see it as they intend me too. Monitors may change in the next few years and I might then too. I mostly use Opera which comes colour managed anyway however at some point last yea/year before I understand it didn't which is when I set up the profiles.

John
-

[Mike Buckley]

For the record: even though a browser may be color-managed, that does not mean that it is displaying colors accurately. To demonstrate that, try a bunch of browsers viewing the same photos and determine which ones don't display your photos on the Internet the same as when using your post-processing software.

I use Firefox because it displayed an image properly when Explorer and Chrome did not. I realize that that's neither scientific nor exhaustive, but based on my anecdotal experience I'm happy with Firefox so far.

[Colin Southern]

I only work in sRGB Colin and can not see me ever buying a ProColour printer and if anyone sends me an Adobe RGB image I wont see it as they intend me too. Monitors may change in the next few years and I might then too. I mostly use Opera which comes colour managed anyway however at some point last yea/year before I understand it didn't which is when I set up the profiles.

John
-

Wonderful for you John - just not quite sure why you're telling me this?

[Macmahon]

Since the web has become so pervasive, many "experts" will tell you to calibrate to 2.2 gamma and 6500 color temperature because this is the basis for sRGB and people will generally be viewing web images with monitors that are somewhere near sRGB.
...
Another note about gamma. Gamma refers to the exponent of the transfer function of a pixel's amplitude in either current or voltage to luminance of the monitor screen. Gamma as used in calibration actually means the gamma correction, the amount of gamma that is eliminated. 1.8 "gamma" results in prints better matched to paper reproduction. Most monitors have a native real gamma that varies quite a bit from the 2.2 assumed for sRGB and each color has its own. Modern LCD monitors are tailored internally to 2.2 using software or firmware. You can still cal them to 1.8 correction. Correction is done so that the luminance is linearized to provide the eye with the signal it expects, since the eye has its own gamma that varies with light level.

Hmmm. Experts do suggest 6500K and a tonal response curve gamma of 2.2, but not just because it's near to sRGB. The tonal response of unadjusted LCD monitors would be such that dark tones, which human eyes are used to being able to discriminate, would be blocked up, and the monitor would give us heaps of separation in highlight tones that our eyes can't make much use of anyway. The tonal response curve correction, with gamma 2.2 is to provide a correction that enables us to discriminate detail in dark tones. sRGB has a TRC with gamma approximately equal to 2.2 for the same reason.
It makes sense to begin a calibration of a monitor by setting the TRC gamma to 2.2 because it's a useful 'dark-tone-separating' place to start. The act of creating a monitor profile with your Spyder or ColorMunki or whatever will, along with colour adjustment, modify the TRC more specifically for your monitor anyway.

If you want to match your screen and your inkjet, your monitor should be calibrated at 1.8 gamma and 9300K.

I understood that a TRC gamma of 1.8 was recommended in the days of CRT Monitors when they were trying to match to CMYK print. Accepting that profiling the monitor will provide a 'more correct' gamma, few 'experts' recommend initial calibration to a gamma of 1.8 now that most people use LCD/LED monitors whose tonal response is quite different.
I'm intrigued by the suggestion of setting the colour temperature as high as 9300K. Most of the debate is around 5000K vs 6500K. 9300K certainly makes colours on my monitor look very strange indeed.

You also have to make sure your printing app, Photoshop for example, recognizes your printer AND YOUR PAPER.

Absolutely!

The color space of choice for RAW conversion, ProPhoto, is 1.8 gamma for good reason.

Hmmm. Actually the default working space for Lightroom, and AFIAK, ACR (Adobe Camera Raw - the raw convertor part of Photoshop) is a modified version of ProPhoto (called Melissa RGB): it has the same colour primaries as ProPhoto, but a linear tonal response curve, i.e. a gamma of 1 rather than 1.8.
There is no logical connection between the TRC of the internal working (or editing) colour space and the monitor colour space. In a colour managed workflow the system converts an image's colour data from the internal colour space to the monitor colour space whatever the gamma values of the internal space and the monitor space happen to be.

People whose opinion I respect think that Ken Rockwell should be ignored for the most part. He lives in an alternate universe.

I'm not sure how much of all of this will help Christina.
I think the bottom lines are:
* Don't unnecessarily throw away colour data in raw conversion (i.e. don't limit yourself to sRGB);
* Use a profiling tool (Spyder, Colormunki, or i1 etc) to calibrate and profile your monitor for it's colour space and tonal response;
* Make sure you correctly invoke the correct profile and media type for your paper/ink/printer when printing;
and everything will be OK!

Cheers

Tim