So yesterday afternoon my wondrous iPhone, survivor of a some-35-mph flight off the roof of my car onto the middle of a busy street in north Dallas and then an hour or so sitting in the middle of that street till I came and rescued it…froze up.

Just as I was getting off the plane at LAX. Just froze up bigger than shi…well, it froze up.

On the rental car bus some guys cheerfully offered the advice to “pull the battery” and when I allowed as how you can’t pull the battery on an iPhone, they further offered the cheerful advice that “iPhones suck.”

Thanks guys.

Anyway, I’ve made a living the better part of my life going to places I don’t know all that well and getting around in them. And I know LA some. But not the part where I am now. And before I got so dependent on this silly little contraption I used to print maps: Airport to hotel; hotel to project; Starbucks in the area.

Not anymore, boy. Just enter it into the wondrous iPhone and let it lead the way.

Well, times being hard for everyone right now and this client not being a huge conglomerate and me working to save him some airfare, I had flown into LAX as it was much cheaper, even though the guy’s shop is much closer to Ontario. I made the hour and-a-half trip on “the 10″ out to this part of the LA sprawl, got off around Ontario Airport, and engaged in an hour-long search right out of “Planes Trains and Automobiles” for either my hotel, or the street my hotel is on, or WiFi that worked.

Ever wonder if they even make pay phones anymore?

Does anyone even carry change to make a call if they do?

Does anyone under 25 even know what a pay phone is/was?

Well I got here, finally. I wound up at a McDonalds where their WiFi didn’t work but the girl at the counter gave me the phone to talk to their “tech support” and after about five minutes on hold it occurred to me I could just call the hotel and get directions.

Moral of the story?

Sometimes ink on paper is better; from now on I’m always gonna print my maps, wondrous iPhone or no wondrous iPhone.

Onward.

So the week is half over and updates from the field have been a little slow in coming…meaning that there are the inevitable kinks to work out of the operation.

However the project at Z Decals is done. And Nick, the owner of Z Decals, is now one happy guy. They do almost exclusively wraps, for cars and trucks and boats, and their design style features a lot of grey. Not even neutral earth tones.

Just shades of grey with vivid color highlights.

If you do wraps and you use grey then you’re well aware of what lots of people call “metamerism” which really isn’t; the tendency of greys to shift to a green cast in sunlight.

There’s a specific reason for that and there are ways to combat it–although unfortunately at some cost in other areas. So I made Nick some laminated and unlaminated profiles with different black generation builds and now he’s got enough arrows in his quiver to nail the bulls-eye on just about any grey project.

And, as is typical with guys running Rolands driven by Versaworks, Nick was so concerned about getting good greys that he made very, VERY clear in our discussions upfront that he was just fine and happy with every other color. “Everything else prints great” were his exact words.

Well, “great” has now been redefined. The profiles in Versa are all made with the same basic profile-making philosophy. And what that philosophy tends to make is profiles that aren’t horrible, but are by no means capable of getting all out of a machine that it’s got to give. We had a set of before-and-after panels we were looking at outside, and while everyone was cooing over the greys, I pointed out to Nick the difference in a red swoosh that ran through the image. Compared to the “after” which was a full and vibrant and vivid red, the “before” looked almost orange.

So where’re all the videos and documentation?

I stumbled a little bit with the Flip camera I’m afraid to admit. I’ve got a bunch of pretty shaky video of the printhead moving and random bits of this and that and a bunch of feet standing around outside looking at comparison images while their voices talk about how wonderful it is to have the problem solved.

So I opened iMovie at Birmingham Airport and figured I’d start watching the tutorials and see if I could paste all the bits and pieces together into something worth sticking on You Tube, but of course the tutorials are all online and there doesn’t seem to be any sort of WiFi at that airport.

So that’s where it stands. Today I’m reading tutorials at home, then off to LA, and I’ll keep the little camera rolling much more at Fineline.

More to come.

So this week I’m going first to Athens, Georgia–which is a little ways outside of Atlanta–and then to San Dimas, California, which is somewhere in the midst of the LA sprawl.

And what’s kind of interesting to me is the similarity of the two projects.

Now, one of the questions I get asked occasionally, or one of the considerations I hear voiced goes something like: Do you work with little guys like me, or is it worth it for a little guy like me to use your servlce?

Well, I don’t think either of the guys I’m going to see would classify their operation as big.

Each is an operation with one Roland printer.

In Georgia, Z Decals is running a VersaArt 64, and in California, Fineline Signs is running an SC 540.

Z Decals is running VersaWorks, and Fineline is running ColoRip.

And the issue bedeviling both of them is neutrals.

Each project is only going to take a day. And both of these guys are existing in today’s economy just like the big guys, just like me, just like you, and just like everyone else. Yet these guys are serious enough about what they do to invest in making how they do what they do as good as it can possibly be. And in the end I guarantee you, it’s going to save them more money than they spend.

Anyway, I thought I’d bring a Flip video camera with me, take this blog on the road, and do a little documenting.

It’s gonna be fun, fun, fun.

If you’ve ever been to a large-format trade show of any kind you’ve seen her.

And one thing good/bad about her, she’s so colorful that even if she’s printed poorly she still looks pretty good.

But do you know anything about her?

Um…if you’re waiting for me to answer, I don’t either.

I have no idea who she is or who the photographer was or for what reason she was taken.

But I do know that she is far from just another pretty face, or just another bright colorful image. She’s so perfect for evaluating ICC profiles that either she was made by someone for exactly that purpose, or she’s one lucky accident.

Want to see just how your profile is clipping greens? Right there in the makeup swoosh that goes from under her eye to her hair.

Magentas? Right there in her cheek.

Wondering about your reds? There in her lips.

Neutrals? Amazing what the little warm grey area in the lower left can tell you.

And the area under her chin will damn near tell you all you need to know about your ink limits and even your black generation.

There’s more. If you have any issues with blue/purple, she’s going to highlight them. Note in particular the transition of shade from blue to purple in the makeup over her eye.

Outside of the Granger Rainbow, she’s the best profile test image I’ve seen, to the point that I’d feel a little lost without her; yet to this day I still don’t know anything about her.

Do you?

As part of my campaign to become interconnected in the whole Web 2.0 world, I actually got a Twitter account.

I said I never would, just because Twitter is such a stupid name and Tweeting is such a silly-sounding thing to do, and also because I refuse to use letters or numbers for perfectly good words.

But all that aside, I did. And so I’ve taken every so often on Twitter to word-searching some terms like “ICC profiles” and “printer profiles”, and of course, “color management.”

So I was word-searching this morning and when I tried “color management” one tweeter had tweeted this:

“I wonder if the engineers who designed color management intended to make it so confusing and frustrating.”

Well, gee…

Know what?

I remember when American Airlines moved to Fort Worth from New York. They went through a process to pick a new ad agency, and part of the winning agency’s presentation was to take a photograph of the meeting room where the principles were assembled to pitch and catch; then at the end of the presentation some two hours later, they handed each of the people present from American a printed image of that same photograph.

And they won the account based in large part of that amazing feat. And that amazing feat made the freakin’ front page of the business section of the Dallas Morning News.

The machine that made this amazing feat possible was the first computer color scanner I ever saw. It lived in a clean room, had its own cooling system, cost half-a-million dollars, and I remember standing with a couple clients damn near slack-jawed the first time we got a demo and saw a background sky changed from one color to another on a screen.

And I don’t know how many hours I spent back in the day at the end of a printing press with clients trying to get color just exactly so.

Now any monkey can get a colorful result. The tools are cheap and any idiot can plug them all together and make them work well enough to get a surprisingly passable result.

But color is still science and it’s still art. Anyone can turn color into numbers, but it gets damn dicey keeping control of the numbers from the starting line to the finish line if you want them to turn out looking professional as opposed to passable.

Fact is that there’s a lot to it, and there’s a lot to learn to make it right. It isn’t that anyone “designed it that way.” It’s just the way it is.

So I was asked the question by a commenter just how familiar I am with the G7 method, and just what I think of it.

Interesting question too. Because as it so happens, G7 has been on my mind a good deal of late.

You’ve probably seen G7 and heard of G7 and probably know that it’s got something somehow to do with this whole field of color management, but do you know what it is, where it came from and how it works?

Well, G7 was developed by Don Hutcheson of Hutch Color, who then–as I’m given to understand–handed over the rights for distribution and promotion and marketing to IDEAlliance.

The name G7 refers to Grey, plus the seven ISO standard colors, CMYKRGB.

So what’s it for and how does it work and what does it do?

Well, back before there were computers or digital images, there were was a move to make standard densities for ink on litho presses. The Holy Grail was and has been in some quarters to make it so that printing was simply a matter of conforming to standards, so that all printing everywhere could be exactly the same.

So all things being relatively equal, if two printers are using inks that conform to a standard, and printing to densities that conform to a standard, then their prints should match, right?

Well, not always.

It gets a little trickier, what with a phenomenon that used to be called “dot gain” but now goes by the more official sounding sobriquet of “Tone Value Increase” or TVI.

And each color on any litho press is going to have some sort of unique “TVI” curve. You can picture it something like this: If you’re not printing any of a color, then since you don’t have any dots, you’re not going to have any dot ga…er, tone value increase; and obviously you also won’t if you’re printing that color at 100%, because you haven’t got any dots to gai…er, tone values to increase in a solid; but where you do have dots, you’ll probably have the most TVI in the mid-range, where they have the most room to gain…er, your tone values have the most propensity to increase.

Anyway, you get the idea.

So if you’re following along here, what happens is that even though two presses might be using the exact same ink and printing to the exact same densities, since they have different TVI characteristics, the internal elements of a same image on the two presses might not be the same.

Okay, now I like to rail against standards a bit, but there are certainly instances when this is important. If you’re an ad agency and you’ve got a buy in several magazines, all you want is to approve the proof and have all the magazines print the ad exactly the same.

So lets fast forward a little to the present day, or at least to an ICC world, and now printed images start as files. And now there are ICC profiles that define the pixels in those files as they relate to the color spaces in which they were created, or to which they have been converted for whatever purpose.

And there are some standard spaces out there like SWOP and GRACol, and wouldn’t it be a wonderful world if all anyone had to do was just use whatever standard profile was called for by the final output, and that would be that, and wherever you printed your image, the standards would be met and it would be exactly the same.

Now here we’re still talking about litho. And at this point if you want to have a proper characterization you’ve got two real choices. You can profile your press and somehow or another convert your file to that particular profile before it prints, or you can calibrate (linearize) your press so that it will print with a standardized profile in an expected way.

And that’s what G7 does. G7 is press linearization, in essence. So that once the press is linearized, it should print files in standard color spaces exactly as expected.

In fact, the G7 manual doesn’t even really recommend you do an ICC profile of a press once you’re done with the G7 process. And if you think of it, why should you?

You know your process colors as defined by ISO standards, and your densities as well, so you know your total gamut is going to be exactly the same as your standard profile.

So G7 definitely has a place in this world…

But I’d argue that place isn’t large or grand format printing.

First reason is that if you’re using a correct and completely color managed workflow, G7 would be somewhat out of place and redundant to start with. There isn’t too much likelihood that your grand format UV machine is going to print correctly if you use SWOP or some other standard print space as its destination space. It’s a given that you’re going to have a custom profile for that machine on that media in the first place.

And it’s a given that you’re going to linearize that machine before you profile it. If you accomplish both those tasks well, the whole reason for the G7 method becomes a little moot.

Now, there are people out there who hold that if you use the G7 method to enhance your linearization you’ll get a linearization so true that profiles truly can be used from one linearization to another…but, that seems more like theory than practice to me. And it still doesn’t make a great deal of sense, since it’s going to take less time to just do a standard RIP-based linearization and then reprofile than to go through the entire G7 routine.

And that of course would be just using G7 to linearize. I’ve seen people touting the G7 process to large and grand format printers, and I do have to say that if that includes attempting to print to “standards”…well, I think that misses the point.

Large and grand format printing aren’t about printing magazines. Large and grand format printing are about big; about bold. About stopping someone on the street in their tracks with a visual and demanding their attention.

If you capture all the gamut your machine has to give in the profile-making process, you can always tell it to emulate a smaller-gamut device if need be for those times when standard colors across a campaign are important.

Let’s say you inhibit your half-a million dollar machine via ink limits so that in the name of standards it can only print the gamut of a web press, can you now use SWOP as its print space?

No. Because it’s not a web press. You still have to do the conversion to your machine space. And what you’ll get is a print that–assuming a well-done profiling job–is a good match to a SWOP proof. However, if your machine is profiled to give you every bit of color it’s got to give, and you send it that same SWOP file, what will you get?

An image you can lay side by side with the other one and it will be identical.

Myself, if I paid a quarter to a half a million dollars for a machine that can print three times the color gamut of a web press, I’m not about to throw all that capability away in the name of “standards.”

Don’t get me wrong here. I think the G7 method is ingenious, and a very elegant way to solve a very old problem. And in solving that problem I’d recommend it above just about any other solution.

It’s just that to me that does not make it one size fits all, or the absolute be-all and end-all of printing device characterization, as many of the people out there hawking it might suggest.

A couple weeks ago I got a call from my friends at eSigns.com just outside Detroit. They’re not only friends but a customer, and I’d been up there a year or so ago to do workflow management for them and to profile their Mimaki JV-33’s.

Seems they’ve invested in a brand new Gandi 1224 UV Flatbed–which is a great printing machine, btw. They’re running it with Onyx and while it was printing pretty much okay, they’d come across one job that was giving them fits. The color in question was a corporate color, a magenta-pink, and running the Gandi-provided profiles, they were getting a sort of lavender-salmon color.

That’s the way it can tend to go, too. You may have similar situations yourself. You’ve got a profile you got from somewhere and most of the time it prints okay or even well, but sometimes there are colors you have to chase for half a day to get, or sometimes you just can’t get there at all.

Now like everyone else in these trying times, the guys at eSigns aren’t spending money needlessly; and at first blush, it might seem like an extravagance to fly someone in to profile a brand new printer that might even have just been profiled by the guy who set it up.

Well, here’s the thing: I’ve known a lot of printer technicians in my day and while a lot of them are very good techs, none of them are color management experts. Many of them even hate color profiling. They’ll make you a profile or two, but by the time they do, the install is done and their flight leaves at 3.35 and as often as not they’re already mentally halfway to the airport.

Well, I flew up and profiled the machine and here’s the difference:

This is an actual and completely unretouched or altered iPhone picture. The one on the right is “before” and the one on the left is “after.”

No editing or altering of the printed image either. The customer file was a .pdf and we just sent it directly to the printer though Onyx. Both images were printed with the exact same file.

What’s obvious here is the background color–which is absolutely nailed here with no modifications to the file on the first try in the image on the left.

If you look at the inset picture though and all you had was the image on the right, you might think it was okay; and this photo loses some of the subtlety, but note the concrete. In the image on the left, it’s neutral and looks just like the concrete it’s sitting on, but in the image on the right, it’s got a little bit of a cast. The same cast you can pick up in the grey background by the dog’s ear.

So what’s the bottom line?

Bottom line is the client loved the color and eSigns has a big new client now. A new client that more than paid for the profiling session. Not to mention that now the machine is profiled to give them these results not just for this client, but on every job they send to it.

Just further proof that cutting corners isn’t always the path to success, even in trying times.

So we all know what an ICC profile is, right? (In this case, ICC printer profiles.)

Well, I like to think of ICC profiles as roadmaps. That’s one my training session analogies. You put them in the glovebox and they’re there whenever any application needs to call on them to get to their particular color destination.

So how are they created?

“Oh! Oh! Oh! Iknow! I know! I knoooooowwwww!…”

Yeah, I see you jumping up and down and waving your hand around. So okay, go ahead and answer.

“Well, you print up some patches, and then you get your iOne, and you look all cool while everyone walks by and asks what you’re doing because you’re a color guru now, and you read the patches into Eye-One Match and when it’s done you’ve got yourself one first-class ICC profile.”

Okay, thank you. What about creating the state of the machine?

“Huh?”

You know, creating the state of the machine. That profile you just made measured something, didn’t it? It measured color values that machine produced on the media you put through it in exactly the state it was in when you made it.

“Ohhhhhhhh…you mean like ink limits and stuff?”

Yeah, ink limits and stuff.

“Oh, well, um….”

Okay, you get the point.

Here’s the bit of it: What’s absolutely unfortunate about the ICC profile making process is that a profile doesn’t have to be a good profile to be a valid profile. You can make yourself a pretty crummy profile with some pretty crummy software and some pretty lousy parameters for the machine (or you can download yourself a crummy profile) and it’ll feed your $250,000.00 printer that crummy information just as if it was prime beef.

See, making the profile is the easy part. Creating the state of the machine, that’s what matters. The “state of the machine” is the entirety of conditions you’re characterizing when you make an ICC profile. And the profile can tell you a lot about itself before you even use it to drive a printer, if you just know where to look.

So, here’s the procedure I use to evaluate every profile I make:

First step is to open the profile in a 3D gamut viewer. I use Monaco GamutWorks, but there are others out there as well.

You’re looking for several things in a gamut viewer. First, you’re looking for the overall size of the gamut. It’s going to vary from madia to media and machine to machine, obviously; but each machine and inkset has a basic “footprint” that becomes pretty easy to recognize after a profile or two, and the gamut needs to look like it’s getting all the machine has to give for the situation that’s being profiled.

Here’s a nice gamut of a large-gamut printer. In this case an Epson 3880. It’s smooth and well-formed, large, and produces color all the way down to its maximum density.

You can see that in the area circled in red. The point along the white line is black, and extending beyond it is the profile’s ability to make color at full density. To get full, rich color throughout the gamut of a machine, this is absolutely crucial. Regardless of where maximum density winds up in any profile, it should not come down to a single black point. The machine should be able to produce color right up to maximum density. And it’s the profile’s job to allow it to do it.

And that is a function of the inter-relationship between single channel ink limits and multi-channel ink limits–the state of the machine.

I’ve come to think of it as the point of precise maximum inking when creating single channel limits. The ideal is to balance the max on each channel exactly so that you’re getting the maximum out of each channel without the combination of any of them putting down too much ink too soon; making the machine reach maximum saturation prematurely so that it will then not be capable of printing full, deep, rich colors.

Assuming the profile looks good in 3D, I then import it into my ColorSync library and view my standard demonstration image through it in Photoshop soft proof. I’m not going to show you all my demo image here, but this is a part of it.

First image is a screen capture of the image with no soft proof applied. Second is the image with the above profile applied as a soft proof.

The area in the upper left is the image I’d use to check every profile I make if I could only use one image to do it. To those who don’t know, it’s called the Granger Rainbow; it’s relatively easy to make in Photoshop, and there’s nowhere in it for flaws in a profile to hide. Note that here it has shifted–obviously–but there’s no posterization, no untoward artifacts; and in particular, in the area circled at the bottom, there are very good transitions all the way across to black, confirming what the 3D visualization showed.

Note also the the Color Checker looks good, the color swatches and ramps look good, and the gray ramps distinguish all the way down to black, and are nice and neutral from L* 0 to 100.

That all done I load the profile in place on the client’s machine and print a test print. If it matches what I see on the screen, then it’s a keeper.

Now here’s another a 3D plot of another profile.

You might be surprised how many profiles there are out there that look an awful lot like this under the hood. Note that in this profile the black comes down to a severe point, which means the machine has no ability to print full colors except in what appear as the outer areas of the profile where it can print to its single channel limits without overinking. This is what happens when single channel ink limits are set too high in relation to the point of precise maximum inking.

Oh, oh, oh….wait. Yeah, I hear you. “Hey! Don’t be talking to me about single channel limits,” I hear you saying. “I am the single channel master. Got my technique down and everything. My single channels don’t be hookin’ or nothing.”

Well that’s all well and good but it’s also beside the point. Each of your single channel limits can be just fine on their own, but the combination of all of them can still cause this effect.

Want even more fun? Here’s what the profile looks like in soft proof.

You can see in the Granger Rainbow exact confirmation of the 3D. Instead of making full rich color all the way to black, this profile runs to black way, way too early. But…

Suppose you weren’t using the Granger Rainbow. Suppose you just had the other elements here on the screen? Color Checker looks good. Patches look good. Ramps look good, including the black. Even the spectrum in the Correct Color logo looks fine. But you go trying to print images with deep rich shadow tones in them, and this profile will let you down every time.

Trust me, there are lots of profiles like this out there. Odds are you’re using some now. (If you’re running a Roland, odds are double.)

Of course if I made this one I wouldn’t do a print test with it. If it fails here it’s going to fail on the printer. Time at this point to scrap it and start over.

Want to have some fun?

Here’s a nice big Granger Rainbow you can download.

Fee free to use it to check out your own profiles and see what you get.

So here it is…

*Drumroll*

The long anticipated Correct Color blog.

Well, long anticipated by me, anyway.

So what’s going to set this blog apart from all the other color blogs out there?

Well, two things, I hope: First is I intend to write it as much in English as possible, and second is…well…

I write a column for Wide-Format Imaging Magazine now and again, and in each one I’ve written I’ve always felt at some point I’m pulling a punch here or there. Not to say I’ve said anything I felt was untrue, but maybe I’ve been a little careful with the shading or left out a name or two. After all, they’ve got to sell advertising to everyone in the industry, and they might stop wanting articles from me if I ragged all over one of their biggest advertisers. Not to mention that my opinions and observations are my own. I didn’t bring anything down from the mountaintop, and of course your mileage may vary.

But here, the opinions and the consequences are all and only my own.

But thing is while that’s all true, I’ve been around a lot. And there’s a bunch of stuff you see behind the scenes in this business when you’ve been around a lot.

And when I see them or as I think of them, I’m going to write about them here.

Like, for instance: There are any number of digital printing forums out there, and on these any number of forums there are any number of posters who’ve got this printer-profiling thing all figured out. They’ll tell you. Just ask them.

Problem is, for the most part, they don’t.

And you may very well not either. Even though you think you do.

Think you do?

Okay, once you made an ICC profile, what steps do you use to check it?

Set aside for now that actually making the profile is the easy part, and here’s a little tip for you: Of all the spectrophotometers out there –assuming of course they’re calibrated and working correctly– just about all of them will yield virtually indistinguishable-from-one-another profiles depending on the profile-making software you run their data through. But while all the various spectrophotometers out there will pretty accurately record the data you feed them, the differences in the profiles made by the different profile-making software packages are huge.

Huge.

There are over 15 basic ICC profile-making engines available out there, plus permutations, and none of them produce identical results. Some of them produce spectacular results. Some of them produce good results. Some of them produce adequate results. And some of them produce terrible results.

Want to know which does which?

Well, watch this space. But for another tip for now, the price of entry for good or spectacular is slightly over two grand. That’s for the software, not including a device. Packages that you can buy with device and software for under fifteen hundred bucks are going to give you results ranging from less-than-optimal to awful.

Okay, you may say, but I ain’t gonna mess with profiling my printer in any event. Why should I? I use Velvette-O-Smooth Premium Satin Matte, and Velvette-O-Smooth.com has profiles I can download for my printer for Premium Satin Matte, and everyone knows Velvette-O-Smooth is the industry leader, so surely their profiles are top-quality, right?

Well, maybe.

Or maybe not.

Ever checked one? How would you know?

Because, back to those things you see when you’re behind the scenes, you’d honestly be surprised. Some of the huge names in media have some surprisingly mediocre profiles available for download. And some have one or two profiles renamed for ten or twenty different media.

So, whether you made them yourself, or whether you downloaded them, how do you evaluate your profiles to see just how good they are?

Well, that’s the first thing I’m intending to write about here. Unless I write about something else first.

But whatever. It’ll be pretty lively, mostly in English, and mostly interesting.

Just watch this space.