From: Andrea de Polo
Date: Wed, Jul 11, 2001, 3:49 AM
RE: [colortheory] More wide/rich color space for archival needs.....
Hello,I want to scan over 500.000 color and B/w images from the 19th and 20th century; I represent a leading photo archive. My color expert suggested to save them in sRGB format since I currently do most of my business over the web and with many pre press and graphic people.
Whoever I know that sRGB clip off some color gamut; probably it will be wise to same my 2000x3000 pixel files as CIELAB, or Adobe RGB and than batch processing the tumbnails and medium resolution as sRGB..... What do you think???
Also, for this consideration, which color space do you believe in considered most appropriate for quality and max amount of color info? Adobe RGB, Apple RGB, etc, etc.....
Thanks; Andrea
--
------------------------------------------------
Fratelli Alinari Photo Archives and Museum
http://www.alinari.com
The world's oldest picture library
tel: +39-055-2395201
fax: +39-055-2382857
-----------------------------------------------
From: Dan Margulis,
Date: Wed, Jul 11, 2001, 1:46 PM
RE: [colortheory] More wide/rich color space for archival needs.....
Andrea writes,>>I want to scan over 500.000 color and B/w images from the 19th and 20th century; I represent a leading photo archive. My color expert suggested to save them in sRGB format since I currently do most of my business over the web and with many pre press and graphic people.>>
This might be a good suggestion if you were selling shoes or financial services. However, your clientele is more sophisticated graphically, therefore less likely to be using sRGB.
>>Whoever I know that sRGB clip off some color gamut; probably it will be wise to same my 2000x3000 pixel files as CIELAB, or Adobe RGB and than batch processing the tumbnails and medium resolution as sRGB..... What do you think???>>
This is not a significant issue. It is true that sRGB has no means of defining very pure cyans. By this, I mean that almost any color where the CMYK values are >90c<10m<10y doesn't exist in sRGB. This sounds much worse than it is. That particular color is virtually unknown in nature or in artwork. You could probably review your entire collection and maybe find two pictures that have this color in them.
>>Also, for this consideration, which color space do you believe in considered most appropriate for quality and max amount of color info? Adobe RGB, Apple RGB, etc, etc.....>>
For this particular application, stressing that it might not be true in other contexts, Adobe RGB is a poor choice. When this type of artwork has intense colors, they tend to be rich and deep rather than brilliant and saturated. Also, retention of detail in colored areas is of prime importance. I would use either Apple RGB or ColorMatch RGB. sRGB is reasonable but I would not use it on the grounds that few of your clients use it and it would open up the possibility of misunderstanding with those who do not understand color management.
Dan Margulis
From: Andrea de Polo
Date: Wed, Jul 11, 2001, 3:04 PM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Dan,Other people suggesting, interesting enough, actually Adobe RGB, so it is interesting that you suggest something else. But again this is probably out of a personal decision, correct?
Do you have any warning is I save for max color gamut as Ekta and than convert for the online and pre press usage into sRGB or Adobe/Apple RGB? I heard that Ekta has most info but it is NOT good for monitor display.....
Andrea
>For this particular application, stressing that it might not be true in
>other contexts, Adobe RGB is a poor choice. When this type of artwork has
>intense colors, they tend to be rich and deep rather than brilliant and
>saturated. Also, retention of detail in colored areas is of prime
>importance. I would use either Apple RGB or ColorMatch RGB. sRGB is
>reasonable but I would not use it on the grounds that few of your clients
>use it and it would open up the possibility of misunderstanding with those
>who do not understand color management.--
------------------------------------------------
Fratelli Alinari Photo Archives and Museum
http://www.alinari.com
The world's oldest picture library
tel: +39-055-2395201
fax: +39-055-2382857
------------------------------------------------
From: Dan Margulis,
Date: Thu, Jul 12, 2001, 10:48 AM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Andrea writes,>>Other people suggesting, interesting enough, actually Adobe RGB, so it is interesting that you suggest something else. But again this is probably out of a personal decision, correct?>>
No, it is because I understand that the character of the images that your company works with is very different than, say, those of a general-practice professional photographer. For the specific type of image you work with, a narrower-gamut RGB is advantageous.
In any case, if you are distributing these images to the general public, Adobe RGB is a horrible choice. Many if not most users wouldn't know how to cope with it and would open it into their own RGB, where it will appear washed out and colorless. The other three choices are enough alike that this objection doesn't apply.
>>Do you have any warning is I save for max color gamut as Ekta and than convert for the online and pre press usage into sRGB or Adobe/Apple RGB? I heard that Ekta has most info but it is NOT good for monitor display.....>>
The idea that any RGB definition would have "more information" than any other is a calibrationist fantasy. Adobe RGB has exactly the same amount of information as sRGB or Apple RGB or ColorMatch RGB. Each, however, contains certain information that one or more of the others lack. Therefore, each one will have certain mild advantages in some cases and certain disadvantages in others.
As a rule, I don't recommend that anyone use RGBs other than the four that Adobe suggests. Even if that were not a consideration, the space you inquire about would be worse even than Adobe RGB for this particular application.
Dan Margulis
From: Andrea de Polo, andrea@alinari.it
Date: Thu, Jul 12, 2001, 11:26 AM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Thanks again Dan for your comment.After reading many emails I think that I could store my files as Adobe RGB and than send out mostly as sRGB; for specific requirements I will convert the Adobe RGB to other color profiles, if needed.
Do you see anything wrong with this train of thought?
Thanks so much for your help :-) Andrea
>The idea that any RGB definition would have "more information" than any
>other is a calibrationist fantasy. Adobe RGB has exactly the same amount of
>information as sRGB or Apple RGB or ColorMatch RGB. Each, however, contains
>certain information that one or more of the others lack. Therefore, each
>one will have certain mild advantages in some cases and certain
>disadvantages in others.>As a rule, I don't recommend that anyone use RGBs other than the four that
>Adobe suggests. Even if that were not a consideration, the space you
>inquire about would be worse even than Adobe RGB for this particular
>application.
From: "Bruce J. Lindbloom", blind@picto.com
Date: Thu, Jul 12, 2001, 11:56 AM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Dan,I'm confused by your earlier message. Here are two consecutive sentences that appeared in that message:
> Adobe RGB has exactly the same amount of information as sRGB or Apple RGB or
> ColorMatch RGB.> Each, however, contains certain information that one or more of the others
> lack.These two statements seem to be in direct conflict with each other. How can the reference spaces contain "exactly the same amount of information" and also each "contains certain information that one or more of the others lack"?
--
Bruce J. Lindbloom, Pictographics Intl. Corp.
From: Jeff Schewe, schewe@schewephoto.com
Date: Thu, Jul 12, 2001, 12:00 PM
RE: [colortheory] Re: More wide/rich color space for archival needs.....
On 7/11/01 5:42 PM, Dan wrote:> For this particular application, stressing that it might not be true in
> other contexts, Adobe RGB is a poor choice. When this type of artwork has
> intense colors, they tend to be rich and deep rather than brilliant and
> saturated. Also, retention of detail in colored areas is of prime
> importance. I would use either Apple RGB or ColorMatch RGB. sRGB is
> reasonable but I would not use it on the grounds that few of your clients
> use it and it would open up the possibility of misunderstanding with those
> who do not understand color management.I would humbly disagree. Apple RGB would be about as bad a color space as sRGB. The main difference is that sRGB has a gamma of 2.2 and Apple RGB is 1.8, but otherwise both spaces have almost equally poor color gamut.
Also, stay away from Lab unless you're archiving in 16 bit color space.
The best choice, from my point of view-assuming you are scanning primarily E-6 chromes-would indeed be Adobe RGB. It will contain most all the usable E-6 film gamut (a bit less than Ekta RGB) while still offering the maximum ability to multipurpose the files. Ideally, I would suggest scanning into Adobe RGB in 16 bit color for maximum quality and tone/color corrections and then archiving an "ideal 8 bit" version of the Adobe RGB file (after all corrections). From there, you can easily transform to either ColorMatch RGB or sRGB for RGB files or any flavor of CMYK you may need to use.
Regards,
Jeff Schewe
____________________________
Jeff Schewe
Schewe Photography
624 W. Willow St.
Chicago, IL 60614312-951-6334 ph
312-787-6814 fx
From: Jerry L. P'Simer
Date: Thu, Jul 12, 2001, 12:01 PM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Hi Andrea,>>I have a collection of victorian images (about 80) that were reproduced as stone lithographs in the late 19th and early 20th centuries. I have scanned all of them and then meticulously retouched the images that I wish to reproduce in print. It is my intent to reproduce them in six colors using various "Hi Fi" color schemes that I develop for every series of images that I reproduce. The completed images, ready for print, are archived in the DCS2 format using single file color composite with six channels each. All the rest are archived as colormatch RGB until I get around to retouching and preparing them for print. After I have prepared them for print I run a batch routine that converts the files for web use. The batch down sizes the images to 72 ppi with some minor sharpening and then generates a medium size and thumb nail size and then saves in the sRGB format using JPEG compression. The images that I have reproduced to date (9) can be seen at: http://www.victorianavenue.com>>
If it is your intent to reproduce the images I would suggest that you prepare the images for print first and then archive as tiff cmyk and then convert for web use. If it is not your intent to reproduce the images yet, I would suggest the colormatch RGB format or the Apple RGB format because you may in the future decide to reproduce the images in print. You can always create images for web use later as well.
Jerry P'Simer
From: Andrea de Polo, andrea@alinari.it
Date: Thu, Jul 12, 2001, 12:12 PM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Jerry,I will well follow your tips; many thanks! Adobe or Apple RGB, this is however another dilemma! Some people says Adobe is better used worldwide while Apple is more on the Mac OS platform..... What you think?
My partner is suggesting to save the files as RGB Jpeg and not CMYK TIFF for the following reason:
1. RGB has more color info and we can always convert to CMYK for example by apply on demand, the proper profile like Euroscale, SWOP, etc.
2. TIFF vs. Jpeg with least compression; our test with images scanned at 2000x3000 pixels, 304dpi show NO visible difference between the two formats BUT with Jpeg we can save over an archive of 3 million images, a lot of space!
Looking to hear, if you can, your last remarks;
Andrea>If it is your intent to reproduce the images I would suggest that you
>prepare the images for print first and then archive as tiff cmyk and then
>convert for web use. If it is not your intent to reproduce the images yet, I
>would suggest the colormatch RGB format or the apple RGB format because you
>may in the future decide to reproduce the images in print. You can always
>create images for web use later as well.--
------------------------------------------------
Fratelli Alinari Photo Archives and Museum
http://www.alinari.com
The world's oldest picture library
tel: +39-055-2395201
fax: +39-055-2382857
------------------------------------------------
From: Kevin Connery
Date: Thu, Jul 12, 2001, 2:52 PM
RE: Re:[colortheory]More wide/rich color space issues
Bruce writes,
>>These two statements seem to be in direct conflict with each other. How can
the reference spaces contain "exactly the same amount of information" and
also each "contains certain information that one or more of the others lack"?>>As I understand it, the difference between the statements is that each 8-bit RGB color space holds the same number of 'distinct' colors as possibilities ("amount of information"), but that what those exact colors are is different.
Think of the two sets of numbers {1,2,3} and {1,5,10}. Both have "exactly the same amount of information"--three thingies--but one of them can describe from 1-10 while the other only describes from 1-3
If you know all your work will fit into the smaller ranged space, you can be more precise when mapping to elements in that space. (That's complicated further by things where the choices might be akin to {1,3,5,7,9} vs {2,4,6,8,10}, where there are some end-point differences while the stuff in the middle is comparable. Many color-spaces fall into this group [in my opinion])
The alternatives are to use larger color spaces (16-bit, 24-bit, or whatever may show up later), or accept some form of trade-off.
If that's incorrect, PLEASE correct me!
--kdc
Kevin D. Connery Tanner Research, Inc. 626-792-3000
Quality Assurance 2650 E. Foothill Blvd. 626-432-5705/fax
Adfui, Feci Pasadena, CA 91107 kevin.connery@tanner.com
From: varis@varis.com, varis@varis.com
Date: Thu, Jul 12, 2001, 5:03 PM
RE: Re: [colortheory] Re: More wide/rich color space for archival needs.....
Hi all,I've been following this thread with some interest because of the wide variety of responses. This issue of deciding what file type to use for archiving has a new wrinkle - now we have to decide on what colorspace as well! I think the tendency has been to overcomplicate things needlessly. The choice of archive space depends a lot on the choices you make in your workflow. If you have no hope of re-scanning originals you will have to save a digital file in some form or another. Once this is decided there are a number of other questions to answer: Do you need to compress these files to save on available archive space? Do you have multiple uses for these files, now or in the future? Do you wish to commit to a particular rendering of these images? If you just can't or don't want to decide then you should probably just save the image file in whatever colorspace you scan in. This insures that you have the maximum amount of original data upon which you or someone else can perform various corrections, transforms or conversions, now or in the future. If you have an input profile for the scanner you can embed that in the file. If you do not use scanner profiles you can always save the files as untagged and decide what to do with the color later.
If you need to determine some "definite" color rendering then you would probably do well to use one of the standard Photoshop RGB workspaces and convert into that to save a color corrected version. Ah... but what space to use? Any of the standard spaces could work - your choice depends on whether you're an optimist, a pessimist or a pragmatist.
The optimist assumes that at least one possible future user of these archive images has the same reverence for the original color as he does and this user might have a new output method that can reproduce many if not all of the original colors of the image. This user would be technically sophisticated and have some method of validating the original colors in some meaningful way. The optimist chooses a colorspace that encompasses the complete range of colors in the original and his only nod in the direction of pragmatism is to choose a standard space - the most likely candidate here is Adobe 1998. Adobe 1998 is wide enough to encompass almost the full CMYK gamut and it stretches pretty far past most colors reproducible in CMYK. The only other standard space bigger would be Wide Gamut RGB and this space is ridiculously big and highly unmanageable in any real world application
The pessimist assumes that most future users of his archive would be unsophisticated. These users would simply open up the image on their crappy PC's VGA monitor and judge the color of the files and proceed to convert (or not) into whatever color output they need at the time. Most likely their most demanding use will be for some flavor of SWOP. The pessimist therefore optimizes his files for sRGB and saves them without tagging to avoid confusing as many users as possible. sRGB, it turns out, is not all that bad to convert from into most flavors of CMYK and any reasonably sophisticated user could deal with the color issues that the more limited sRGB gamut presents.
The pragmatist tries to balance the demands of both of these extremes. He must guess at the most likely end use for these archived files and try to optimize for that use without needlessly limiting the potential color range or increasing the probability of introducing some color compression artifacts or posterizing colors during the conversion process. He assumes that the user will probably judge these images on a reasonable monitor if they care about the color at all and most users will convert the files for whatever final use with some modicum of competence. The most likely choice here would be ColorMatch RGB - based on the ideally calibrated monitor, this space is not so large that it is likely to contain colors that are not reproducible by the majority of output methods. ColorMatch is closer in character to a well calibrated "Macintosh" monitor and is less likely to look wrong to the majority of service bureaus because they are used to working with Macs (whether or not they utilize a colorsync workflow) The color range defined by this space is the largest possible that can be seen on a monitor and it affords the best chance of success for someone who wants to judge the colors visually. It may not contain the full gamut of CMYK but the colors it misses are not frequently encountered in photographic images anyway and in most colors it exceeds the gamut of CMYK.
Me - personally, I'm an optimist. I archive my files in Adobe 1998 or sometimes a custom workspace . I create original artwork that I anticipate reproducing with fine art printmaking processes and I will most likely directly control their use. For all commercial uses I generally convert the files to the appropriate colorspace myself- not willing to trust some strangers ability to make critical color judgments on my behalf or correctly apply color management to insure the best interpretation. When I'm gone and my children inherit my image assets I assume that they will care enough about my work that the best possible reproductions will be specified and used. The rest I don't care about.
regards,
Lee Varis
varis@varis.com
www.varis.com
888-964-0024
From: "Jerry L. P'Simer"
Date: Thu, Jul 12, 2001, 6:21 PM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Andrea de Polo wrote:> >2. TIFF vs. Jpeg with least compression; our test with images scanned at 2000x3000 pixels, 304dpi show NO visible difference between the two formats BUT with Jpeg we can save over an archive of 3 million images, a lot of space!>>
True indeed. But, you should keep in mind that each time you open an image and resave in the JPEG format you throw out a little more image data each time that you save that you can not get back. If it is your intent to work on the images after scanning, and at a later date, than I would suggest that you still use the TIFF format with LZW compression until the file is completed. LZW compression will not compress as much as even the highest quality JPEG compression but you won't lose any data this way.
As for RGB flavors, this would depend on your work flow and the final use intended for the images. I personaly use the colormatch RGB format. I work in a mac based work flow and colormatch serves my purposes just fine.
Jerry P'Simer
From: David Riecks, david@riecks.com
Date: Thu, Jul 12, 2001, 11:33 PM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
At 06:08 PM 07/12/2001 +0200, you wrote:>1. RGB has more color info and we can always convert to CMYK for example by
apply on demand, the proper profile like Euroscale, SWOP, etc.>2. TIFF vs. Jpeg with least compression; our test with images scanned at
2000x3000 pixels, 304dpi show NO visible difference between the two formats
BUT with Jpeg we can save over an archive of 3 million images, a lot of space!Andrea:
My suggestions would be to save as RGB, colormatch RGB or Adobe RGB, as plain ol' TIFF (no LZW compression).
In the upcoming months you'll have the option to save into the new JPEG200 format (JP2) which will include a "lossless" compression format. In addition they will be wavelet encoded, which means that you can say goodbye to the "blockies." (what we will save hello to is not completely understood yet);-).
In the meantime, convert to sRGB, save a version as TIF, then as JPEG for web access. When the new JPEG comes out convert all the old TIF versions to the lossless version of JP2.
For more info on the new JPEG, see (http://www.jpeg.org/JPEG2000.htm).
David
David Riecks * david@riecks.com
701 W. Washington St * Midwest/Chicago ASMP
Champaign, IL 61820 *
ph/fax 217-239-FOTO(3686) * http://www.riecks.com/
From: Dan Margulis,
Date: Fri, Jul 13, 2001, 12:34 AM
RE: [colortheory] Re: More wide/rich color space for archival needs.....
Jeff writes,>>Also, stay away from Lab unless you're archiving in 16 bit color space....I would suggest scanning into Adobe RGB in 16 bit color for maximum quality and tone/color corrections and then...>>
Well, since you visit only occasionally, let me restate a challenge that is no doubt nauseatingly familiar to other list members.
For a long time now the conventional wisdom has been that one should work in 16-bit or some other form of high-bit color. No doubt this is theoretically true, and I used to believe that is was true as a practical matter. However, some years ago I acquired a large number of 16-bit images from various vendors and beat them up with a series of ridiculously complex moves. I then compared them to the same images, converted to 8-bit prior to applying the same moves. The resulting files were different, mind you, but damn if I could find even the slightest reason to prefer one to the other, no matter how extreme the correction, no matter how exigent the output conditions.
Since this isn't the result I expected, I have for several years asked here and elsewhere for those who advocate these methods if they might not be able to provide me, say, two or three sample original high-bit images, with a record of what moves were applied to them, so that I could verify that there is a quality gain, however slight, in applying them to a 16-bit image as opposed to an 8-bit one.
Since I have been making this request, what I have received is a large number of histograms purporting to prove that 16-bit is better, a large number of assurances that "16-bit has worked better for me", and a large number of excuses for why the images in which it has worked better are either unavailable or under NDA. As time has gone on I have grown to suspect that perhaps the reason no one can supply such images is that there are no such images to supply.
Can you help me out? If you or anyone else can supply such example images and they really demonstrate the merit of working in 16-bit, I'll be glad to let this list know and if the differences are significant I'll print them in my column at a size large enough for people to see. If you or anyone else has such images, just let me know and I'll give you shipping instructions.
Two provisos: the files have to be photographs, not computer-generated art. Also, if the changes you propose to make are truly bizarre, I reserve the right to make the comparison by converting the files from 16-bit to 8-bit and then back to 16-bit before applying them.
Dan Margulis
From: Dan Margulis,
Date: Fri, Jul 13, 2001, 12:34 AM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Bruce writes,>>These two statements seem to be in direct conflict with each other. How can the reference spaces contain "exactly the same amount of information" and also each "contains certain information that one or more of the others lack"?>>
Theoretical answer: (see post by Kevin) any RGB variant comprises 16,777,216 possible specifications for color. Any variant will contain certain colors that cannot be exactly duplicated in certain other RGBs. Any variant will lack the ability to exactly duplicate certain colors that are achievable in certain other RGBs. A number of people blithely assume that wider-gamut RGBs are inherently better, refusing to believe that when they add color definitions they also have to subtract some.
Real-world answer: (see post by Lee Varis) variant definitions of RGB will behave differently in response to the typical kinds of image manipulation users may apply. The effect on overall quality in one's choice of RGB is small in the overall scheme of things but it isn't zero. A user whose workflow designates Apple RGB, for example, is going to get results of a slightly different character than one who has decided on Adobe RGB. Not necessarily better, not necessarily worse, just different. Which one is "better" depends on a) the skill level of the user; b) what the images will be used for; c) the character of the images themselves.
In Andrea's case we don't know much about a) or b), but we do know that her images are atypical and therefore c) comes into play. The type of images she handles tend to work better in narrow-gamut RGBs. Therefore, assuming that the only two storage spaces available in the world are sRGB and Adobe RGB, I have little doubt that she will get slightly better results with sRGB. However, that is not the same recommendation I would make to Jeff Schewe for his own work.
Dan Margulis
From: Dan Margulis,
Date: Fri, Jul 13, 2001, 10:27 AM
RE: Re: [colortheory] More wide/rich color space for archival needs.....
Andrea writes,>>After reading many emails I think that I could store my files as Adobe RGB and than send out mostly as sRGB; for specific requirements I will convert the Adobe RGB to other color profiles, if needed. Do you see anything wrong with this train of thought?>>
I have no serious problem with it other than that it creates unnecessary work and opens the possibility of conversion errors. If you have decided to send out sRGB, I'd store in sRGB also. For the specific type of images you work with, again stressing that it would not be so for the average user, you will probably get better quality results working in sRGB than Adobe RGB, if those are your only two choices.
Dan Margulis
From: Todd Flashner
Date: Fri, Jul 13, 2001, 11:26 AM
RE: Re: [colortheory] Re: More wide/rich color space for archival needs.....
> Can you help me out? If you or anyone else can supply such example images
> and they really demonstrate the merit of working in 16-bit, I'll be glad to
> let this list know and if the differences are significant I'll print them
> in my column at a size large enough for people to see. If you or anyone
> else has such images, just let me know and I'll give you shipping
> instructions.How severe such moves are you willing to stand behind?
I'm still something of a digital newbie, so my experience is limited, but this is what I've seen for myself so far. I expect it may change as I test more.
I scan with a Leafscan-45, which allows me to capture a 16-bit HDR RAW file (no adjustments made to white point, black point, or gamma). Such files are rather dark, tonally compressed, what have you. If I convert to 8-bit a *that* stage of the game, by the time I've made it into a decent looking image I've got severe posterization.
I noticed this early on in my efforts, and stopped proceeding that way pronto, so I don't have an archive of such images to share with you. In fact most of my scanning is in Grayscale, on BW negs, but the same holds true. But if you think that your argument would hold up even under those extreme conditions, I could send you portions of my HDR files for you to play with.
However, if I use the scanner software to yield a decent looking 8-bit file to begin with, (vs, starting with a 16-bit HDR file and manually adjusting wt pt, bk pt, gamma,) then making similar moves to the images, each in their respective bit modes, as you say, the results may be slightly different, but not objectionably so.
Seems to be a matter of degree we are talking about, no?
Todd Flashner
From: Broudy, David
Date: Fri, Jul 13, 2001, 12:15 PM
RE: RE: [colortheory] More wide/rich color space for archival needs.. ...
> In the upcoming months you'll have the option to save into
> the new JPEG200
> format (JP2) which will include a "lossless" compression
> format. In addition
> they will be wavelet encoded, which means that you can say
> goodbye to the
> "blockies." (what we will save hello to is not completely
> understood yet);-).whee, another file format!
one of Kodak's film-scanning apps has an option to save "JPEG Lossless", but since this app is relatively old and creaky I doubt it is actually using JPEG2k. nobody I've asked at Kodak seems to know what it is, but Photoshop can open these so I have to assume it's just some application of JPEG that's less-lossy than... something.
is this option you write of a Pshop plug-in?
From: Aaron Kiley
Date: Sat, Jul 21, 2001, 9:17 AM
RE: [colortheory] sRGB gaining momentum
I've noticed the new Nikon D1X, and some graphics programs like Qimage are setting their defaults to sRGB. Should I stick with Adobe RGB for press?
Also, my client got a new printer. This printer is requesting that I deliver RGB instead of CMYK. What are the pros and cons of this.
... Aaron
From: Dan Margulis,
Date: Sat, Jul 21, 2001, 10:31 PM
RE: [colortheory] sRGB gaining momentum
Aaron writes,>>I've noticed the new Nikon D1X, and some graphics programs like Qimage are setting their defaults to sRGB. Should I stick with Adobe RGB for press?>>
Assuming that press is the only destination for these images, then it depends on your color correction skills. If you aren't particularly experienced and just want "pleasing color", you're likely to get it more easily with Adobe RGB. If you *are* proficient with curves, Adobe RGB is a poor choice for CMYK destinations.
>>Also, my client got a new printer. This printer is requesting that I deliver RGB instead of CMYK. What are the pros and cons of this.>>
If the printer eventually needs a CMYK file, then this request means one of three things.
1) The printer thinks that they have the expertise to make a really good sep out of an RGB file; 2) The printer has been burned too many times by photographers who didn't know what they were doing, concludes that all photographers are incompetent, and asks for the files in RGB in the hope that the printed result will be merely mediocre instead of horrible; 3) The printer doesn't like photographers horning in on their business and wants to keep them in their place, namely RGB.
So, the pro is, if you think the printer is apt to be much more skilled than you, he should get a better result. If you are more or less equal skill, then it's a con. There is no perfect RGB>CMYK; there will always be sacrifices and compromises. You can't blame the printer if he decides to make different sacrifices and compromises than you would yourself.
There is one other con, which is easily avoided: AFAIK you could count on the fingers of one hand all the printers in the U.S. who know anything about RGB workspaces. If you are using Adobe RGB the chances are extremely large that the printer will open it up in some other RGB and it will look flat and colorless. Unless you are positive that the printer knows the right way to do this, send him the file in LAB, not RGB.
Dan Margulis
From: pfigen@keyway.net, pfigen@keyway.net
Date: Sat, Jul 21, 2001, 10:39 PM
RE: Re: [colortheory] sRGB gaining momentum
> If you aren't particularly
> experienced and just want "pleasing color", you're likely to get it more
> easily with Adobe RGB. If you *are* proficient with curves, Adobe RGB is a
> poor choice for CMYK destinations.Why, if you're adept at Curve base corrections, would the choice of Adobe, be such a detriment compared to sRGB or ColorMatch? I agree that someone who isn't very good might have better results in Adobe RGB, but how exactly does negatively affect the proficient?
Peter Figen
Peter Figen Photography
Los Angeles, Ca.
From: Tara Keezer, taralynnk@earthlink.net
Date: Sat, Jul 21, 2001, 11:30 PM
RE: [colortheory] Re: sRGB Gaining Momentum
On Saturday, July 21, 2001, at 11:11 AM, Aaron Kiley wrote:> Also, my client got a new printer. This printer is requesting that I
> deliver RGB instead of CMYK. What are the pros and cons of this.If the client is insisting you do exactly what the printer requires, add a waiver to the client's proof approval, indicating that you will not accept responsibility for final color output, since it was taken out of your hands.
Personally, I trust myself over an unknown tech and technology. I'm not all that eager to add to a service bureau's bottom line just because they want to handle the color conversion or they want to pay off the fancy new RIP they just installed.
Cynically yours,
Tara Keezer
From: Stan Schwartz, snsok@swbell.net
Date: Sat, Jul 21, 2001, 11:31 PM
RE: RE: [colortheory] sRGB gaining momentum
Dan and others, what about the various RGB color spaces for preparing images for the various printing machines such as the Fuji Pictrograph that require RGB input?The first image I had printed by one particular service bureau, who requested it in RGB, looked dreadful. They told me afterwards that the sRGB color space just wasn't big enough to produce great results on the Pictrograph. They suggested Colormatch RGB. Another service bureau offering Pictrograph prints that I tried suggested the Adobe RGB.
I have sent off images for printing done in both those color spaces. I am
not clear on how to decide which to use.Stan
From: Chris Murphy,
Date: Sun, Jul 22, 2001, 3:32 AM
RE: RE: [colortheory] sRGB gaining momentum
\>The first image I had printed by one particular service bureau, who
>requested it in RGB, looked dreadful. They told me afterwards that the sRGB
>color space just wasn't big enough to produce great results on the
>Pictrograph. They suggested Colormatch RGB. Another service bureau offering
>Pictrograph prints that I tried suggested the Adobe RGB.The Fuji Pictrography has it's own RGB space. Each Pictrography has a slightly different kind of behavior as well, probably because of variation in the auto-scanning densitometers connected to them use to calibrate them. While the device does a good job making sure a specific Pictrography has consistent output, it does not ensure consistent output among all Pictrographies.
Ideally, the service bureau would accept either ColorMatch or Adobe RGB images, and when they receive them, convert them to "Joe's Lab Pictrography RGB" space using a suitably made profile. That way you, as a customer, don't have to deal with the peculiarities of their Pictrography. I've profiled a number of Pictrographies and this works very well. A couple of companies I've worked with will give a ~20% discount to people who have a calibrated/profiled monitor, work in Adobe RGB, and will convert their own images to the RGB space needed by the Pictrography using an ICC profile supplied by the lab doing the prints.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Dan Margulis
Date: Sun, Jul 22, 2001, 11:18 AM
RE: Re: [colortheory] sRGB gaining momentum
Peter writes,>>Why, if you're adept at Curve base corrections, would the choice of Adobe, be such a detriment compared to sRGB or ColorMatch? I agree that someone who isn't very good might have better results in Adobe RGB, but how exactly does negatively affect the proficient?>>
Adobe RGB devotes considerably more of its space to relatively bright colors. Also, far more of its colors are out of the CMYK gamut altogether. In principle an expert should be able to achieve more or less the same results working in any RGB but in practice images corrected in Adobe RGB will usually wind up marginally more colorful.
If we use Adobe RGB in preference to, say, sRGB, for preliminary color moves, we from time to time will get important brightly colored areas that don't have enough detail. The flip side is that if we use sRGB we from time to time will get files that seem to us colorless.
Pepping up the dominant colors of an image in CMYK is pretty easy. Inserting detail in bright areas, OTOH, normally calls for channel blending into the weak color, which is much harder.
Don't get me wrong, it's not a huge deal, but for CMYK destinations I think that intermediate to expert users will get somewhat better results with any of the other three standard RGBs than with Adobe RGB.
Dan Margulis
From: Dan Margulis,
Date: Sun, Jul 22, 2001, 12:12 PM
RE: RE: [colortheory] sRGB gaining momentum
Stan writes,>>The first image I had printed by one particular service bureau, who requested it in RGB, looked dreadful. They told me afterwards that the sRGB color space just wasn't big enough to produce great results on the Pictrograph. They suggested Colormatch RGB. Another service bureau offering Pictrograph prints that I tried suggested the Adobe RGB.>>
The choice of RGB space has an impact on the final result, but as long as nobody screws up, it isn't a big one. If you were saying, my images are basically OK but I want to take the next step, then maybe we should talk about changing your RGB definition. If you say that they're "dreadful", that isn't the fault of your RGB. Either you are doing something seriously wrong or they are, or both.
A useful first step would be to convert one of your own files to LAB and give it to them in that form. If it still looks terrible, that probably means that the fault is at your end and you need to look at your entire setup. If it looks considerably better than when you had given it to them in RGB, further investigation will be necessary to find out what went wrong at their end.
Dan Margulis
From: Aaron Kiley
Date: Sun, Jul 22, 2001, 2:58 PM
RE: [colortheory] Re: sRGB gaining momentum
Thanks for all the expert advice. I'm going to try again to convince them to let me send CMYK files. That failing, I'll watch them take a stab at my corrected imbedded sRGB files and if they fail, I'll have lots of ammunition for getting my way next time. (also time to do a proper test or their press)What if I work in CMYK then convert back to RGB. Is there a potential loss of quality when my RGB files get a final conversion to CMYK by the press people?
One more question. I have a Nikon D1X camera. I've heard it's best to set the camera RGB type to Adobe. Correct me if I'm wrong, but this seems a good idea for collecting the largest gamut.
... Aaron
From: Stan Schwartz, snsok@swbell.net
Date: Sun, Jul 22, 2001, 6:00 PM
RE: RE: [colortheory] sRGB gaining momentum
That is very interesting. Thank you for the explanation.With the Adobe RGB color space, the Pictrograph from one vendor looked
pretty good in terms of color saturation but not color accuracy.I went to the Fuji website in the United Kingdom, and downloaded a Fuji
Pictrograph output profile. I used that in Photoshop 6 for "soft proofing."
I did the color adjustment so that the soft proof looked accurate and then
asked them to printed straight rather than do a profile conversion. By
doing that, I did get better color accuracy. I guess I must have been lucky.I asked this vendor they could supply a profile--they said no.
Stan
From: Stan Schwartz, snsok@swbell.net
To: Dan Margulis, 76270,1033
Date: Sun, Jul 22, 2001, 6:08 PM
RE: RE: [colortheory] sRGB gaining momentum
Dan,I understand that the input for this particular machine has to be RGB.
Stan
From: Chris Murphy, lists@colorremedies.com
Date: Sun, Jul 22, 2001, 6:58 PM
RE: RE: [colortheory] sRGB gaining momentum
>I went to the Fuji website in the United Kingdom, and downloaded a Fuji
>Pictrograph output profile. I used that in Photoshop 6 for "soft proofing."
>I did the color adjustment so that the soft proof looked accurate and then
>asked them to printed straight rather than do a profile conversion. By
>doing that, I did get better color accuracy. I guess I must have been lucky.No, it's probably describing Pictrography behavior better than ColorMatch RGB or Adobe RGB is all. It's not dead on since it isn't a custom profile for the specific Pictro you're printing to, but it's not way off either.
>I asked this vendor they could supply a profile--they said no.
Keep looking around, they're out there, especially with Lambdas, Lightjets, and Pictrographies. While I don't expect you to argue with your vendor, my argument would be as follows (from a consumer point of view):
"MY responsibility as a customer is to make sure that I prepare my files correctly for output. I should have a calibrated and profiled display, and use a suitable standardized RGB space for editing my images in, such as ColorMatch RGB or Adobe RGB. I should also embed that profile into my image (an RGB TIFF in this case) so that the vendor who prints it will know whether I used ColorMatch RGB or Adobe RGB. I also recognize that there will be shifts in saturation from monitor to print, but I should expect the hue to be correct. Tonal range will be compressed, but I can live with that as well."
Those are customer responsibilities. Note how you are *NOT* responsible for the behavior of the vendor's printer. What your vendor is asking of you is for YOU to be responsible for the peculiarities of their Pictrography. Their responsibility should be to do one of two things: a.) Accept an image with a profile embedded in it (such as ColorMatch or Adobe RGB) and convert it using their Pictrography profile. b.) Give you their Pictrography profile and let you convert it before submitting it for printing.
If they are unwilling to do this, and you feel like you're paying for iterations you should have to, then find another vendor. While it works for some people, I don't usually recommend it - you could make a profile for their Pictro yourself. The reason I don't recommend this is because it makes you responsible for their printer behavior instead of them doing their job. What it sounds like is they are perfectly happy with you getting results you aren't completely pleased with, then having to color correct the image manually until it prints correctly, while having to pay for each test print.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Stan Schwartz
Date: Sun, Jul 22, 2001, 11:20 PM
RE: RE: [colortheory] sRGB gaining momentum
Those are good suggestions.I do like the Pictrograph print as a good balance of quality and cost. If you or anyone else knows of a good source for Pictrograph prints who would work in the way you describe, please let me know on or off list.
Thanks,
Stan
From: Chris Murphy
Date: Mon, Jul 23, 2001, 12:19 AM
RE: RE: [colortheory] sRGB gaining momentum
>I do like the Pictrograph print as a good balance of quality and cost. If
>you or anyone else knows of a good source for Pictrograph prints who would
>work in the way you describe, please let me know on or off list.In my local area, Photo Craft Laboratories will do a discount (wholesale pricing I think) for their Lightjet output. Mike's Camera has a Pictrography and while they don't have explicit discounts you could probably tell them you're using a calibrated/profiled monitor and can provide them images in either Adobe RGB or ready to print using their profile. If nothing else you'll be much closer on the first print than a conventional workflow with these kinds of devices. Both are in Boulder, CO.
I'm sure Andrew has a few in his area, possibly even nationally that are using profiles for Lightjet/Lambda/Pictro output.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Dan Margulis,
Date: Mon, Jul 23, 2001, 1:03 AM
RE: [colortheory] sRGB gaining momentum
Stan writes,>>I understand that the input for this particular machine has to be RGB.>>
It does. By supplying the vendor with an LAB file you force *them* to convert it into RGB. This eliminates the possibility that they are misinterpreting the flavor of RGB that you are supplying them, which could account for your bad results. They can't misinterpret an LAB file.
Dan Margulis
From: Andrew Rodney
Date: Mon, Jul 23, 2001, 9:33 AM
RE: Re: [colortheory] sRGB gaining momentum
on 7/22/01 10:18 PM, Chris Murphy wrote:> I'm sure Andrew has a few in his area, possibly even nationally that are
> using profiles for Lightjet/Lambda/Pictro output.Two that do very nice work and supply profiles:
Nancy Scans in New York
Pictopia, a web based lab with a very new Lightjet (www.pictopia.com).
Pictopia charges $10 a square foot!Andrew Rodney
Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.