Untitled Document

For the next Professional Photoshop, I am planning an experiment. I am going to make a CD of some good RGB images that I believe are challenging to print. Most will be intended to be sRGB but untagged, some will have sRGB or other profiles embedded.

I will pretend that I am just getting into digital photography and am experimenting with Photoshop for the first time. I will take the CD to a dozen different drug stores and mom-and-pop camera shops that advertise instant processing, and see what happens. Then, a week later, I'll send the same CDs to the same shops and see if there's significantly different results.

Unfortunately, I've fallen seriously behind, and I will be on the road and unable to run these tests before I actually draft the affected chapter. So, I'm going to assume that I know the answer but be ready to throw the whole thing out if it turns out that the tests show something different. But before taking that gamble, I'll ask if anybody on the list has tried something similar or might have a different opinion than the one expressed below.

What I *expect* this will show, based on a limited sample plus some word of mouth from amateurs, is:

*There is considerable variation from shop to shop--more than there would be in the CMYK world between one commercial printer and another.

*The day-to-day variation in a given consumer-oriented shop seems to be fairly reasonable, indicating that the machinery self-calibrates well, because it sure doesn't look like anybody's assigned to maintain it.

*Some really weird output has happened to me with blue skies.

*Nothing that I have seen pays attention to any embedded profiles. It seems like everything is assumed to be something like sRGB.

*The chances of getting any sort of sensible advice from a shop employee are slim.

Does anybody think that the actual tests will show something else?

Dan Margulis

___________________________________________________________________________

Your assumptions are largely correct and I can tell you exactly what you will find. It all depends on where you go, what kind of equipment is there and who is manning the controls. Of course, this was always the case with "machine" printing, even with film, but it's worse now. Frankly, how much can you expect when the typical order is selling for literally pennies a print. Price is generally the only consideration when people shop for consumer grade prints, so expect very little in the way of quality control and intelligence. The industry has been successfully dumbed down to the level of Wal-Mart and Costco.

If the order is for more expensive enlargements in a better shop, it should depend on your instruction to the order taker. First, you should be asked to give the instruction to either "print without correction" or to "make the best looking print". In the later case, results will obviously vary, even with the same operator, from day to day. The former case sounds easy, but it's actually harder and it depends on the shop's capabilities.

Although the machines are quite good at maintaining a linearized gray scale, none of these printers recognize profiles. Also, they're not set up to reproduce sRGB or any other profile accurately, so it doesn't matter what you embed. In fact, the typical minilab software will alter the image all on its own before the operator even gets it on the screen. Even if he thinks he's printing without correction, he can't. Two identical machines on two versions of software may yield different results from the same file, even though they're both calibrated and working properly. Better labs have the ability to turn such "intelligence" off, but most shops never do because it isn't profitable. Most can't turn it off.

In a better shop with better equipment, your files will be opened in Photoshop and converted to a profile that was custom made for the printer, resized to 400ppi at the final size and this new file then sent to the printer in a "repro" mode where the printer's software is entirely locked out of the equation. You should get a call on the un-profiled images to verify their origin. In this scenario, you will get exactly the same image every time, month after month, and it should be nearly as reliable and accurate as a contract proof.

Most operators know nothing about any of this and some don't even have the capability let alone the knowledge or desire to work this way. Unfortunately, most customers don't know the difference either, nor do they care, because in this business today, price is the only consideration.

john castronovo

___________________________________________________________________________

Dan,

As one of those "drug store" chains with a photolab in every store, we have implemented ICC based color management across the board. Whatever flavour of RGB our customers decide to use will be recognized and converted to sRGB. We're not using custom profiles for the printers, although that could be implemented.

Control strips are run every morning to help ensure consistency from day to day. The most common cause for inconsistent results from printing the same file on different days would be the operator. If one operator figured a particular correction would better enhance an image compared to another, then you may receive two different prints. Ask them to print with "no adjustments" and this does not occur.

I received a brochure from Costco about a year ago targeted to pro photographers. It had recommendations about RGB working spaces and pointed their clients to "drycreekphoto.com" for custom printer profiles. I've printed the same file on different occasions and received slightly different results. Nothing huge, about on par with the variation we see on press with our flyers.

IMHO, yes the variation from shop to shop is huge depending on the type of stores you include. sRGB is pretty much the assumed standard for the Noritsu and Fuji printers, but it depends on the front end driving the printer and some labs are recognizing embedded profiles. The level of staff knowledge also varies widely with most in the "very limited" camp.

I haven't personally experienced any funky blue skies, but the metamerism (or likely more accurately, it is color constancy) on the Fuji Crystal Archive paper is awful. Makes the Ultrachrome inks look great.

Darren Bernaerdt

London Drugs Ltd.

(although the opinions expressed above are entirely my own and not the

official corporate line)

___________________________________________________________________________

On 3/28/06, Dan Margulis wrote:

That's an interesting parallel. At first this sounded like one of those things you see on the news where they send the same car to different mechanics to find out who's cheating customers – but something tells me you've got a bigger axe to grind.

Your expectations seem to be right on the money.

My question: I wonder what would happen if you gave them LAB?

-----

J Walton

___________________________________________________________________________

Anecdotal evidence from using the same high-street shop a number of times:

- There is some variation in colour from time to time: from red cast (& oversaturated reds), to greenish grays

- I've always used sRGB, or removed the profile altogether. Prints appear quite highly saturated so there may be an expectation that files should be Adobe RGB. In any case, attempting to make an allowance for the daily variation, there doesn't seem to be any change between embedded sRGB profile and untagged images.

- I haven't had experience of pure intense blue skies yet. They doesn't exist here :( I don't think I've asked them to print out-of-sRGB-gamut images either.

- Neither have I spoken to the staff about the kind of correction they make. On the few occasions when I've tried, it looks like I've never been able to speak the same language. For instance their idea of an "uncorrected print" is when they let the machine do the corrections. They don't call them "corrections" though, just what the machine *has* to do.

- The pro lab I use is a bit different but request Adobe RGB (never tried another profile or an untagged image to see what they say)

It's just one data point. My strong suspicion is that like the wet labs of yore, there may be ghastly operators at work, or hidden gems, and you won't know until you try them all. They all play in the same price range so there isn't very much to tell them apart; possibly word of mouth from trusted persnickety finicky hair splitters ;)

Regards,

Francis Corvin

___________________________________________________________________________

Dan,

Frankly I don't understand the point of doing this altogether. What exactly are you trying to achieve? You're writing a book on PROFESSIONAL Photoshop. I have images that are difficult to print and they fall into two distinct categories: (1) images with seriously out of gamut colours that suffer from gamut compression unless I "tinker", and (2) images that have considerable detail in levels L = 3 to 20 that I would like to see better differentiated, and here again I have adapted or developed some ways to separate those tones, but there is probably still room for improvement on both accounts. The idea that Costco or Shoppers Drug Mart could even begin to handle such images is simply beyond the pale. What I hope to get from the next edition of your book is fresh insight on how I can deploy my battery of Photoshop, Epson 4800 and all the colour management science behind them, and in which I have invested so much time and money, to handle such images that have problems printing properly. Experiments with a variety of local drop-off processors won't do it for me. Maybe you are trying to prove some arcane theoretical proposition with this, but unless I'm missing something it will be largely irrelevant to my interests.

Cheers

Mark Segal

___________________________________________________________________________

On the other hand do this same test with a couple of images placed on a page prepped for offset and I'm there. With all the printers on this forum we should be able to see lots of variations and pitfalls.

Lee Clawson

2/\V/\7 Studio

___________________________________________________________________________

Hi Dan,

I think that what we are missing here is your explanation as to why you want to do this. You know what your target readership is so I am sure this exercise has to do with something applicable to the

professionals who buy your books.

All the best!

Paco Márquez

661 McKinley St.

MIramar

San Juan, PR 00907

787-721-8554

787-587-7384 Cel.

http://www.pacomarquez.com

___________________________________________________________________________

The greatest variation I have experienced is with the large chains, whose results are all over the place. Independent labs, such as those that offer professional film processing, have been accurate and consistent.

The machines are more stable than the workflow. You will find great variation whenever user adjustments are made, and reprints will seldom match when the workflow includes user adjustments.

From the large chains, I have gotten both over-saturated and weak sky blue.

Yes, sRGB is a good assumption. One confusion is whether the printing system makes conversions or not, and if so, by what method. I have found that ColorMatch RGB on the Fuji Frontier gets a "better", less cool neutral but is overall a bit flatter in contrast than sRGB.

The chances of a shop employee or shop owner getting straightforward information from the OEM is slim. They are quite tight bout their proprietary systems.

Sure, one thing to check would be if there is any scaling with regard to image size. Given a 4x6 image file @ 300ppi, I'll bet that the print you get back has subject matter that is a bit larger or smaller. This is totally no the case in offset.

I think that the test results might be insightful beyond what I have already experienced. Especially if you could get a knowledgeable rep from and OEM or two to get involved.

Henry Davis

___________________________________________________________________________

On 3/29/06, MARK SEGAL wrote:

Obviously Dan is the best one to answer t his question, and I'm sure at some point he will.

But if you boil the situation down to its basic components, there are plenty of similarities between handing files off to a drug store print process and handing files off to a random sheetfed printer. The process may differ, but you are basically giving tagged (or untagged) files to an unknown person with unknown process control. Just as you cannot choose to processes your photos, many of us do not get to choose who prints our CMYK files.

And yet we still need to provide a good finished product. At some point we need to protect ourselves from the truly inept while still giving a quality printer something to work with.

Unless he has some working space experiment working, I'd guess that's his angle on this.

-----

J Walton

___________________________________________________________________________

Henry, this is interesting, as both spaces are idealised editing spaces, R=G=B in both, so neutrals should not be affected...it sounds like calibration drift or perhaps some undocumented glitch/conversion etc. Is this repeatable at this shop, what happens if you use Apple RGB or Adobe RGB etc.

While on profiles being used or not, Adobe Acrobat and Adobe Reader use profiles and it would appear that Adobe Reader has hardwired profiles to assume untagged data (my guess is the assumption is sRGB, US Web Coated SWOP v2 and Gray 2.2 gamma). So one can get variation depending on source colour modes and the profiles assigned/assumed to these objects. We often get PDF file supply for print (but we are not a photo lab) and there can be lots of fun with 100K or 0CMY100K being treated different in composite output and other colour issues.

Regards,

Stephen Marsh.

___________________________________________________________________________

dear Dan,

I would consider the on line stores as well.

My guess is that you are increasing the size of your audience, just following photoshop sales.

Pursuing that idea a little further, I would think that a representative chapter on home printers: epson, canon and hp printers is also in this category.

on the personal front, the printer companies almost give away the printers with the expectation that they will make it up on ink sales, but since I purchase third party $1.60 (canon) or $2.00 (epson) ink cartridge replacements that appear to be spot on color-wise, water and fade proof, and cost pennies per print (I get my paper at Costco) with consistent enough color, I can't see spending 20 cents for a print, but would benefit from your informed understanding to manage the colors even better than I do now -which, at default settings, are most of the time good enough.

I do notice that the gamut of these printers appear to be larger than standard CMYK printers, especially the more recent models, which, I know, opens up a major challenge in what color space to use. I would guess that a lot of us could benefit from your wisdom in this area.

onno de Jong

___________________________________________________________________________

In a message dated 3/28/06 11:38 PM, Darren & Leanne Bernaerdt wrote:

Hi.

I don't understand: if your chain is implementing ICC-based color management "across the board," how can it be that there are no custom profiles for the printers? Shouldn't that be considered a crucial step?

Regards.

--------------

Marco Ugolini

Mill Valley, CA

___________________________________________________________________________

As someone who is responsible for the operation of a Fuji Frontier, I generally concur with jc castronovo. It is a good idea to keep in mind that digital minilabs are designed for the mass market retailer. While the Frontier represents more than half of the digital minilab installations, most machines operate in much the same way regardless of manufacturer.

The machines themselves are quite capable of very consistent work. To bear witness to this, it is now a common practice to print black and white images on color photographic paper and get remarkably neutral prints. Bear in mind that photo paper has no black plate, the image is constructed exclusively of CMY dyes. Digital minilabs can do this out of the box with no special adjustment. This task was challenging at best with analog equipment, and not really practical to do in production.

Of course, like any printer, the actions of the operator, or lack thereof, have an impact on the result. It has been standard practice within the photo industry for years to calibrate your printer and processor daily. In the case of the Frontier, the operating manual instructs this, and the software makes it simple and easy to do without any particular knowledge. It takes about 20 minutes to accomplish, with maybe 5 minutes of active operator involvement, tops. A couple of years ago we were asked to complete a survey for Fuji. One of the multiple-choice questions was "How often to you process a control strip?" The answer options were:

1. Weekly

2. Monthly

3. Never

4. Don't know

Daily was not offered as a possibility. From experience, the machines do not drift much, and while you can get by without monitoring them a lot, they are not immune to variability. If you put this into an environment where the operator has no understanding of color, and may in fact be prohibited by his management from making adjustments, you do increase the chances for variability no matter how good the equipment may be.

With respect to the files themselves, the Frontiers assume sRGB and convert from there to a proprietary internal working space. Fuji has not published any information about this, but conversations with engineers and empirical evidence leads me to believe that this is a consistent and predictable process, like doing the exact same profile-to-profile conversion every time in Photoshop. This is a reasonable approach in the intended market since the vast majority of consumer digital cameras do not embed a profile, and if they bother to specify a colorspace at all, it is invariably sRGB. In some versions of the Frontier software sold outside of the USA, it is possible to bypass this conversion, but the difference seems to be trivial. The USA consumer market versions don't offer this option. It is possible to profile the printer, and there are professional and third-party software packages that enable a fully color-managed workflow. The procedures are by no means standardized, so what you will find in actual lab sites will nearly always be unique. With film, you could put a roll on the counter and the lab would know what to do with it without anybody having to say anything. Such standardization has yet to evolve in digital photography. As a lab, we never know what to expect in a file.

You may also run into the effects of image enhancing routines. When printing files from the majority of the digital cameras in the field today, a straight, unenhanced print suffers in comparison to a print from film on the same system. The usual complaint is dark, flat, dull. Most minilab manufacturers are now using scene content analysis along with the EXIF data to do all kinds of local and global "enhancement". Kodak Perfect Touch and Fuji Image Intelligence are examples of this. These should be repeatable for a given image, but you can find different results between seemingly similar images. I am starting to see evidence of similar enhancement strategies within original files coming from newer digital cameras, which may cause some overenhancement since the printer software won't be anticipating this. The minilab manufacturers usually offer the ability to turn the enhancement on or off, but it is not unusal for mangement to lock the settings in order to avoid having a skilled operator around. If the settings are not locked, it is not unusal for them to be left at whatever the keeps the current operator out of trouble the most.

Some posts have questioned the relevance of this project. Speaking for my industry, many inkjet printer users fail to consider the costs of their time in buying paper and ink, and making the prints, and any prints that they may throw away because they aren't right. Some folks just aren't interested in dealing with printer hardware. Retail printing offers a good quality product for a fair price. Well, should offer, but that is up to the individual store. Many actually do a very good job. The size capabilities of retail minilab equipment is increasing, and for very short run printing offers an attractive and convenient value proposition for even the most advanced Photoshop user.

Randy Wright

Specialty Color Services

Santa Barbara, CA

___________________________________________________________________________

Randy,

I think you get it right. Minilabs, film and now digital do fine. And I agree with you, "Retail printing offers a good quality product for a fair price". I wish I could say the same about the non-retail offset printing industry.

My test would be to send the same file to a number of offset printers using the same press and paper then see what the results are.

Lee Clawson

2/\V/\7 Studio

___________________________________________________________________________

The Costco store in LIC (possibly the closest to Dan's playground) has a downloadable profile for their Noritsu.

I used it a couple of times and the prints came out fine (a bit darker than it looked on my monitor, and somewhat flatter) but I didn't expect Dugall quality, and for $3 a 12x18" satin Fuji Crystal Archive it's a big bargain (what's up with that paper? looks cheap : ).

I printed 2 images up on a sheet and cut them in half for 2 9x12" model headshots. (the models are happy, I'm not sad…)

A lot of wedding photographers use Costco or Sam's Club even if they don't admit it, and they are a big segment of Dan's books fans.

Laurentiu Todie

___________________________________________________________________________

Marco,

Let me clarify - in terms of input (IOW - receiving files from customers), we recognize embedded profiles. This applies to our workflow whether it comes in via our website or via media handed to our photo counter staff or if a customer uploads the images from their media via a kiosk in-store. From there, the embedded profile is maintained whether it is being directed to our RA4 based printers or our inkjet printers. With our wide format inkjet printers, ICC profiles are utilized for the output. With our RA4 printers, custom printer profiles are not currently in use by the staff in the lab. If the customer's files are not in sRGB, they are converted to that profile. Apologies if my earlier ambiguity made my post unclear.

In my personal opinion, if the majority of the amateur digital "point-and-shoot" camera users understood the concept of white balance, the color in their images would be far more accurate than anything ICC profiles currently bring to the table.

Darren Bernaerdt

___________________________________________________________________________

I find this discussion on this list, this subject interesting. This is the first time I have seen this about photo labs. I use the drugstore/mom and pop labs for family/emergency pictures (1-hour). Three or four, I tried. I always go with the unlucky expert attitude. One will always get some variation from lab to lab, even at different times. Wither they use custom profiles for the machine(s) or not. But not like the cmyk press variation. And their pretty good considering (price, turnaround) they are a drugstore/mom and pop type lab.

The problem I run into is not so much with blue skies, but with high-key (97L, 0a, 0b) backgrounds with people (fleshtones) in the foreground. And that tells you something right there. The color of the background will go to the negative "a" side, another lab will go to the positive "a" side. Enough that my mother-in-law can tell the difference. She has no photography or printing background. She did not look like a "silver fox "in the photos. Hey. Its ST. Paddys' Day.

What I have to do, what I'm told from the lab manager is when I bring my files in for printing, note, if sending them electronically or on the bag where you put the cd in, that they are "professional photos". Yes, I did not know when you shoot on a 97 L, 0a, 0b BG. you have to list it as "professional photos". I have to remind my mother-in-law about this as well, when she sends in her pictures with 97 L, 0a, 0b BG's. Or I should have had my forehead tattooed in 50 L, 127a, 127b with "professional photographer" across it (out of gamut for my head) when submitting the cd. Which I take is that some are going to have to be printed optically or I should say "with corrections", even though I did this in P.S.(checked neutral balance in several areas w/ eyedropper) and put "no corrections" when printed.

Also, they do not need an embedded profile, either way they assume sRGB. Which some labs will recommend sRGB as the workspace or convert to sRGB, if there in another profile. This information is given, even to consumers' who somewhat understand color management. I have sent the same photo, one in sRGB and one Argb and the Argb look different printed than sRGB. But not enough to get bent out of shape about it.

Also, they can not be in folders (when using P.S., anyway), as the software they use does not open folders. One thing I forgot. They have to be in jpeg format (compression size optional). And this is how I prepare them for these labs. Its not like I fight them on this stuff. As I know their set up for the average customer in mind. Now, their was a mom and pop type lab that I used, catered to more the professional level, as well for consumers', where they can print for other color spaces (either Argb or sRGB, must be tagged if other than sRGB), except L*A*B, because they want to know how to handle the conversion. They were not too hip and savvy on the L*A*B thing.

Files can be in either jpeg, or tiff (recommended, no lzw, IBM-pc format only, as one knows the Mac can read either). and have to say as for neutrals (other colors as well), as Tony the tiger would say: "They're Grrrreat !" This even goes for running second prints at different times.

I find that the cosumerlabs that use custom profiles are somewhat better for neutral balance. The skin is printed more neutral, more like Veri-color III was to film. Where another lab that did not use custom profiles, skintones printed more like the pro line films...pro 100, more vivid colors. Overall....These labs do a very good job of printing for the price their asking for the prints. Most have good customer service. Where, your not happy with the printed result, they will correct it. Even if you state "no corrections", the first time. Which is good for people, lets say, my mother-in-law, where she has a digital camera and does not use a software program for pre-flight and is not hip to ink-jets and/or don't want to mess with em'. Just pull the card from the camera, pick the pictures to be printed. They will even size them for you for 4x5,.5x7, 8x10 format. At no extra cost.

As one of the lab manager's told me....We're not competing or ever competed with the professional labs. Its the "press and print" (popular with the consumer market for good results and the price factor for em') ink-jets that we're competing against. And some labs, drugstore type, will accommodate printing for professionals, with a discount. With pretty good results.

John Opitz

___________________________________________________________________________

What people have to understand is that getting acceptable results by simply sending sRGB to one of these printers is simply an accident, and not by design. Such minilabs generally don't recognize profiles and they have no internal profile of their own. LUTS for gray scale calibration and gamma are not profiles. Owners of these machines who care enough to make their own profiles or have them made by a professional have to purchase additional software to do the conversion from the original file's source profile to the printer's profile. Some labs provide their custom output profiles to their customers and ask the customer to do the conversion before sending the fiels in to print. This will yield the most accurate colors, but it costs more than what most people are willing to pay, so the usual workflow is to let the operator or the machine correct the files on the fly. I've had professional photographers complain that we altered their files, and other pros complain that we didn't. You have to ask and be flexible.

john castronovo

___________________________________________________________________________

I think this is exactly why sRGB was designed. It gives everyone a "baseline" standard, and one that tries to be reasonably close for mass-market devices that are not calibrated.

So when Joe Snapshooter sends a JPEG straight from his point and shoot, or after some messing around on his laptop, he has a reasonable chance of getting an acceptable print.

But when I manipulate a file on my color-managed workstation, processing a file from a high-end DSLR in a managed workflow, the prints I get are hardly an accident. When I find a vendor who keeps his machines reasonably well calibrated to factory standards based on sRGB, I can expect to get prints that are more than just acceptable.

This has been my experience. I'm getting better and more consistent prints from Walmart and Costco than I ever did in the old analog days from mass-market photofinishing. I only wish that routine CMYK jobs from offset printing were so accurate and consistent.

Tom Judd

___________________________________________________________________________

On 3/30/06 Lee Clawson included the following suggestion in a post referring

to Dan's suggestion about sending files to retail outlets for digital file

printing:

Now that's a great idea! Maybe Dan would be interested in trying that one out for demonstration purposes. It would certainly give his readers a clear idea of just how much variation they can expect in a commercial litho job. It's one thing to read about the kinds of problems one is likely to encounter, and quite another to see examples of the actual output. Granted he has done this in a limited fashion in Professional Photoshop, but it would be helpful to see what happens with a variety of images--all printed from the same files. One bad result is interesting, but maybe the printer was just having a bad day, or equipment problems. Would the quality be consistent for several images printed by the same printer? That would give us a much better idea of how much variation we can expect from printer to printer.

Howard Smith

___________________________________________________________________________

I am inclined to think that the results of such a (print shop) test wouldn't surprise anyone on this list. A downside of testing numerous offset printers is the obvious one of expense. Testing photographic output at say, 29 cents a pop, is a relatively low cost experiment. Although with that said, I don't think those results will surprise anyone either.... I'd expect a "wide variety of variability".

Les De Moss

DigiGraphics LLC

___________________________________________________________________________

Tom,

I apologize for continuing to kick a dead horse (yes, I'll stop, but one last kick please...) this is my point too.

Why aren't we discussing a test like this for offset printing ???

Lee Clawson

2/\V/\7 Studio

___________________________________________________________________________

It would make sense to do so.

I've only been hanging around this list for a couple of years, but I would bet that it originally served only CMYK printing. Given the explosion of inkjet printers, I'm sure that there are now many times the number of people printing via inkjet than producing conventional print jobs. So it's quite reasonable to expand the list audience to include everybody who's interested in controlling color from computer screen to paper.

Tom Judd

___________________________________________________________________________

A list member sent me the following off-line. Because it contains comments bearing on topics touched on by John Castronovo and others, I pass it on.

Dan Margulis

*********************

To: Dan Margulis

Re: Has Anyone Tried This

Date: Thu, 30 Mar 2006 14:24:35 -0000

Hi Dan,

I would be glad to test your CD's for you. Sounds like a good idea.

I've been using the same color lab for twenty years plus. The

quality of the work is excellent. The lab does everything right.

Daily color checks; good quality control.

But even then you get variation because each individual doing the

printing sees different and has different levels of skill. Recently

a new clerk was printing too dark. There was a photo journalism

student working there that did super excellent printing.

Even with digital prints going from my MAC to the Noritsu I've had

fairly good results. But then they haven't been too willing to work

with me on the differences.

I dread the day this lab will close because it will be almost

impossible to find this quality. And the pressure on this lab from

Wal Mart, Walgreens, Target, etc. is taking it's toll.

The test Cd's are a good idea. If I'm trying a new lab I test first.

I get something printed and don't tell them I'm a professional. I

tested a digital print several weeks back (It has been published in

Kansas City magazine and Eventing magazine with excellent results)

to see how this lab handled a persons work. When I got the print

back there was an obvious green cast. And the 12 X 18 print was

rolled into a small tube. Lots of creases.

If you would like me to test the cds I would be glad.

___________________________________________________________________________

Mark Segal writes (and Lee Clawson and Paco similarly)

Thank you for informing me of the book's title, but I wonder how you can speculate at this point as to what use I may intend to put this information to. There is a *great* deal that professionals can learn from discussing the problems and practices of nonprofessionals. My writings often feature extremely poor originals, not characteristic of what professionals work with, because the same moves that make such images obviously better also make excellent images subtly better. I have written extensively on how Auto Levels and similar commands work, even though I strongly recommend against their use, because if you don't understand why they're effective on certain kinds of images and not others you will have trouble understanding when to make more sophisticated moves.

The idea that you SHOULD be able to take your images to Costco was forcefully advocated in the early to mid-1990s in the paper "Color Management Systems for Pushbutton Color", in which it was argued that "color is rapidly becoming a commodity" and that skill in operating such systems would be unnecessary in the world to come. I took a good deal of heat at the time for suggesting that color management systems do not work in the absence of effective process control, which I stated might or might not be found at Costco. This was a distinctly minority view then although ten years later it seems obvious.

That said, the statement quoted above represents the type of attitude that I would like to combat. It smacks of the same type of thinking as "the idea that you can color-correct in sRGB is beyond the pale", or "the idea that you can get good results without a calibrated monitor is beyond the pale" or "the idea that you can get predictable results without profiling your camera is beyond the pale." Or, earlier, "the idea that you can even begin to get film-like quality from a digital camera is beyond the pale", and even earlier, "the idea that you can even begin to get high-end quality on a Macintosh using Photoshop is simply beyond the pale."

It's *not* beyond the pale. Assuming similar quality paper and similar colorants, the quality of the input images is vastly more important than the difference in predictability between Costco's system and a better-managed one. A person skilled in color correction will have no problem producing better work at Costco than a less-skilled one with a better machine. The Costco work will look better almost all the time if the colorants are equal, and it will still look better most of the time even if the better system uses higher-quality colorants. Similarly, if the paper Costco uses is better than the stock in use on the more reliable system, the advantage of the better stock will outweigh the disadvantage in predictability more often than not.

If you want to know how to use all these weapons without the headaches so frequently associated with them, you need to stop thinking of specific products or some propellerhead's nostrums and start thinking of what you are trying to achieve practically. To my mind this means looking at every aspect of the workflow with a critical eye, without preconceptions. If you believe you should work with a calibrated monitor it should not be because you think that working with an uncalibrated monitor is beyond the pale. It should be because you have carefully considered the consequences of working with an uncalibrated monitor. What bad things do you expect will happen, specifically, and to what categories of image? What types of images will improve if you have better calibration? How do you know when your calibration is good enough? What types of problems can never be solved no matter how much time and effort you put into them?

If you have actually thought about these things and are comfortable with your answers, then you'll find the rest of process control easy, just as I do. That's the philosophy I'll attempt to teach in the appropriate chapters of the new book. If not, you shouldn't be surprised if somebody else shows up with better results from Costco than you do from your high-priced gear.

Dan Margulis

___________________________________________________________________________

In the 29 years I have operated a professional color lab, many discussions have taken place regarding the validity of blind tests. It seems straight forward enough; run something -film, proofs, an enlargement- and see what you get.

The reasoning goes like this: without first announcing who you are, what you do, or what you expect, the work received will be produced according to a given lab's 'baseline'... without any special handling, attention or concern for the preferences of a particular client. I think that line of thinking is mostly correct.

I believe the exercise *can* separate labs into broad categories of unacceptable or acceptable, for *overall* quality of the product, price, turnaround, packaging, etc.

Beyond that, experience tells me that neither a lab or a photographer operates independently of the other, and that true success depends on a close partnership between the two. This is true because of the very issues we discuss on this list concerning countless pitfalls in our perception of color, density, and contrast and how best to achieve output that matches up with our own subjective tastes and requirements.

"Good" professional labs, as judged by their clientele, manage to hit that mark either by chance (i.e.: a great blind test that happens to suit a particular client's tastes), or because they have taken a particular client's preferences into account and incorporated them into the production of their work.

This is not to say that top professional labs do not follow controls, procedures and standards in how they go about producing their work, they do. But subjectivity is an unavoidable fact of life in the color industry, and the best labs will put their baseline standards aside in favor of their client's... when they know what that is.

If we know that a client prefers 'very saturated, over-sharpened' images... that's what we strive to produce. If we don't know that, and provide him with images of more conservative proportions, he may reason that we're not a 'good' lab. And for him, that would be true. For both he and us, the lack of communication would be unfortunate.

I am not in any way arguing against running blind tests in order to choose a lab. But there is more to consider if you are looking for the best lab to suit your imaging needs.

I am suggesting that once an 'acceptable' lab is selected from such a test, obtaining the very best quality means taking the next step of forming a relationship with those who will be producing the work, because in the end it is nearly impossible to achieve 'quality' without first knowing how that is defined, and in our industry it is defined in a thousand different ways...

Les De Moss

DigiGrahics LLC

___________________________________________________________________________

Man! I've seen the same sep printed in different places in the same magazine or brochure where they don't match each other by a country mile.

john castronovo

___________________________________________________________________________

----- Original Message -----

From: "Marco Ugolini"

Of course there are custom profiles for these machines. We make our own and they profile very well, just so long as any front end processing is turned off. Many machines will correct for image biases if you don't take steps to prevent it. Manufacturers don't supply profiles and it would be of little use because they have no control over the choice of paper, chemistry and gray scale aims that each lab may independently employ.

john castronovo

___________________________________________________________________________

Dan, et al,

As a suggestion and in the spirit of being "fair and balanced", why not use the ISO 12640-2 XYZ/RGB (sRGB color space) natural images as the standard to be sent to all these labs? If you're familiar with these images, they are a good cross section of the types of images that one might experience. They are not designed to fail in any way but at the same time they do represent some challenges such as flesh tones, neutrals, saturated colors, one image that's predominately

"warm" that might fool a machine analyzer, etc.

If folks aren't familiar with this image set, I could upload lo-rez versions for viewing only (they're copyrighted). I just received my set the other day (along with the 12640-1 CMYK set). I started with the XYZ versions, converted to LAB and did a slight a/b color contrast enhancement and slight L* tone move (the images as supplied are exceedingly dead and flat). I'd be happy to provide mine as long as whoever is sending the CD to the LAB can show proof of ownership if the ISO cops come after us. :-)

Just a thought.

Terry Wyse

_____________________________

WyseConsul

Color Management Consulting

G7 Certified Expert

704.843.0858

http://www.wyseconsul.com

http://www.colormanagementgroup.com

___________________________________________________________________________

Thank you, Darren.

Of course, ICC profiles are not meant to create much more than a match between a verifiable expectation (for example, an image as it appears on a calibrated and profiled display) and a result (as viewed on an print made with a profiled inkjet printer, for example).

If the image has a cast, or a poor white balance, the print will have them too. ICC profiles are not meant to fix a poorly-executed image.

Regards.

--------------

Marco Ugolini

Mill Valley, CA

___________________________________________________________________________

Thanks, John. That's not surprising, but it's nice to receive this information from someone who has actually seen the variations. It makes me wonder how so many jobs are printed well--or maybe they're not and we just don't realize it because we didn't see the original images. With that in mind, maybe our real objective is to get prints produced that get the job done instead of trying to get prints that are close to 100% accurate. If a flower arrangement, a Playboy model, or a street scene looks good in print, should we really worry about how well the printed output matches the original? It's a nice goal, but maybe impractical to aim for accuracy instead of for acceptability. Yet another thing to consider is that viewers are not scientific in their evaluations of printed images, and probably don't really care all that much anyway. How many of us have spent hours or even minutes studying an advertisement for Campbell's soup or a glossy ad for a new Ferrari? We glance at the picture, maybe tear it out of a magazine and put it on a bulletin board, or just toss it once we've gone through the material.

Where the fine arts are concerned, it's another matter. But again it's really more a matter of producing work that our clients like rather than trying to get exact color matches. If we take a photo of some red roses, that red can vary considerably and probably will still be pleasing to most viewers who see it.

Just random thoughts.

Howard Smith

___________________________________________________________________________

Without a doubt Les. A blind test without communication is meaningless. What we need to do is to send the files along with complete instructions on how they're supposed to be printed.

There are three possible general instructions:

a) print so that it looks best according to the technician's skillful eye

b) print so that the embedded profile is honored and converted to the output profile of the printer for the best visual match

c) just drop it in the chute and see what comes out

All of these are valid workflows in certain circumstances and there are variations for each of them. The paper must also be standardized. Don't expect Fuji Crystal Archive luster and Kodak gloss to look the same.

What it all means is that the very thing you want to know by this test is impossible to learn without a good deal of communication when the order is placed. The irony is that the best labs will offer the most variability. Take away the communication and the best lab may appear like the worst.

john castronovo

___________________________________________________________________________

OK Dan, Fair Ball - challenging posts get challenging answers!

Of course it is true - in an objective and literal sense nothing is necessarily and for all time "beyond the pale" in this world - only certain things get there in one's mind by virtue of one's own experience and objectives, which needless to say are not necessarily eveyone else's experience and objectives - witness numerous contributions to this thread alone. Re your last paragraph, on the whole and to a considerable extent, "I have", and "Yes". But notwithstanding all that, for many purposes of photography "process control" is a combination of art, science and intent, hence considerably broader than what the corner mini-lab brings to the table even if all of them were perfectly synchronized. It is in that sense I truly wonder about the role of this topic in the context of "Professional Photoshop" as we know it - "ah, but therein perhaps lies the rub" - are we being treated to a new and different Professional Photoshop, not as we know it? Looking forward..............

Mark Segal

___________________________________________________________________________

From: Les De Moss

Subjectivity _should_ play a very small part, and then only when one is pushing the gamut limits of the printer.

You're suggesting that it's up to the lab to do part of the customer's editing for him. If I want very saturated, oversharpened images, it's my job to create those effects in Photoshop, before given them to the printer. What I think photographers want from the lab is the same thing they want from they're own printer: prints that look like the screen. I've been able to achieve that with my own printer--if I can't with a lab's printer, then there's something wrong with the lab.

Ciao, Paul D. DeRocco

___________________________________________________________________________

Hello Dan,

Because it may give you some additional insight from an avid consumer of your educational materials, further to last night's post I would like to explain a few things relating to your last two paragraphs which I've extracted below for ease of reference. I suspect I am probably one of many people with similar experience and objectives amongst your readership, so I thought it perhaps useful to you that dwell on this a bit.

(1) When I buy a specific product, it is only after careful research and consultation with people I trust who know it better than I do. Once I buy it, I need to think in terms of that product's capabilities and limitations. I'm not married to Epson - in fact if the new generation of wide-format Canon pigment printers turn out to be several steps ahead of my Epson 4800 in respects that matter to me I'll replace it. Turning to propellerheads - it is the same issue. Regardless of who writes what, I evaluate the content based on what else I know, find out elsewhere or for some tangible reason makes sense to me, not the writer.

(2) Thinking of what I am trying to achieve practically: of course. To work through only the "for instances" you raise below: I started working with a calibrated monitor after having the usual issue of mismatch between the soft-proof and the printer. So I read alot about colour management and all the "gurus" I consulted kept emphasizing the importance of a properly profiled and calibrated monitor. So I invested a few hundred bucks in ColorEyes Display and a Monaco Optix XR and wow - did that ever improve things. So yes, I had the "before and after" experience with monitor calibration and therefore I believe in it - not to say that by accident an uncalibrated monitor couldn't give fairly decent results, but that isn't the point. In my case I know monitor calibration helps with colour casts and luminosity rendition - the latter especially for images with alot of three-quartertone information I wish to reproduce. And I know whether my monitor calibration is good enough when I view the prints under D-50 Solux illumination compared with the soft-proof I see on the monitor - recognizing of course there are inherent differences between reflected and direct illumination that with current technology can't be completely arbitraged, and recognizing that artistic judgment intervenes about what looks right or most effective on paper. And speaking of awareness about limitations, I do know, for example, that with current technology there are gamut compression issues that cannot be solved to my total satisfaction, but that situation is a moving target and it gets better every year.

(3) So as you see, I have thought about many of these things - there's more than the foregoing to think about of course, and one needs to know what to think about before thinking about it, so to the extent I don't know everything there is to think about in this business I'm 99% likely to be missing some beats. But I'd rank myself some way up the learning curve and yet process control, while easy alot of the time, can still be challenging for certain kinds of images - and I know them when I see them. It is for handling problem images, improving dexterity with all imaging issues and sweating the last drop of quality out of images that are straightforward to start with that I turn to books such as your's, Eismann's, Kelby's, Willmore's, etc. So when I say that Costco is irrelevant to my objectives for reading a book like Professional Photoshop - it isn't that I'm snobby about commercial mini-labs - in fact I've seen myself they produce alot of work that is perfectly adequate for a great many people and purposes - that's fine, but so what? My underlying point about relevance is that like many people I'm using Photoshop as a personal experience in artistic and technical achievement that is augmented with learning materials like yours. Of course the subject of quality and consistency from mini-labs is most likely a topic of interest in its own right to people directly concerned with that issue. But unless the excursion into mini-lab syncopathy bears some kind of direct relevance to one's usual reasons for using Photoshop in photography, it just leaves me wondering why get into it. As I concluded yesterday, perhaps in the fulness of time you will make the light dawn......................

Best regards,

Mark Segal

___________________________________________________________________________

paul writes

hello paul, That is a rather ambitious comment... have you printed many files/scans from other absentee image providers. Profiles or no profiles, would you not consider that your setup is a rather 'closed loop' arrangement?

I believe that there are too many variables to expect any lab/pro or discount pharmacies to deliver consistent work unless that photographer spent the time necessary to match his/her images to the labs average output...no suprise there.

My guess would be that if you submitted a soft/flat 'bluish snow scene with evergreens barely exposed on the distant mountain side that four out of five different labs would create a similar 'better' image.. more tone range, nice clean neutrals, deeper blacks... basically a nice bw version with no green or blue haze. nice/better? but not what the customer wanted or expected.

jim bean

___________________________________________________________________________

Terry Wyse writes,

This thread is branching out into areas never contemplated. I am not interested in testing individual labs or printers. If I were, I would never recommend a single run of images--the test would have to be repeated at least three times on different days of the week and preferably different shifts. Any lab can have a bad day, and the fact that one batch comes back inadequate means very little.

If you want to test a lab's capabilities the way to do so is by sending them live work. If it comes back unacceptable then you throw it out and you haven't spent any more money than by sending them a test and throwing that out.

My initial question was more about what result a user might expect if he chooses one of these low-cost outlets at random. I note the suggestions from Lee Clawson and Howard that I should be doing the same sort of test with commercial printers. Everybody knows what the result of *that* test would be--I've shown lots of examples of real-world variation, including one graphic in Professional Photoshop of the same file reproduced by six different printers who all claimed to be SWOP-compliant, and another showing how much variation SWOP itself considers acceptable.

In the initial post I suggested that probably there's more variation between work done at Walgreen's vs. Costco than there is between commercial printer A and commercial printer B. The reason I asked for opinions is that I don't know for a fact that my suggestion is true, whereas I am quite confident I understand what sorts of variations can be expected if you send CMYK to an unknown commercial print shop. And, at least one list member has said that he disagrees with my suggestion.

The following items *do* seem to me clear, however,

*If you go to one of these vendors the chances of getting a visually pleasing, if somewhat unpredictable, result are pretty good.

*A lot of professionals in fact use these sources even if some of them are too ashamed to admit it.

*The Costcos of the world do not care if they match the output of the Walgreens or London Drugs of the world; however they do care that they can match their *own* output if you bring the job back for a repeat.

*The chances of finding somebody who knows what they're doing to help you at one of these vendors is pretty slim.

*Anybody who thinks that these low-cost vendors are not a serious threat to quality-oriented photo labs probably also does not believe that DVD is a serious threat to VHS or that consumer-level digital cameras can produce professional-level images.

Dan Margulis

___________________________________________________________________________

From: jimbean

No. I don't run a test print, then tweak the file according to how it looks wrong, and then print again. I calibrate and profile my monitor to 5000K, have halogen lighting adjusted to 5000K, and use profiles on my printer. (I built my own printer profiles, but found they were virtually identical to the stock Epson ones.) And my prints look as close to the monitor as I could possibly expect, given the intrinisic differences between the media. So I'm running completely open loop.

I should be able to get the same results from a lab if they give me an accurate printer profile that I can use to soft proof. And if I don't soft proof, I should still be able to get as good a result from a lab as I do with my own printer without soft proofing, which is fine for most narrow gamut images.

I have only one experience with an outside lab (getting an oversized print) and it was horrible, but that _was_ seven years ago. On the other hand, I know serious amateur photographers who routinely use Costco to do their printing, and are very happy with the results. I don't expect perfection, but these experiences lead me to believe that running open loop with a lab is feasible, and that when people get lousy results, it's a quality control issue, not artistic differences or intrinsic limits in the process.

Ciao, Paul D. DeRocco

___________________________________________________________________________

From Paul D DeRocco

In the traditional film-based, analog workflow that preceded digital imaging, this was *always* the case. An image could not be reproduced without lab technicians making decisions as to *how* the image *should* be reproduced. In the analog world those decisions were confined largely to color and density. Contrast was controllable only through paper selection, more involved film masking techniques, flashing and so on.

Of course, that is not what we are talking about here in the digital age, because the digital photographer has these correction and editing tools at hand on their own workstation. Evidently you know how to use those tools to achieve the results you are looking for.

In your situation then, achieving photographic output that best matches your screen does not require the lab to edit your images. It does however, require you and your lab to determine how your work will be produced, that is, on what equipment (Chromira, Lightjet, Pictro, Film Recorder) and what materials will be used (Kodak Endura, Fuji Crystal, Agfa... gloss, matte, lustre) the appropriate resolution of the file, and in what color space your images should be submitted. All of these decisions have an impact, for better or worse, on the perceived quality of the final output.

Within the context of this thread on testing labs, the simple and valid point I am making is that without communication, the results of printed output are a crap-shoot, because variables exist in the lab environment that are simply out of your control unless they are determined discussed, and selected in advance. To blindly test a lab without understanding and participating in these decisions would be to leave critical decisions on these matters to someone else.

I'll also add that many professional photographers have yet to fully make the transition (in their expectations) from analog to digital in terms of what they expect from their lab. Quite often images are submitted for a particular service as *print-ready*, meaning that the lab is to output the image as-is without intermediate editing or corrective work. If the output then fails to satisfy the client, too often we are asked why we didn't "take a look" or "make a phone call" or take some other action to prevent the image from actually being printed as-is.

In other words, the lab is held to a level of responsibility that is a hold-over from the days of analog printing... if it doesn't look right, the lab failed in their judgment. Nowadays, if the photographer is to create, edit and correct his own work (in essence, be his own lab technician), then the responsibility for the quality of the output he receives rests solely on his shoulders. They just can't have it both ways. Obviously this is true only for labs that practice consistent process control and calibration to known standards such as sRGB or A98.

Agreed. We all want to see on print what we see on screen. Just as we want a slide or transparency to print exactly the way it appears on a light table. Both are laudable goals, both are entirely impossible based on physics alone. Comparing an image that is rendered on different materials and illuminated under two vastly different methods (transmissive and reflective) and expecting them to provide a visual match is utter, uneducated nonsense.

But I am splitting hairs here. With the right equipment, materials and workflow, the results can and should be visually pleasing if not visually outstanding... as you have indicated in your statement regarding your own printing.

Is something wrong with a lab if they can't do what you say can do using your own equipment and m