Saturday, May 10, 2014

Getting the most color for your ink, part 1

So, my wife says to me the other day, "Hey Math Guy, you should do a blog about ink mileage - you know, making the most efficient use of pigments." I love it when she calls me Math Guy. It's so much more intimate than John the Math Guy. "Surely you've thunk some thoughts that no one ever thought to thunk before." I don't love it quite as much when she calls me Surely.

But she does have a point. I do have a few thoughts. I have been waiting for an opportunity like this to share them with an eagerly waiting world.

Dot gain is your enemy

I remember years ago hearing a competitive pride in the press room when it came to printing sharp dots. Mushy dots were a sign of a sloppy pressman. Really good pressmen, running on really good presses would produce sharp, crisp dots. Really good pressman would make Felix dots. All pressmen wanted to be Felix. All presses wanted to be Felix presses. All printing inks wanted to be Felix inks. All fountain solutions wanted to be Felix solutions.

Felix halftone dots versus Oscar halftone dots

And so it was the goal to make crisp, Felix dots, and press crews worked at reducing dot gain.When a press crew managed to bring the dot gain down by one point, there would be huge bonuses and wild parties and groupies everywhere. Many of you remember those days. Mick Jagger would show up, and the roll tender would get a call to be on the Carson show. You get it. The whole magilla.

Then those darn pre-press guys started getting involved. Stick-in-the-muds, every one of them. They complained about needing to run a different plate curve for every job. The pre-press folks didn't buy into the whole press room machismo thing. They didn't want to be on late night TV. They just wanted consistency.

And so it came to be that the printing pundits made the rounds, popping in on the Tonight Show and getting little blurbs on the bottom of page 17 of the tabloids. The message went from "Dot gain and Communism are the enemy" to a more subtle one. "Dot gain, just like Kim Kardashian, is inevitable. We can't get rid of either one of them. All we can do is try to control them."

In light of all of this, what I am about to say is heresy. When it comes to ink mileage, dot gain is your friend. Yes, it needs to be controlled and kept consistent, but more is better.

Kim Kardashian proudly displaying high dot gain

Why is there dot gain?

A point to consider: stochastic printing (FM screening) has high dot gain. That means you need to adjust the plates curves to make it print like conventional printing. Get it yet? You need to bring down the 50% when you're printing stochastic. Have you caught my point yet? You need to put less ink on the paper when you are making a halftone. Less ink for the same amount of color.

I think that the fact that stochastic screening requires a different plate curve is common knowledge, but I'm not sure that everyone has connected the halftone dots. Stochastic printing requires less ink. High dot gain means less ink.

But... maybe the "less ink" part is not obvious. Maybe I need to expound on a question that has baffled philosophers of printing science for decades. Why is there dot gain?

Every Phy-Ed major knows why there's air - to fill up volleyballs!

I know of three explanations for why halftone dots come out fuller than one would expect: more ink, more diffusion, and more squish. The physics is probably correct behind all of them, but it is likely that one or two of them are the major factors. 

More ink

The simplest explanation is that the plate simply delivers more ink. The more ink, the more dot gain. As it was explained to me by Herr Gutenberg, it all had to do with ink/water balance. When you put a little more water on the plate, it will crowd out the ink in at the edges of the dot, and there will be less dot gain. If there is a bit less water, the ink will have the upper hand at the edge of the dot, and the dot will grow.

Based on this, the model is this: more dot gain -> more ink -> more color. Simple enough.

This all makes sense, but I have tracked press runs while adjusting water up and down within reasonable limits. I saw a lot of change in dot gain over hours of press time, but little of it was correlated with the amount of water. I don't think that the "ink/water balance at the edge of the dot" theory is the big explanation for dot gain. 

More diffusion

In 1936, Yule and Neilsen came up with the idea that there are two parts to dot gain: physical and optical. They said that the dot on the paper is indeed larger than the dot on the plate, but that there was a second effect. The paper between the dots takes on some of the color of the ink because of light diffusing into the paper.

I won;t go into much detail explaining it here. You can look at my previous blog post for that. I just want to say here that the Yule-Neilsen effect gives you a little extra color for free. That apparent tinting between the dots acts like more ink, more ink that you get for free.

More squish

Noffke and Seymour came up with a little different explanation in 2012 - dot squish. (Some of you may recognize the name Seymour. He has this blog?) A pristine silo of ink is first deposited on the printing plate, and then that nasty old press comes along and presses it flat. Note that there is no change in the volume of each dot, just it's shape. 

Halftone dot transmogrification under imply pressure

But what of the color? Does squishing the dots change the richness of the color? Well, yes. I go into more detail in the blog on the Noffke-Seymour effect, and Pat and I went into excruciatingly painful detail in the TAGA paper

(By the way, the call for papers for the 2015 conference is out. The conference is set for gorgeous downtown Albuquerque, March 22 through 25. Email me if you have any questions, or want to discuss an idea.

Does this get you more halftone for your money? Let me motivate the idea a little bit by considering the extreme. Let's take that silo to the extreme. Keep the silo the same volume, but picture it becoming more of a needle - a very tall spire of ink that has a very, very tiny footprint on the paper. Being very tall, the microdensity of that ink is very high. The color at the top of that spire is very rich. But it covers an infinitesimal amount of paper, so the overall reflectance is pretty much the same as the paper. 

That tall narrow spire is the very least efficient use of ink. It is the cleanest, crispest dot possible, but it is absolutely lousy when it comes to ink mileage.

Dot gain is your friend

The first of the three explanation for the cause of dot gain predicts that dot gain is "cost neutral". You put in more ink, and you get just that much more color. The other two explanations predict that dot gain is like getting a little extra pigment in your halftone for free. My own observations are that the first explanation of dot gain is not the major effect.

So, my conclusion is that dot gain is your friend. According to the Yule-Neilsen model, whatever it is on press that causes more spread of light into the paper, like higher line screen or stochastic printing (or perhaps some reformulation of the paper?) can reduce the cost to print a halftone of a given color. 

According to the Noffke-Seymour model, whatever it is on press that causes the halftone dots to spread out more, like decreasing viscosity or increasing pressure, can reduce the cost to print a halftone of a given color. 

Moral of the story -- An efficient halftone dot is a happy halftone dot


  1. Seems like this would be somehow related to ink strength. High strength ink would be able to be squished more and still produce good color. Squishing low strength ink results in lower chroma, so you need to put down more. And then you've lost your reduction of ink usage. Or I missed something important. (posted by Dave Wyble, who does not need to be anonymous but has no accounts on the list)

  2. Thanks for the comment, Dave.

    The theory behind this is based on a few assumptions: 1) a 50% dot takes up 50% of the ink and then it gets squished out, 2) the squished out ink roughly follows Beer's law, and 3) the reflectance of the whole is the weighted sum of the reflectance of the squished out dot and the surrounding paper, with weights determined by how far the dots squish out.

    You can see the full development in a previous blogpost and the full TAGA paper:

    For the TAGA paper, we tried the model on CMYK print for web offset (AM and FM screening, heatset on coated stock and coldset on uncoated stock), flexo, gravure and inkjet printing. The model fit pretty well for all of these.

    I have not (yet) tried this on any spot color inks. I see no reason why it shouldn't work, with the possible exception of inks that have some opacity (which means they don;t fit Beer's law well). Perhaps pastel inks might behave different as well? Dunno. But, as yet, I have only conjecture.

  3. Ahhhh.... I remember the groupies.... Good times for sure!