Prints and Adobe RGB vs. sRGB

-->Should I not be working in the Adobe RGB color space because I'm on a PC?

The choice of a working space has nothing to do with the operating system!

-->Or should I say, is it smarter to run at 2.2 6500k sRGB until I transition over to MACs?

The TRC gamma and the white point of the display has NOTHING to do with the TRC gamma and white point of your working space. You can use whatever working space you wish! That being said, Mac users should calibrate the TRC gamma of their displays like their PC brothers (2.2). But again, this has nothing to do with the RGB working space you edit your files in. With a good ICC profile that describes your display, you can work on either platform and produce the same color appearance from the same working space.

-->d strange results with different labs. Either the photo looks like it was printed in sRGB from an embedded Adobe RGB color space or it's always 25% darker in the actual print than on my monitor.

The issue was the lab isn't color managed and when they got a file that wasn't in sRGB, their heads exploded. With someone that is working properly with color management, you can hand them data in either sRGB or Adobe RGB and they will properly convert that data to their printer color space (NO output device other then a display can output sRGB).

-->They claim that no matter what color space you submit your photos in (adobe RGB, sRGB) they will properly convert the image so no prior conversion is needed on the photographers end. I was skeptical from the start when I read this, but simply didn't know enough about color management to turn away from their services. Me and my wife prefer to work in Adobe RGB from the camera>photoshop>to print.

You need to have a conversation with them and see if indeed they do use proper color management or simply tell you this. Most Frontier labs want sRGB because they are too lazy to setup the machine to work properly with true output profiles so they want data in sRGB so the system can simply assume that all files are in that color space for the eventual conversion to the print color space. That CAN work but it's not ideal and if they are doing this, giving them a file in Adobe RGB will produce quite unacceptable color output.

-->if I drag an image from the Adobe RGB CRT over to the sRGB LCD...

I seriously doubt you have an Adobe RGB (1998) CRT so you need to clarify what you're referring to. There are about 500 such displays in the entire world (I have one) and they cost more than most cars.

You might want to read the following article (it's general enough that much of it covers some stuff you're talking about, just ignore the stuff about the E1):

(link) vnu_content_id=1000734256

I thought that this was the reason why Frontiers want sRGB files: The jpegs from consumer digicams are in sRGB colorspace.

You said your monitor was calibrated. How? And how long ago?

Interesting letter to editor in last month Photo District News, the writer ranting about how most experts suggest working in the AdobeRGB Colorspace, and goes on to say AdobeRGB is in fact, an extrememly wide color space that cannot be matched by any print nor other output deive, including both offset presses and inkjet printers. He continues saying the big divide at present is that so many digital experts have been touting Adobe RGB as the space to work within, though almost every part of an image file will be rendered "out of Gamut by doing so. Just check the Gamut Warning under View in Photoshop for any such AdobeRGB file if you are in doubt, etc.etc.,,He goes on saying the leading digital lab in Atlanta who say point blank that nearly every printer and diital house in the buisness works in sRGB, despite what we're being told by the so called experts, personally I shoot in sRGB and do corrections in sRGB and submit as request by my lab, the largest on the central coast, in sRGB. I guess my question is if the big dogs and color experts cant get on the same page, what chance does the average digital photographer have of steering the proper course.

> Interesting letter to editor in last month Photo District News, the writer > ranting about how most experts suggest working in the AdobeRGB Colorspace, and > goes on to say AdobeRGB is in fact, an extrememly wide color space that cannot > be matched by any print nor other output deive, including both offset presses > and inkjet printers.

Kind of nonsense. NO output device can produce sRGB but a display that's producing sRGB. It's based on the behavior of a display with very specific aim points (gamma, white point, chromaticity values even the ambient conditions this display is setting in). Adobe RGB (1998) also have (or will have) a defined reference medium (Adobe is working on this). A reference medium is simply the specification under which this color space is supposed to be viewed, NO printer can produce either sRGB or Adobe RGB (1998)! However, when you bring images into these color spaces (a processes known as encoding), you get to decide the gamut volume you can use for eventual output. Many, many output devices have some gamut areas that exceed sRGB (and in some cases even Adobe RGB (1998). Even if 10% of some areas fall outside sRGB, is that important to you? Having that 10% clip to sRGB when you could have used it on your output device may or may not be critical to the quality of the output. But if you clip the colors, they are gone, you never had the opportunity to use those colors.

> He continues saying the big divide at present is that so > many digital experts have been touting Adobe RGB as the space to work within, > though almost every part of an image file will be rendered "out of Gamut by > doing so. Just check the Gamut Warning under View in Photoshop for any such > AdobeRGB file if you are in doubt, etc.etc.,,

The gamut warning in Photoshop is pretty useless. If you're talking about seeing a gray (by default) overlay of the colors that can't be printed, that's there for old timers who used say the sponge tool to manually desaturate those portions of the image to fall within gamut. That was fine in the 20th century, but today we use good output profiles and robust gamut mapping to do this. That's why it's so important to have good output profiles for the process and to pick the rendering intent that YOU prefer. That does a far better job of gamut mapping (either clipping or compression depending on what you pick).

> He goes on saying the leading > digital lab in Atlanta who say point blank that nearly every printer and > diital house in the buisness works in sRGB,

That's total BS. There isn't a printer on this planet that produces sRGB. What these printers do is ASSUME that the color space is sRGB so they can convert the data to the actual output color space. If they profiled the device and provided you the output profile, you could do this yourself, see a far more accurate soft proof and decide how to map out of gamut colors using the rendering intent YOU prefer. I repeat, there are NO output devices that produce sRGB (or for that matter Adobe RGB (1998) or any other working space that's based on the reference media of a display).

> despite what we're being told by > the so called experts, personally I shoot in sRGB and do corrections in sRGB > and submit as request by my lab, the largest on the central coast, in sRGB.

That's fine and if you are happy with the output, end of story. However, your capture device and most output devices CAN produce more colors then you are funneling the initial data into (sRGB).

> guess my question is if the big dogs and color experts cant get on the same > page, what chance does the average digital photographer have of steering the > proper course.

This guy doesn't sound remotely like an expert to me. Who was it?

I always thought an RGB monitor had a larger colour space than either Adobe 1998 or sRGB and the reason for using a reduced colour space is based on the fact that a reflective medium like a print is not able to diplay the wide range of colurs that a monitor can. The other reason for using sRGB is that with a reduced range of colour, a onitor does not need to be terribly well calibrated for the colours in the image to fit into its gamut. someone else mentioned that out of gamut warnings in PS did not help that much, however they are speaking of RGB images. Remember that PS was designed for print and CMYK images where the colour gamut is much reduced over sRGB. Many inkjet printers can take advantage of a wider colour space than can photographic printers and as such you have to pick the working colour space to match the output you are planning for. for example I am shooting a wedding tomorrow that will be printed in a lab. All images will be kept in sRGB colour space, However if on the weekend I go and shoot images that are designed to be printed on my Epson 2200, they will be edited in Adobe RGB

Marrisa, Before you change to a decent lab that can demonstrate at least a grasp of colour management, ask your current one for an icc profile of their printer, ink and paper. you can then use this to proof your image in photoshop prior to submitting them the job. The sRGB or Adobe RGB spaces are commonly used because their wide gamut means that they can hold images from many different sources and allow you to make quite large changes in an image whilst being confident that the colours held are a true representation.

-->I always thought an RGB monitor had a larger colour space than either Adobe 1998 or sRGB.

Nope. Most current technology displays (CRTs) behave (or can be set to behave) as sRGB devices. That's what sRGB is and was designed for.

If you view an image in sRGB outside an ICC aware display and it looks good, the display is behaving closely to the sRGB specifications. Or to put it another way, if an image looks good in the sRGB color space, that image truly is in sRGB being viewed on an sRGB device (the display). The reference medium that defines sRGB is a display in a quite exacting environmental condition (ambient light). This is another reason that when someone tells you their printer produces sRGB, they are full of hot air. It's simply not possible. The data has to be converted and rendered to the output (print) based on it's reference medium.

Most Frontier labs want sRGB because they are too lazy to setup the machine to work properly with true output profiles

Then why don't you set up your own Frontier shop Andrew vs accusing business owners who are risking their own capital of being 'lazy'? This is akin to complaining that amatuer mini-labs don't use professional paper, but you don't use professional labs because they are too expensive.

I'm not thrilled with having to downgrade into sRGB from AdobeRGB either because it does result in a reduced printed gamut range on Fuji CA, which to my frustration always seems to happen with my macro work. Reds and magenta's especially get castrated.

However, if I'm being that particularly anal about an image, I'll flip it over to a LightJet lab that's profiled out, which most are, and use them vs hassle a shop that's handling a 95% clientale that has no clue what AdobeRGB is.

Is there a US online lab that does online proofing, physical proofs, online orders from my customers and runs strictly in Adobe RGB??

Isn't Mpix.com a profiled shop?

Mpix.com (among others) do color profiling but AFAIK do not support AdobeRGB.

PDN reader writes: "AdobeRGB is in fact an extremely wide color space that cannot be matched by any print nor output device, including both offset presses and inkjet printers." The new Epson R1800 is claimed to produce 97% of the AdobeRGB colorspace.

-->PDN reader writes: "AdobeRGB is in fact an extremely wide color space that cannot be matched by any print nor output device, including both offset presses and inkjet printers." The new Epson R1800 is claimed to produce 97% of the AdobeRGB colorspace.

This is a straw man argument. The PDN reader isn't providing the right answers here. Are there any output spaces that fully fit within the Adobe RGB (1998) gamut? Most likely not. But the shape of that gamut is very simple since it's based on very simple parameters which describe how an idealized display behaves. Are there colors in an output and capture space that fall outside Adobe RGB (1998)? Most certainly. Even if only 10% of the output space in various areas falls outside of Adobe RGB (1998), those are the 10% of colors at the extreme saturation you will NOT reproduce. This issue isn't "let's fit all of the gamut from one into the other." it's "what part of the color gamut do you want to reproduce. If you plot in 3D the gamut of an output device, a very large portion might fit easily inside Adobe RGB (1998) with space left over. But if say a good portion of saturated greens fall outside AND reproducing those greens are important to your image, then you need a larger container (working space) gamut.

As for sending Adobe RGB (1998) to a lab, this isn't any different then sending sRGB. You basically have two options here. You send them the data ready to go to the output device IN the output color space for that device (you need an output profile). It doesn't matter what the original color space is, you are providing the file in the color space for that device.

Other option is you send a tagged RGB file and the lab either recognizes it and does the conversion for you or like most dumb labs, they assume some color space (they don't recognize the color space) and unless you send the file in the color space assumed, the output is ugly.

It's not at all difficult or even time consuming to setup a device (even a Frontier) to recognize an embedded profile (be it sRGB, Adobe RGB (1998), or any flavor of RGB) and convert that data to the output space. But if the device assumes everything is sRGB, then sending anything but sRGB will hose the output. It's as simple as that.

And again, the output device isn't producing sRGB (or Adobe RGB (1998), it's producing the color space of that specific output device.

So if one is happy with Frontier printing places (which only accept sRGB files) it is a complete waste of time capturing adobeRGB in camera and working with adobeRGB in PSCS. It would be better to have a complete sRGB workflow from beginning to end, is this correct?

Only if your lab demands files in sRGB. If you have a profile for your lab, you can still shoot in Adobe RGB, convert to their color space, and avoid sRGB altogether.

-->So if one is happy with Frontier printing places (which only accept sRGB files) it is a complete waste of time capturing adobeRGB in camera and working with adobeRGB in PSCS. It would be better to have a complete sRGB workflow from beginning to end, is this correct?

Not at all. Other than the time it takes to convert the data (and the tiny loss of data), capture and convert to Adobe RGB (1998) or wider, convert to sRGB for the labs. Once you initially convert into sRGB, the additional colors you MIGHT be able to use today or in the future is gone forever. Of course you can hold onto your RAW files and now you have all the color potential of the capture device...

First let me say thank you to everyone who gave feedback. Andrew especially.

So basically if I can find a lab who correctly handles/identifies Adobe RGB and sRGB files as two different types of files, then I should be okay? If you go to printroom.com and click on about us, and then click on the color management link on the left. They clearly state that they do the proper conversions for both color spaces. So either they're just bs'ing or they still don't have an accurate enough conversion going on?

And they do have an icc profile to download. But, when I use it to softproof I do not see the changes that I see in my physical prints. I see my reds lose richness and look a little washed out and orangeish. Then my black levels also lose punch, they drop down a few notches in darkness/richness. So I'm unsure how this icc profile that they gave me to use is actually accurate because I see totally different results in my physical prints such as oversaturation and sunburned skin tones.

And no, I don't own an actual Adobe RGB CRT monitor :). I simply meant that I work with an extended desktop with two monitors and one of them is a CRT working in Adobe RGB while the laptop is for my palettes; in sRGB.

I'm still unclear as to why I should continue to work in Adobe RGB. It just seems like there's not much reason if every printer out there is converting to it's own color space that has far less color range than even sRGB (from what I've read and seen from most people's posts). Maybe I'm still missing something.

Andrew, what color space do you primarily work in? What would be the real pros and cons of just working in sRGB? Why work in Adobe RGB if it's only going to look that way on my monitor and not in the final print?

I want to focus my time on becoming a better photographer, not on all this friggin' color management stuff. Of course I understand that it's important for a digital photographer to know all of this. But a simpler, more universal system should be in place so photographers can shoot, image, and print without having to worry about their prints looking inaccurate to what they saw before leaving their computer. It's also irritating having to research the globe to only find that everyone is divided about the subject.

It doesn't seem like Mpix.com offers online proofing and ordering for wedding photographers, but thanks for the link. Does anyone know if collages.net (burrell) works with both sRGB and Adobe RGB files? And converts them correctly?

Here is the definition of sRGB colorspace.

The title is A Standard Default Color Space for the Internet - sRGB. Note that it says "Internet", not "Printing".

sRGB is predicated on a number of assumptions:

• People may have inexpensive, small-gamut monitors. sRGB chooses this gamut because it is "the lowest common denominator". You may easily have a monitor that outperforms sRGB.
• The gamut is tailored for CRT display, not printing or digital cameras.
• The authors assume (correctly) that most people do not have a calibrated monitor, that digital image files do not have embedded ICC profiles, and most image viewing programs do not know about ICC profiles.
• They also make the (rather poor) assumption that most people do not want ICC profiles in their image files, viewing programs, or operating systems. While it may have been valid to think that ICC profiles add too much extra overhead ten years ago, computers now have enough speed and storage to make ICC profile overhead insignificant.

The sRGB standard therefore assumes that digital imaging is an uncalibrated system for most people. The goal of the sRGB standard is to impose a colorspace that everyone gets by default. That way, images on the worldwide web will display reasonably accurately on people's CRTs. Since the CRT is uncalibrated and unprofiled, and the display software is not performing a colorspace conversion, the gamut of sRGB is equal to the gamut of a cheap CRT (circa 1991). This limited gamut is "baked in" to digital image files.

Unfortunately, because of this need to address the "lowest common denominator", sRGB is a compromised colorspace.

For photography, there is no reason to restrict yourself to such a limited colorspace. Your capture device (digital camera, or scanned film) can capture a wider gamut than sRGB. Photoshop handles colorspaces in a competent manner. Your modern monitor exceeds sRGB, and you can get a reasonably-priced calibration kit. Printer gamuts do not match sRGB (and future printers will improve).

What Andrew suggests is that you do not compromise the quality of your results by using sRGB, when better tools are at your fingertips, and there is little or no cost in using them.

> So basically if I can find a lab who correctly handles/identifies Adobe RGB > and sRGB files as two different types of files, then I should be okay?

IF you want to send data that's in an RGB working space (which is not an output space) to a lab and they are going to do the conversion to the output space AND they say they will accept (recognize) sRGB or Adobe RGB (1998), you're OK. You don't get to soft proof the image to the printer or pick the rendering intent. > go to printroom.com and click on about us, and then click on the color > management link on the left. They clearly state that they do the proper > conversions for both color spaces. So either they're just bs'ing or they still > don't have an accurate enough conversion going on?

No, they are (hopefully) recognizing the embedded profiles from those RGB working spaces and doing the conversion for you. Better would to get the actual output profile but we're not there yet... > And they do have an icc profile to download.

Great!

> But, when I use it to softproof > I do not see the changes that I see in my physical prints. I see my reds lose > richness and look a little washed out and orangeish. Then my black levels also > lose punch, they drop down a few notches in darkness/richness. So I'm unsure > how this icc profile that they gave me to use is actually accurate because I > see totally different results in my physical prints such as oversaturation and > sunburned skin tones.

Well there are a few ways to setup a soft proof (like using the Simulate Paper White/Ink Black etc) and of course you are expected to be viewing the prints under a controlled light (a CCT 5000K light box). If all that's setup and the soft proof is still "off" it could be the result of the display profile or the output profile (or both). But we are not there yet...

> And no, I don't own an actual Adobe RGB CRT monitor :). I simply meant that I > work with an extended desktop with two monitors and one of them is a CRT > working in Adobe RGB while the laptop is for my palettes; in sRGB.

No I don't think so. Both displays are probably much closer to the behavior of sRGB but the bottom line is you need to calibrate and profile the displays. > I'm still unclear as to why I should continue to work in Adobe RGB. It just > seems like there's not much reason if every printer out there is converting to > it's own color space that has far less color range than even sRGB (from what > I've read and seen from most people's posts). Maybe I'm still missing > something.

That's not the case. There are lots and lots of printers that have areas that exceed sRGB. > Andrew, what color space do you primarily work in?

Adobe RGB (1998) or ProPhoto although I have "legacy files" made in the 90's in ColorMatch RGB.

>What would be the real > pros and cons of just working in sRGB? Why work in Adobe RGB if it's only > going to look that way on my monitor and not in the final print?

That's not the case. You can work in Adobe RGB (1998) and use some of the colors you have at your disposal. If that's not an issue, use sRGB. > It doesn't seem like Mpix.com offers online proofing and ordering for wedding > photographers, but thanks for the link. Does anyone know if collages.net > (burrell) works with both sRGB and Adobe RGB files? And converts them > correctly?

Burrell doesn't have a clue about color management (unless they've done a total 360 in the last few months or so). They expect you to tweak (read screw up) your display to match a print they send you. Really bad idea!

Marissa,

You may find the following web-site useful - it graphically (no pun intended) represents various printers colourspaces in comparison to Adobe RGB.

http://www.drycreekphoto.com/tools/printer_gamuts/

I have yet to try Drycreek, but had downloaded their instructions here:

http://www.drycreekphoto.com/custom/customprofiles.htm#Profiling_kit_downloads

They provide a target to test your printer first before ordering a profile. The document also cautions that some colors cannot be produced on some printers, regardless of profiles and color spaces, etc. I give them a lot of credit for stating this.

So my background is math and computer science, and there's an important observation regarding the sRGB/Adobe RGB religious wars that seems completely obvious to me, yet almost never gets mentioned by photographers or color geeks:

Everybody knows that AdobeRGB represents a wider gamut, but everybody also seems to assume that this means it is "better" somehow. Why? Using a larger color space doesn't magically give you more colors to work with. It spreads the same number of colors across a wider area. This is because no matter how many bits per channel you work in, there are a finite number of states that you can use to represent colors. If you insist on stretching your color space to include lots of primary greens that can never exist in nature and will never be printable, you have to devote some of these states to representing those unusable colors. Those states are wasted, leaving you with fewer to allocate to the colors that can actually be used. The wider your color space, the coarser your color gradations will have to be and the more banding.

Thus it seems to me that the ideal working color space would be one that represents the full gamut of your destination output device, and not one color more. If sRGB is larger than your printer's actual gamut, you are already wasting some of your workable bits and using an even larger space would just waste more.

Now if the argument is that one should preserve more information in order to take advantage of future improvements in printer technology, fine, but I find it hard to believe that most working professionals (most of the people that care about color calibration that is) really care about that. How many event photographers or journalists care about making prints of 10 or 20 year old pictures? Fine art photographers maybe. I'm none of these things so I don't know.

The concept of relative size is not adequate in color spaces because they overlap. Rather, one must ask larger or smaller in which part of the gamut? The issue of posterization has been discussed repeatedly, but does anyone read the archives?

Bear in mind that technology is developing. The gamut you can get using a new inkjet is much greater than what you could get ten years ago. A few years from now, if you need to make a print of a current picture (e.g., because the print faded!) you will be obviously be using newer technology, so why limit the images to a small color space?

Thanks Mark and Robert for the drycreek links, I'll definitely check those out.

Andrew, what do you make of Michael Dickerson's post? This sounds like it makes sense but at this point I've swayed in so many different directions I don't know what to believe.

Oh, and I picked up a 12x18 engagement shoot print today at my local pro lab and it looked exactly like what I had on my monitor. They prefer working in Adobe RGB. Thing is, they charge quite a bit more for prints than online labs so that's why I'm trying to still find a pro online lab that accepts and prefers Adobe RGB. An online lab that caters to the pro, not the consumer point and shoot end.

Oh, I contacted mpix.com about what printers they use and what color spaces they prefer. This is what they told me...

Yes, I have attached our ICC printer profiles for soft proofing in Photoshop. Uploaded files should be in sRGB color space with no embedded profiles. Mpix outputs with Durst Theta 50 and Kodak RP30 printers. profile's internal name (MPRP30E1_02_26_04.icc and MPThetaE5_02_26_04.icc) Mpix printers output in sRGB color space. Which color space is correct? ICC color spaces are simply mathematical representations of a particular device?s color gamut or the color gamut of the objects you are attempting to replicate. Fundamentally understanding color space differences may seem simple. Wider spaces encode a wider range of color than smaller spaces. So, on the surface, it would seem to make the most sense to use the largest color space possible. sRGB is one of the smallest color spaces, so why is it our default color space? There are two main reasons. First, we choose sRGB as our working, preferred color space because most of the current digital cameras capture in sRGB. Second, the digital photographic printers we use are also sRGB devices. Using a wider color space is moot if you are capturing in sRGB and outputting to sRGB.You can not increase the color space of an image since the color space of an image is defined at the moment of capture. Translating sRGB images into and out of other color spaces is not only time consuming but of little benefit.We use sRGB because it simplifies the amount of work that is necessary for you to prep a file for submission. As always our goal is to partner with you and offer solutions that both meet your high quality standards and workflow as much as possible. All orders coming in are reviewed by a color tech and given a color correction on a calibrated monitor before it is submitted to the printer. We do offer a 'no color corrections' list were your files are sent directly to the printer without any adjustment from our tech's. I would be happy to flag your account for no color corrections if needed. Mpix Customer Service

> Using a larger color space doesn't magically give you more colors to work > with.

No it doesn't, it provides the potential./

>It spreads the same number of colors across a wider area. This is > because no matter how many bits per channel you work in, there are a finite > number of states that you can use to represent colors. If you insist on > stretching your color space to include lots of primary greens that can never > exist in nature and will never be printable, you have to devote some of these > states to representing those unusable colors.

But the colors DO in many cases fall into the scene and CAN be output. 100% certainly not. Again, if 10% of the colors you could capture are able to be placed in the Adobe RGB (1998) container but not in the sRGB container and you want to reproduce those 10%, one allows this, one doesn't. It's pretty simple.

R0/G255/B0 is just a number and doesn't tell us how to reproduce that color. We need a scale. How green is G255? Well in Adobe RGB (1998) it's the same values as in sRGB but the scale tells us that the green is more saturated in Adobe RGB (1998) than sRGB.

> Those states are wasted, leaving > you with fewer to allocate to the colors that can actually be used. The wider > your color space, the coarser your color gradations will have to be and the > more banding.

And why with progressively larger color spaces you want to be working in more than 8- bits pre color. But at least in the last 8 years, Adobe RGB (1998) has proven to work quite well with 8-bit data. > Thus it seems to me that the ideal working color space would be one that > represents the full gamut of your destination output device, and not one color > more.

Exactly but that's simply not possible for a number of reasons. For one, the gamut of the working space is based on (almost always) the reference medium of a display. The reference media and gamut of a printer isn't anything like that. So there's always a bit of a disconnect.

>If sRGB is larger than your printer's actual gamut, you are already > wasting some of your workable bits and using an even larger space would just > waste more.

IF but I doubt that's the case in most output devices for the reasons above. The sRGB color space is based on the behavior of a display (a very specific display condition). There may be some output devices who's gamut fully falls within sRGB but most do not (and in some cases that's even true with Adobe RGB (1998) but the issue is less; more output gamut is contained in that space). But you'll likely never get a prefect fit. The SWOP gamut (as defined by SWOP TR001) fully fits inside of Adobe RGB (1998) but not sRGB. So there's at least one example. > Now if the argument is that one should preserve more information in order to > take advantage of future improvements in printer technology, fine, but I find > it hard to believe that most working professionals (most of the people that > care about color calibration that is) really care about that.

Then they are pretty dumb! If you look at the progress made in digital imaging and output in the last 15 years, this would be a grave mistake.

> How many event > photographers or journalists care about making prints of 10 or 20 year old > pictures? Fine art photographers maybe. I'm none of these things so I don't > know.

Speaking as someone that spent a large amount of time making a living as a photographer and how spends most of his time working with that group, I'd say a lot. Why paint yourself into a corner with your data. It's like you can go back to your original chrome and rescan it. When you press the shutter and tell your digital camera to toss the RAW data and funnel that color into sRGB, you're stuck with what you get FOREVER.

If photographers didn't care that much, you wouldn't see the discussions of the archival nature of print and even film clogging up the net as you do. Photographers DO care!

-->First, we choose sRGB as our working, preferred color space because most of the current digital cameras capture in sRGB. Second, the digital photographic printers we use are also sRGB devices.

Write them back and tell them I told you to tell them that's nonsense.

(Getting offtopic, sorry, will stop).. but thanks Andrew; I was expecting flames, not well reasoned argument containing useful information I didn't know. I didn't know printers commonly produced gamuts that don't fit inside sRGB for instance, since the labs claim they do not (as Marissa is finding).

My camera (*ist DS) lets you capture in jpeg/sRGB, jpeg/AdobeRGB, or raw. I experimented with all these things but was skeptical that the camera is actually capturing a wider color space in AdobeRGB mode (as opposed to performing the exact same reduction from RAW and just mapping the results to AdobeRGB numbers). If the recorded data really is different then I should rethink the color space issue, since I'm also one that believes in discarding no information as you suggest. Unfortunately, raw files are too big for me most of the time (I mostly get to take pictures when traveling and the Pentax camera's raw files are 10-11 MB).

-->My camera (*ist DS) lets you capture in jpeg/sRGB, jpeg/AdobeRGB, or raw.

All of the above are initially RAW. When you pick sRGB or Adobe RGB (1998) you're telling the camera to render the color into an image and then encode that data into a working space. This is a two step process. The rendering is totally up to the manufacturer to decide how they think you'll like the color. That's why if you setup two different cameras from different manufacturers (even differing models from the same) shot the identical scene and ask for say sRGB, you'll see two different looking images. In a way, that's OK (assuming you like the rendering). It's a lot like when we used to pick a specific film due to our preference of how it rendered the scene. Fujichrome and Ektachrome of the same scene don't look the same. Which do you prefer? So this rendering (on the fly) is totally out of your control. The encoding into a working space is not at all ambiguous; it's standardized. If you shot the scene and rendered that from two cameras, the encoding into sRGB would be identical (but of course the images would look different). Of course, if you shoot RAW, you get full control over the rendering (from what's called scene referred to output referred). IOW, when you shot an scene, you're not recording the colorimetry of the scene (the measured color), you're recording the scene as you want it to appear, in this case in sRGB. So the camera isn't producing either sRGB or Adobe RGB (1998) but you can take the original RAW data and render that and encode that into those spaces. I guess if someone wants to say, based on how this all works "the camera creates sRGB" OK. But that's not how it has to work and in reality, there's a huge amount of differences (based on the rendering your manufacturer feels is "correct").

-->I experimented with all these things but was skeptical that the camera is actually capturing a wider color space in AdobeRGB mode.

The best way to see this in action is to shoot a scene in RAW with a very wide gamut and dynamic range. That's not too hard to do. Then open in Adobe Camera RAW and look at the histogram and how it indicates clipping as you change the four RGB working space. Also as you control the rendering (with all those nice sliders). I've seen with me 300D many scenes where even Adobe RGB (1998) will show color clipping while picking ProPhoto RGB will contain the data. I set this working space and bring the data into 16-bit and of course I control the rendering of the data (scene referred to output referred).

I have read this thread with great interest and it is nice to see the subject getting such an informed airing.

Labs are among the greatest culprits when it comes to spreading mis-information largely because photographers tend to assume that their lab is staffed by experts in colour management and that what they say must be correct.

My experience is that many lab staff do not understand colour management but propagate whatever myth best suits them, possibly with good intent (I am charitable). Happily I print all my own work (Epson 4000), no need for a lab.

One friend of mine is a skilled photographer but more artist than technician. His lab has told him to use his D1 set to sRGB (high class weddings and social photography) as that will give him the best colour. They also suggested that the different versions of Photoshop (v6, v7 & CS) all handle profiles differently and would cause his output to vary in colour depending on the version of Photoshop in use (no wonder he is confused).

The reality is that if he sends sRGB images to the lab, the prints are ok. If he sends Adobe RGB images to the lab, the colour is poor. In his mind this means that Adobe RGB is bad and sRGB is good (re-inforced by the lab). I have just spent a couple of hours explaining to him why that is not the case. I suspect that his lab is geared up for sRGB and not converting the colour space.

I have asked him to take three identical photos, one RAW, one Adobe RGB, one sRGB to view in Photoshop CS. Hopefully he will be able to see how data is being lost as the colour space decreases in size (Gretag Macbeth software includes the ability to view relative colour space, fascinating to compare monitors with printers with adobe with sRGB etc.)

He is then to send three photos to the lab for printing, one in sRGB, one in Adobe 1998 and one converted from Adobe 1998 to sRGB. Theoreticall all three prints should come back looking about the same.

I expect, however, that the Adobe print will be very different, indicating that the lab is not managing colour properly.

My advice to him is to use Adobe 1998 as his working colour space and when he copies the prints to cd for sending to the lab, run a batch to convert them to sRGB. That seems a reasonable compromise.

Les