You've left a few potentially important details out of your scenario, but I will give you my thoughts. Are your B&W images 4/C CMYK, grayscale, or RGB? It can make a difference.
When editing images for my own fine art printing, I always tag it with a profile. In fact, I do this whether my image is grayscale, CMYK or RGB. Of course, I am doing the printing myself, so I have full control over the project. I use a Canon 6100 and custom RGB profiles (which I build myself), so most of my B&W or toned monochrome prints are RGB. Without a source profile, Photoshop will have to assume or assign a profile in order to make the final conversion.
I have used B&W RIPs before, such as QTR, and they don't speak ICC. They simply take the raw umbers in the file and output them to the printer, passing them through specific inking recipes (which they tend to call profiles, but this is very confusing to users, because they can be confused with ICC profiles, which they definitely are NOT). In this case, the RIP simply ignores any embedded ICC profile and just outputs the numbers.
More traditional RIPs (Poster print, Studio Print, ColorBurst, etc) can properly interpret incoming CMYK data with an ICC profile and make whatever conversions are necessary to give you an accurate print. Of course, accuracy depends on having accurate profiles for monitor, printer, paper, ink, etc.
So, if we are talking photos destined for fine art printing on an inkjet, I normally would tag my files with an ICC profile.
I recently designed a job for a client using InDesign. It included a lot of photographs, illustrations, and native components created using Photoshop, Illustrator, and InDesign. The problem was that the customer had not selected a commercial printer, so I had no idea where the job would be printed, on what type of press, ink limits, paper stock, etc. What to do? I chose a middle of the road profile (IDEAlliance SWOP2006_Coated3v2) which has an ink limit of 300 and assumes a #3 coated sheet. I left this profile active in InDesign. I edited my files in Photoshop and Illustrator and tagged all images with the same profile, and placed them into InDesign. So, all components had the same "generic" CMYK profile based on a middle of the road industry standard (I use the term standard loosely). Having all components tagged enabled me to generate an accurate, high quality color proof on my inkjet (assuming the final press will adhere the the above SWOP standard).
Before submitting the file, I converted it to PDF/X1-a using Acrobat Pro. This strips all profiles from the file, but names the intended color profile in the PDF file. So, I was supplying an untagged file, but if they wanted to, they could see what the intended profile was. I stripped the profile because I did NOT want the unknown commercial printer to convert my carefully generated files. One concern was that all items on the black plate (black text, lines, vector art, etc) might somehow be converted through Lab to a new CMYK profile, thus giving me CMYK separation of my blacks, potential registration problems, etc. To cover my butt, I supplied my color proofs with the file and told the client the final job should look very close to my proofs if the printer did his job right. It also gave the printer a guide print to help them get it right. So, to me, that is the major reason to supply an untagged file.
Sorry for the long diatribe.
Thanks Lou for explanation.
What I left out was that these are photos that will either be on single pages of a book or part of a photo signature. I never know in advance who the printer is so can't and don't want to create a product designed specifically for their presses. The photos are CMYK or greyscale.
The operative sentences in your explanation--for me-- is "So, I was supplying an untagged file, but if they wanted to, they could see what the intended profile was. I stripped the profile because I did NOT want the unknown commercial printer to convert my carefully generated files."
So, if I understand right, I want the printer to print WYSIWYG. If there are any problems I see in the proof (almost always too dark or too light), I fix it myself.
I hope I have it right that I should then choose not to use any color profile.
First, it is important to know whether you are printing on a web press (high speed press using rolls, usually large quantities) or a sheetfed press. If you don't know, then it is best to base things on the lowest common demoninator, being web press. It is also important to know whether you are printing on coated or uncoated stock. Coated papers can usually take more ink than uncoated papers. For coated stock, you should be able to use a total ink limit of about 300 max and be safe on nearly any decent press. I'll assume it is coated stock if you are doing high end photography. Of course, if it is grayscale, the highest you will ever reach is 100%, since you will use only one ink.
By picking an "industry standard profile" (such as US Web Coated SWOP v2, or SWOP2006_Coated3v2), and stripping out the profile, you are essentially instructing the commercial printer to use your numbers and not to muck with anything. This is a good thing IF the printer prints to a standard, has good process control, good gray balance, etc. If they don't, you may have a color cast (at least with CMYK). Using Heavy GCR will help avoid color shifts on press if you are doing B&W work with 4/C process inks.
WYSIWYG isn't really applicable when talking about RGB, CMYK, or even grayscale unless you also include a profile (ie, the intended output intent). CMYK is partial information. For example, take an RGB file, open it in Photoshop, then assign sRGB. Now assign Adobe RGB. Now try ColorMatch RGB. The appearance of the file as displayed on your monitor changes from one assigned profile to the next. The RGB numbers PLUS the profile gives you unambiguous definitions of color for each pixel in your image (they relate back to Lab, which is device independent and unambiguous). CMYK is the same. The press, paper, inks, process control, humidity, phase of the moon, and mood of the operator all come into play. I will say that CMYK without a profile is generally much safer than RGB without a profile, especially on press, because at least there are long held conventions about total ink load, gray balance, target densities, etc.
If circumstances allow, it is helpful to be able to submit an accurate hard copy proof to the commercial printer when they run the job. They can use this as a guide, and if their proof looks a lot different than yours, then at least they should be alerted that something is not right. If you cannot submit a proof, then at least try to get the commercial printer to supply YOU a proof of some or all of your images, so you can sign off on it before the job runs. If you can't do either, then hopefully, they will adhere to an industry standard, maintain good gray balance, process control, density, etc. A PDF/X1-a file gives them untagged CMYK numbers, but it also tells them what the output intent is.
No! Include a profile in Photoshop and InDesign. Always include a profile in Photoshop and InDesign. I think Lou would agree.
Untagged files are mystery files.
PDFs supplied to printers are a different story. You can indeed submit CMYK PDFs with untagged data. But that option is available because PDF is an output format, not a design format.
And even when submitting an untagged PDF it has to be converted to the proper destination. Best example, PDF/X-1a to the appropriate press CMYK. All of the data is untagged but has an Output Intent.
But please don't go stripping profiles from your design. That's insane. Appearance is lost.
Just take a very simple grayscale image. Duplicate the image.
1st copy Convert to Profile. Go to Gray Dot Gain 10.
2nd copy Convert to Profile. Go to Gray Dot Gain 30.
Notice how the appearance stays the same. Now on each copy, go to Assign Profile, and select Don't Color Manage.
Now you've got one dark image and one washed out image.
That's why you need CM, to communicate appearance. That's just grayscale. Imagine how screwed up things stand to be when you introduce the color variables.
I am a color management junkie, and I always prefer to have my files tagged, whether RGB, grayscale or CMYK. I also prefer to work with color savvy printers who understand and use ICC color management. So, yes, I always edit files and tag them, and my InDesign files are also color managed. This also allows me to do accurate soft and hard proofs (given the limits and accuracy of profiles, etc).
When I design a file for an unknown printer, I generally try to provide a PDF/X1-a file, since I have no idea who will print it, their RIP, their capabilities, press, etc. The only thing I am reasonably sure of is that it will be printed on a sheetfed press on coated stock (could be #1, #3, #5 or whatever).
I have been priveleged to be able to pick my printers in the past, and only worked with those who demonstrated a good understanding of color management, had high quality, good process control, etc. And I almost always attended press checks to make final tweaks during the print run. It's only been in the last few years that I have been thrust into the murky world of designing jobs and not knowing who would print them. I can't say I like it either, but that's life.
It was my understanding (perhaps wrong) that if I supplied a PDF/X1-a job, the printer would print my CMYK numbers without converting to a new CMYK profile. A conversion could turn my lovely K-only plates into 4/C rich black (yuck) and completely change my black generation. Sometimes I want Heavy GCR, sometimes I don't...it depends on the image. That, to me, is the whole point of supplying an untagged file (even though the output intent is known). Knowing the ouput intent, the commercial printer can make an accurate proof on their proofing equipment. Then, they can do whatever magic they need to do during the press run to match their proof. I would think doing a conversion, without instructions from the customer places a printer in an awkward position, because if they convert and print, and the job is not acceptable, they are holding the bag since they altered the file. Applying curves to their plates to match a SWOP standard is a different story, since it is not really a conversion, per se, but tweaking their plate making apparatus to make sure they achieve industry standard densities and a good gray balance on their press.
Perhaps I am missing a few pieces of the puzzle on the printer's end, since I am not a press expert, so I would appreciate being enlightened if I am in error. I want to fill any gaps in my knowledge, and I'm sure there are many.
Dear Lou and Rick,
Thanks very much for this discussion. I am very much a newbie when it comes to color management. You have given me much to ponder.
I did not say earlier that the work I submit is always in a PDF. However I do not use the PDF/X1-a or other built in settings.
I'll be interested if there is anymore you have to say. Actually, are there any good "Introductions to color management" books out there? Or articles?
We've barely scratched the surface on the subject of Color Management. Color perception is a complex field, and there are many factors that contribute to what we see and how we see it.We've hardly touch on the subject of viewing light source, color temp, light intensity, etc.
Real World Color Management by Bruce Fraser, Chris Murphy and Fred Bunting is an excellent book on the subject. It gets pretty deep pretty fast, but presents the subject well. I recommend it for anyone interested in getting their arms around the subject.I also have a few articles on the subject on my website, which you are welcome to download if you wish. (http://www.dinagraphics.com/color_management.php). Click the links on the left side of the page to read or download the PDF files. Perhaps they will help ground you in the fundamentals.
Color Management takes time, effort and study to master. I've been working at it for 10 years and am still learning. The good thing though, is that my monitor to print match is excellent, and my inkjet color proofs are usually a dead ringer for those I receive from qualified commercial printers who know what they are doing (not all do, unfortunately). I got into this stuff to get better control over my own photographic prints and for press work, since my proofs didn't match the proofs or press sheets coming from the printer (and my boss complained, rightly so, that my blues looked too purple).
Hope this helps.
I am in complete agreement with everything you said.
If a PDF/X-1a is submitted to a printer, it should go through the printer's RIP with numbers preserved. It is a solid output standard. The only thing is, the designer needs to be certain that the CMYK destination used is a fair representation of the print condition.
The point I tried to make before was that a designer should not strip profiles from all the design elements. All Photoshop images should be tagged. All Illustrator documents should be tagged. All InDesign documents should be tagged. In the case of ID, 2 tags, RGB and CMYK.
Thinking beyond PDF/X-1a – let's assume the designer supplies native files. The print industry standard is to preserve CMYK numbers, unless total ink limit is a factor. In the case of tagged RGB images, they will be converted to the InDesign document CMYK space (that is definitely open to argument, and I could dive deeper but for now I won't). In the case of untagged RGB images, they will assume the RGB space of the ID file and be converted to the ID CMYK space.
In the horrible instance of untagged RGB images and untagged ID, they will assume the printer's working RGB and be converted to the press CMYK.
If a designer wanted to conform to a complete late binding workflow and supply native files, I would suggest designing using all source color (RGB or Lab). In the case of RGB all data should be tagged.
The InDesign document should have no CMYK information, and no CMYK tag (unless the designer knows the destination, but then, is it really late binding?).
The printer assigns a CMYK destination and processes the job.
So far in all my years of prepress I have never processed a job with that scenario. But in my mind that's what late binding is all about. The designer does not know the destination and submits all tagged source color.
That is one problem with PDF/X-4. It preserves source color but has a CMYK output intent. The printer could choose the alter the CMYK destination, but if he did that without consulting the designer first there could be disagreements. (Again, that standpoint is open to argument).
I'm glad to hear I got it right!
I have thought about designing jobs (when the final printer is unknown) using InDesign and Lab or RGB and submitting files in that format, but I always chicken out (of the two, I'd probably choose Lab, since it is unambiguous). I'm just not comfortable that the as yet unknown commercial printer will know what to do with the files. I also worry about them keeping "black only" items on just the black plate.
This path also denies me the opportunity to do final tweaking in CMYK to make my images look their best. I far prefer to convert to CMYK myself, use the black generation and rendering intent I consider appropriate for each image, then tweak color, contrast, adjust total ink (if necessary), etc, all within Photoshop. I always felt that supplying PDF/X1-a CMYK files to an industry standard would generally be the safest way to go. Since essentially all my work is sheetfed on coated stock, I'm usually safe with US Web Coated SWOP, SWOP2006_Coated3v2, etc (assumes it will be printed in the USA).
The fact that you have never seen a late binding workflow in all your years of pre-press accentuates that fact. I know some people are doing it, but I'll wait until it becomes more mainstream before walking down that path. There are plenty of printers out there who won't even accept RGB files (not sure if it is because they don't know what to do with them, or whether they don't want the responsibility of converting and taking any blame if the color doesn't meet expectations).
Thanks for your response. You clarified a few points I felt semi-confident about. I'd love to spend a month working for a competent commercial printer.
Where do you work, by the way?
There are definitely huge problems with late binding.
Right now I am working on a job with an impossible blue. And the source files are sRGB, which you think would make things easier but not really.
The beautiful royal blue always goes purple with a straight conversion. Lighter shades go violet. Overcorrect and the lighter shades go too cyan.
I work for a mid size printer in US, I'll be happy to discuss more via PM.
Your scenario is one reason I work in CMYK. I can control the final look there and tweak it to get as close as possible to the desired color. I try never to show RGB files to clients, since I don't need to build up their expectations only to be disappointed later. Who needs that headache? If I do supply them an RGB file, it is sRGB and I make sure all colors fit in a standard SWOP space.