You should contact the lab beforehand, and find out what they expect to see in file parameters. Adobe RGB (1998) is pretty much the standard for print media, but I've heard that some labs want to see sRGB, which is the standard for the web.

Thomas.

The simple answer is YES you loose something. Bill is correct both JPEG and sRGB are Web standards. And they are used for putting images on web sites for one reason, because the file size is much smaller, and the reason is because it takes less time to transmit and display the images. Both JPEG and sRGB obtain smaller file sizes by throwing out information by SIMPLIFYING the information. The result is if you maybe have a dozen shades of red in an image you can make it smaller in file size if you reduce the number of shades of red to half a dozen. In reality for instance a picture you take of your family in Adobe RGB reproduces each member with a distinct skin tone, when the photo is converted to sRGB all the member of your family in the picture have the same generic skin tone. And like JPEG once you convert from Adobe RGB to sRGB the loss of color information is permanent, you don't get the color information back if you convert back to Adobe RGB.

All fleshtones are within the gamut of sRGB. It's certain rich greens and blues that can get clipped as seen when capturing brightly lit flowers and high gloss lustre finish expensive paint jobs on customized cars.

Adobe RGB has a gamut that is 50.6% of LAB which is 97% of the visible spectrum, while sRGB is 35%, a significantly smaller number of distinguishably different colors. The effect is not just in any one part of the gamut, but that the number of distinctly different variations in color for all is reduced. The difference is practically if two people reproduced by Adobe RGB have subtly different complexion colors, and if converted to sRGB the two complexions would be less distinguishable or would look the same.

I've never seen complexion shifts when converting from AdobeRGB to sRGB. And I've also been told by Bruce Fraser that fleshtones fall well within even SWOP color space, far narrower than sRGB.

Can you link to an image that demonstrates this shift in fleshtones. I'ld like to see it.

Tim,

Bruce Fraser has his good points but he is not a photographer, never has been, and never will be. I have nothing against him personally, and he has contributed a lot of useful information and understanding. But many who work inside of the color management industry have a nickname for Bruce Fraser, Andrew Rodney and a couple of other self-styled color management guru's, the "Pixel Mafia".

If you just look at a comparison of a 2D representation of sRGB colorspace superimposed over Adobe RGB you might logically assume the loss of colors is just around the periphery. But a 2D representation is a gross distortion because colorspace is three dimensional and when you decrease the volume of the space you reduce the number of color variations across the board - the color palette is more limited. Instead of 50 variations of any color there is just 30. The results is a lot of the subtle variations in color disappear.

The same principle is applied to the extreme in motion picture animation where the color palette is arbitrariy reduced for production efficiency and consistency to very few variations of each color. And that is why I associate sRGB with cartoon color.

The palette of photographic reproduction is limited enough compared to the real world of color we see with human vision, so why should one want to reduce it further unless you are publishing on the web where it is essential to keep file sizes small.

I understand completely about 2D and 3D gamut plots. I was talking about fleshtone complexion changes you said would happen when converting from AdobeRGB to sRGB.

I've never seen this happen and thought you could show an image sample to demonstrate this. Since I've already asked for this from you without a reply I assume you don't have an image sample.

And as for the Pixel Mafia, I don't see much of them posting any longer on forums lately. They don't like to explain the technology very well when I ask them a question. I don't know if it's because they want you to buy their book or they have nondisclosure contracts with the major imaging software/hardware corporations. I've seen them hold back on the Adobe forums with these kinds of excuses. I do know they are very well informed with lots of inside information they sometimes like to slip out if you're lucky enough to keep after them on the forums.

If you're speaking of seeing a difference in the Photoshop work space, you'll be hard pressed to see an actual difference on your monitor because Photoshop reads the color tags and displays the images properly. You'll have to actually make prints to maybe see a difference.

Tim,

Why should I take you seriously when you ask for an image sample to be seen on a web site. You cannot reproduce the range of Adobe RGB (1998) on the web anyway.

>>I do know they are very well informed with lots of inside information they sometimes like to slip out if you're lucky enough to keep after them on the forums.<<

They are? How do you know? Because they say so? The main criticism from the color management industry's color scientists is that the Pixel Mafia does not know or understand the science. I sure do not know the science, although I have read a lot of what has been published over the years including the most recent edition of Richard Zakia (Professor Emeritus at RIT) book Perception & Imaging.

If you want some good information support on color management here are a couple of sources: there are a lot of articles by Steve Upton President of Chromix.com on their web site; and their is a lot of information at http://www.brucelindbloom.com - there is also copious material on http://www.apple.com/colorsync/

What I mean about the Pixel Mafia being well informed I was referring to up and coming improvements in digital imaging technology in general.

I cut my teeth on digital imaging starting back in '98 reading Bruce Fraser's Real World Photoshop. I chose his book over the WOW books because he has a background in commercial prepress technology and input/output system workflow which is my background.

But thanks for the feedback. It's an interesting subject nonetheless.

Wow, so much about so little. Since we all perceive color so differently ("color is the most relative of all media"), we are never going to solve these issues unless all our eyes are on the same screen or print. Even then we could argue about the number of different skin tones in a particular image. Everyone could go on forever throwing stuff against the wall but there are no clear answers to many of these questions.
As far as I'm concerned prints are still the most important product of a photographer and their gamut is so small compared to the screen that it almost doesn't matter what color space you look at it in. Then there's the problem that the warmest screens are usually 6500K and most galleries have 3200K bulbs. High quality RAW files are even clipped by the gamut of Adobe RGB never mind ever seeing those colors in a print. Methinks t'would be far better to expend one's energy on imagemaking than trying to figure out what doesn't matter.
There have been some important matters discussed. Everytime you change the colorspace you do hurt your picture. Anytime you change a jpeg and save it you lose as that's the way the format was designed. On the other hand, if you save your Photoshop image as a maximum jpeg at full resolution for transport the print will be virtually identical to the uncoverted one. In case you're still skeptical about jpegs, for those with broadband, Adobe provides in Photoshop's automate functions, the ability to transmit the highest quality file complete with password protection. sRGB represents the lowest common denominator, save it for your snapshots or the web.
NRennnie

>>As far as I'm concerned prints are still the most important product of a photographer and their gamut is so small compared to the screen that it almost doesn't matter what color space you look at it in.<<

NRennie,

Sorry but measurement indicates that an Epson R2400 reproduction gamut exceeds sRGB in two of three dimensions by a considerable amount. So ignore gamut at your peril.

Color temperature matching in print output is easily accommodated in a custom printer profile for any display illumination color temperature.

Digital camera raw files are not usually clipped by Adobe RGB (1998), which will accommodate most of the color range in an Ekatchrome transparency, which it was originally designed to do.

>>Everytime you change the colorspace you do hurt your picture.<<

Changing colorspaces (assignment) does not affect the RGB values in a digital image file unless you actually CONVERT profile to profile from a larger space to a smaller one and then re-save the file.

David,

I am still shooting and saving (in camera) my files in the highest possible JPG quality format but I am seriously considering (finally!) moving to RAW -- hence my interest in this thread.

So, from what I can figure out so far, the best way to put as many "quality odds" as possible on the photographer's side towards a higher quality print would be to stick to the following workflow:

a) Capture in the Adobe RGB (1998) color space
b) Save the photo in the RAW mode
c) Process/Edit the photo in the Adobe RGB (1998) color space
d) Save the edited photo/image in TIFF/non-lossy format (or?...)
e) Save a copy of that file in the highest quality JPG format for compatibility with most labs' requirements.
f) Try to find a lab that works in the Adobe RGB (1998) color space instead of their usual sRGB.
g) Then wait to be delighted or disappointed ... ( )

Did I get things right or did I miss or twist something somewhere?

Thanks, David, for the clarity of the explanations.

cambug

cambug,

Item d): or save in Photoshop's lossless PSD format, either flattened or not (which allows you to save with the layers intact, but takes more disk space).

Frans Waterlander
pixographer

Frans,

Yes, from a purely photographic point of view, PSD seems more logical (since the editing is done in that format) but what happens IF the file contains some ITPC data? (not just the Exif data, but caption, keywords, etc. as well)

I always thought that TIFF retains that data but that PSD doesn't.
[I know for sure that if I load a JPG file that has ITPC data in it in Photoshop, once I edit and save that file in the PSD format, that data is gone/lost.]

Or is there a way to protect/preserve it?

cambug

cambug,

I am not familiar with ITPC; if you have Photoshop, you can try it out.

I know that Photoshop will not keep all the Metadata that is contained in my Nikon D70 RAW file when converting from RAW. However, once in Photoshop, it lets you add file information (File>>File Info...) with the headings Description, Categories, Origin, etc.

Frans Waterlander
pixographer

Cambug,

OK, you have the workflow figured out through "D", but the next logical step is to use a 7-8 color photo printer to reproduce what you have on file.

If you are going to send your images out for machine printing by a snapshot lab, why go to all the trouble of using Adobe RGB, Raw format, the work of manual color correction and adjustment editing. You might as well use sRGB and JPEG in the camera if you are going to have snapshot quality prints made.

cambug,

David Brooks' response implies that photo labs in general cannot provide you with high-quality prints of your digital images and this just isn't true. There are many photo labs, brick-and-mortar and on-line, that can do a superb job and often for less money than if you do it yourself. Personally, I do my own printing because I like to and have never tried photo labs, but many enthusiast and professional digital photographers use these labs. Do a search of this and other photography forums and/or post your question and you will have no trouble getting recommendations for high-quality, low-cost labs.
I can recommend a couple of other photo forums to search or post the question if you email me at frans2001@netzero.net.

Frans Waterlander
pixographer

Frans,

There is a distinct difference in both quality and cost between having machine prints made by a consumer photo finishing lab and having custom prints made by a professional photo lab. I did not intend by my remarks to put them all in the same bag. And as a matter of fact most professional photo labs also make machine prints at close to consumer cost per print. If you want quality prints from TIFF files in Adobe RGB colorspace, you pretty much have to deal with professional-level services and pay the rates custom services require.

David,

Cambug didn't say that he was going to use what you refer to as a "snapshot lab". He said "a higher quality print" and "a lab that works in the Adobe RGB (1998) color space instead of their usual sRGB". That doesn't sound like a "snapshot lab" approach. Yet your response did not make the distinction between a consumer photo finishing lab and a professional photo lab. By commenting only on "snapshot labs" your reply is incomplete and misleading; cambug needs to know that high-quality prints can be had at prices often lower than for do-it-yourself prints.

Frans Waterlander
pixographer

Frans,

I think you misread both what cambug said and what I replied.

Then if he needs to know that higher quality prints can be had for less than what he can print for himself, shouldn't you provide the identification such a resource and one that will also NOT require JPEG and will work in Adobe RGB?

David,

"I think you misread both what cambug said and what I replied."
And what, may I ask you, did I misread? Both cambug's and your words are pretty clear!

"Then if he needs to know that higher quality prints can be had for less than what he can print for himself, shouldn't you provide the identification such a resource and one that will also NOT require JPEG and will work in Adobe RGB?"
I have already indicated resources for cambug to use, but since you ask for it specifically, let me respond specifically:

Cambug, you should search the forums at www.popphoto.com and www.photo.net and if you don't find the information that you are looking for, post your questions on these forums; chances are you will get very useful feedback. Both forums have many posts from knowledgeable enthusiasts and professional photographers that get high-quality prints from sereveral labs. Many of them no longer print themselves because they get top-notch results at the same or lower prices.

Frans Waterlander
pixographer

Frans,

In other words you can only provide what you've read that's hearsay by others to substantiate the claims you make? That's neither real nor very assuring. That's like recommending a second hand car to someone you've never driven and know nothing about.

In other words, David B. Brooks, I am explaining that there are alternatives to your one-sided, incomplete and misleading comments that would have cambug believe that the only choices are inferior prints by a "snapshot lab" or printing yourself. I am NOT making any claims or recommendations. In an effort to point out other possibilities, I am relaying as accurately as possible feedback from many photo enthusiasts and professional photographers and I fail to see how that gives you the justification to badmouth me.

Frans Waterlander
pixographer

Frans,

You jumped in and found fault with my response to another individual message, so you get some of the same back. Do you have anything to complain about when I merely pointed out your information is second hand?

David,

Glad to see that you are still your amiable self.

Frans Waterlander
pixographer

Getting back to Thomas's original post:
>>Lets say I have shoot RAW files in Adobe RGB, and then open them in CS2. If you increase to 16 bit, but convert to sRGB in Photoshop, will you loose anything?<<

The reference to "Increase to 16-bit" is flawed to my mind.
You can't (properly) increase an image to 16-bit colour since the required extra colour data (over 8-bit) is already missing. ??
You can, however, convert a raw file to a 16-bit tiff. Is that what you meant to say?

Pip22,

If you are shooting in Raw then you have the choice in Camera Raw to output to Photoshop in 16-bit preserving the bit-depth of the original Raw file which is now always greater than 8-bits, so all the information captured by your camera is retained.

If you convert Raw saved with Adobe RGB space in camera and then convert to sRGB in Photoshop you will loose as much as 30% of the color information captured by the Raw file.

Quote:
Can you link to an image that demonstrates this shift in fleshtones. I'ld like to see it.

I don't think this can be demonstrated either. The areas where Adobe RGB tends to matter is at the extremes, where colors tend toward pure Cyan and Red...not in the area of skin tones which are complex colors well within the display gamut of sRGB.

Quote:
e) Save a copy of that file in the highest quality JPG format for compatibility with most labs' requirements.

Why convert to JPEG at all? Is it required by your service? While a high quality JPEG (12) is likely to produce almost nothing in the way of visible artifacts, I am not sure why you would choose that over a non-lossy format for submission or archiving. Space should not be considered at a premium these days.

Quote:
In other words, David B. Brooks, I am explaining that there are alternatives to your one-sided, incomplete and misleading comments that would have cambug believe that the only choices are inferior prints by a "snapshot lab" or printing yourself.

I agree with Frans.

Quote:
Both JPEG and sRGB obtain smaller file sizes by throwing out information by SIMPLIFYING the information. The result is if you maybe have a dozen shades of red in an image you can make it smaller in file size if you reduce the number of shades of red to half a dozen.

This seems not clear, and pastes sRGB as an inferior color space. sRGB has as many colors as any other flavor of RGB, those colors are simply mapped differently. Where Adobe RGB may battle for colors at the extremes, there are subtleties of color in the nearer range which it logically has to neglect to allow for expansion in the extremes. I don't know that I would be quick to assert that any conversion is ultimately lossless.

Quote:
If I upload my pictures to some commercial printer/developer, will they recognize the format so that I don't get washed out colors on my print?

All printers and services will not handle color management in the same way. Your best bet to achieve consistent results is to test. The easiest way to do this is to submit a test print as Adobe RGB (tagged) and the same file converted to sRGB (tagged), then submit an sRGB file untagged. If the Adobe RGB and sRGB come out nearly the same, the service is likely honoring your tags. If they come out notable different and the untagged file and sRGB are the same, then sRGB is assumed. If all come out differently, there is another profile assumed. Regretfully not all services know how to best handle files -- and not all "technicians" are long on the job.

>>I don't think this can be demonstrated either. The areas where Adobe RGB tends to matter is at the extremes, where colors tend toward pure Cyan and Red...not in the area of skin tones which are complex colors well within the display gamut of sRGB.<<

Not true at all. It can be readily demonstrated, but not in images that are reproduced on the web because of the inherent gamut limitation of the internet and what is displayed by a web browser.

If you want to demonstrate this the best example is a photo of two people with subtly different complexion color that can be seen in a photograph in Adobe RGB colorspace, and then reduce the colorspace to sRGB and the complexion tones will be less distinguished. When you reduce the size of a color gamut you reduce the number of different distinguishable color variations, so subtle differences across the gamut are less distinguished when a conversion from Adobe RGB to sRGB is made.

Many of the labs like those available on-line often utilize Fuji printers many of which only support sRGB and these labs often do require files submitted in JPEG format.

It is inferior. Adobe RGB has a LAB Gamut Efficiency of 50.6 % compared to only 35% for sRGB. Reference: RGB Working Space Information: http://brucelindbloom.com

Let me rephrase: I am positive that you are incorrect. There is no way for an image editing program to know what two different complexions look like and 'normalize' them. Skin tone is well within the gammut of sRGB. If you are experiencing compression, you are not using color management correctly -- and are probably assigning rather than converting profiles.

I would be glad to see you post an image where this compression occurs. Frankly, in about 20 years of working with digital images, I have never seen it happen when files are handled correctly.

Quote:
Many of the labs like those available on-line often utilize Fuji printers many of which only support sRGB and these labs often do require files submitted in JPEG format.

Just because some labs that you deal with require JPEG files (and I can't imagine why they have this technical limitation) does not mean that all users should automatically convert to JPEG. That is just not good advice. For someone all agog and a-twitter about detail lost in converting to sRGB (and I think your thoughts are mal-aligned there), you certainly make no bones about saving to a lossy (meaning: loses detail) format.

Adobe RGB is compressed when viewed on a monitor, so you never really see the actual color you correct. That is a limitation of monitor display. how is it that you can correct what you cannot see reliably? I have my methods, but I wonder what exactly you are looking at when making your comparisons.

There are also, and I repeat, the SAME NUMBER OF COLORS (16 million+) in AdobeRGB and sRGB...each space having detail enhanced in some area. While there may be limitations in the range of sRGB (so that colors fit to what can be reliably displayed) it is more the rare case that image color will require the enhanced range -- or more likely that the result will be demonstrably different. there are also limitations in Adobe RGB, these limitations are just different (read: 'different', not 'better'). The bulk of color in the visual spectrum is adequately reproduced by sRGB. Your comparison of percentages is misleading. There are many wavelengths that we don't even see, and they far outnumber the 'visual spectrum'. The fringes of the visual spectrum are not necessarily important to image detail or base perception, and these likely make up the bulk of the difference between the color spaces--the percentages do not directly reflect the importance of colors in the range of those sets.

I am not saying at all that there are no advantages to Adobe RGB, but that you are likely over-rating them. You don't have to believe me, but I do think you will have a lot of trouble posting images that support what you say. Thank you for the link to Mr. Lindbloom's site. Perhaps you are also familiar with Dan Margulis? He and I both come from pre-press backgrounds and share many similar ideas on color.

Broader is not always better.

I have been reading Mr, "contrarian by the numbers" Dan Margulis off and on for about 15 years. It is really interesting you should cite him as a reference considering all he did was poo poo color management until very recently.

However he, and his latest book was cited here in the forum as an excellent resource for color-blind photographers who need to do color with Photoshop.

Apparently by flaming me you think you can get even for my turning you in to management for placing a commercial ad in this forum. It won't work because most of the regulars already know where I am coming from, and all but a few will avoid getting involved as they do not appreciate flaming contests. So I will not clutter up the forum with any more responses. But if you want to continue to try me keep up the personal attacks, that is also not tolerated by the moderator and can get you expelled from the forum.

Quote:
if you want to continue to try me keep up the personal attacks, that is also not tolerated by the moderator and can get you expelled from the forum.

There was nothing in the form of personal attacks at all in my other responses. The information you are peddling is not quite right, and I am doing what i would do on any forum to help clarify for anyone who is listening to you. I don't see a horde of people agreeing with you--in fact several others before me have mentioned that they don't believe you are right. I don't think you have the fan club you expect.

I see you have not posted examples. I would be glad to see them, and it would prove your point...IF you can produce them.

About Dan...he was around before color management, and so was I. People who treat it like a godsend are just flat wrong, as somehow there were those of us who got results before color management. i do not poo-poo it, and i don't believe Dan does either...but probably neither of us treat it as a holy grail. There are ways to use it, and ways not to: and not all of them require following the same rules you seem to be doling out. There is a way to separate the sense of color management from the nonsense and make it useful. I do not believe your advice is well developed. I believe I have stated facts as to why.

If you are such an expert you should know that the internet does not support the display of images with a gamut as large as Adobe RGB, and compressing an image small enough to be uploaded to this site does not display anything but cartoon color.

So far your criticism has not been substantive beyond your opinion and is really not worth the effort of responding. I'm not moved to any effort in defense as your arguments are empty of fact. I'll not waste any more time with this pointless exchange as your purpose is obvious and transparent.

Quote:
If you are such an expert you should know that the internet does not support the display of images with a gamut as large as Adobe RGB

And likewise, an expert should know that you can set up a download of any type of file you'd like all safely zipped up so the other party can download and print or examine.

It isn't worth a reply, because you don't have an answer.

Quote:
your purpose is obvious and transparent.

Yes, and that purpose is to get you to substantiate your claim that skin tones are affected when correctly handling sRGB file conversions, or that it is critical to use AdobeRGB in every case (and I don't care if you would now like to limit that to print images--though the orginial premise was not limited that way).

It is not flaming to request substantiation of something I have not seen. Perhaps you know something I don't. But perhaps you are trying to come off as knowing more than you do about this area of imaging where your real strength lies somewhere else. For example, I would guess you are probably a better photographer than I am. In the case of this thread, however, I think you either have a misconception that you can easily correct (you obviously have seen evidence of your claim, or you wouldn't suggest it). If there is evidence I would be glad to see it and adjust my own thinking. If not, perhaps you can adjust your so the next time you phrase your opinion it is a more well-honed one, and in that way more beneficial for people to read.

I have substantiated my claim and provided the research results to others expert in color management. But, you I would not give the time of day.

It stands unsubstantiated then...you pointed to a website that has some numbers on it. You have no examples, because there are not any. You are mad at me because you are wrong.