HOW *I* DO IT :: Coloring

Let's color! We'll use that Robocop strip again, which I've recolored and turned into a print you can get on Inprnt.

Okay so where did we leave off last time? Oh yes I remember now, we had a super duper sharp black and white bitmap file, 1200 dots per inch of sexy high res magic. If I remember right, I had convinced you--it was tough, you're pretty stubborn--that thresholded black and white files, where each pixel is either black or white, are the way to go for black and white printing. You were like "whoa but it looks all chunky and stairsteppy when I zoom in" but then I was like "relax baby shh you're looking too close, this is HIGH RES we're talking about here." And I was right! What you worry might be jagged edges will not only look smooth as silk in print, they'll look a LOT smoother than they would otherwise. Thresholding is your friend!

When you tell a printer, whether desktop or otherwise, to print something with sharp, exact edges, the printer knows exactly what to do--it's the same with color as it is with black and white, although of course it's a lot more complex in places.

Before I start, props to King Colorist Alec Longstreth, who taught me about 80% of this technique. I asked him for help, and he actually called me up and talked me through coloring a panel over the phone, which I'll be grateful for until the day I die. If you're not already a fan of Alec's, please let me direct you to his site, where you can find his excellent Phase 7 comic series. Not to mention his work coloring Aaron Renier's critically acclaimed The Unsinkable Walker Bean graphic novel series.

And as in previous posts, please remember that this is just one of a million different approaches to making, scanning, coloring, printing, and selling your own comics. The only rule that ALWAYS applies is that any way you can make it work, make it work. Get out there and get your hands dirty and figure it out.


So last week we ended with a bitmapped file. Bitmap is just a file format, basically you're getting rid of any information that isn't "these pixels are either black or white." That means no layers, no funny stuff; and while the file will be editable, you can only do a few different things--mainly turn pixels either black or white. Bitmap is really best for transmitting files--my black and white bitmapped Robocop comic from last week, a 7.5" wide, 1200dpi image, is just a little over 1 megabyte, which is nothing. But in order to color it, we've got to unbitmap it, assuming we did in the first place--for instance if you go straight through to coloring, or unlike me don't like keeping a high-res copy of the art at that stage. I'm a digital hoarder.

With your bitmap file open, go to Image > Mode > Grayscale, which will switch it to a grayscale file. That means now you can do a lot of different things to it, layers, shades of grey, etc., but you'll notice the file jumps in size right away. If it asks you for ratio, it should default to "1", which is what you want just in case.

1200dpi is too big for a color file--you'll end up with a massive 800MB monster on your hands. So we'll give up some resolution to keep the file size and manageability down. Go to Image > Image size. In the dialogue box, make sure you check "Resize Image," then change the dpi to 600. Photoshop will essentially redraw the thing--it'll look like it shrinks, but it's the same physical dimensions. The number of dots per inch has dropped, thus your monitor displays it differently.

Another way to change the resolution is with the Crop tool--in fact, I need to crop out all those margins and stuff--I don't need them for print, and they'll only jack up the file size. I can always add them back in later if I feel like it--it's just whitespace, after all.

With the Crop tool selected, draw your marquee as tight as you can get it to the image edges. I print my strips at 6 inches wide in my minis, so I set my crop width to 6in, and the resolution to 600. Once I hit enter, boom! 600dpi image, 6 inches wide, ready for magic.

Now that it's in grayscale and at the right size, convert it to a CMYK color file by going to Image > Mode > CMYK Color. What we're essentially going to do is VERY simple, probably a lot more simpler than most coloring methods--I have a lot of trouble not getting obsessive over things, so the less things for me to worry about, the better. That's why I like Alec's method so much. Plus it goes well with my personal coloring aesthetic, which tends toward flat fields of bright colors, with no gradients or lens flares or modeling. Not that I hate that stuff, although in the wrong hands it can look pretty terrible. But I just like to keep it simple, and I'm not a very good colorist so the simpler I keep things, the less potential mistakes I can make.

So let's get started: go ahead and Save As this file as some new version, so you're original scan and everything is separate, in case you totally screw up and need to go back. Then create 6 new layers, and name them like this, top to bottom:

In case you can't see the image, it's ART, TRAP, PLAN, FGRND, BGRND, WHITE. Click on your Background layer, then go to Select > Color Range. In the dialogue box, put your cursor over something you know is solid lineart, and click it to select all pixels of that color (black). Make sure your Fuzziness is set to 200, although honestly I'm not sure why. That's just how Alec trained me. When you hit Enter, you'll have all the black pixels selected. This is another great reason to do "thresholded" work--it's very easy to create very exact selection areas, which means you get no fuzziness, anti-aliasing, and embarassing gaps between things. All those little stairsteps lock together nice and sweet.

Now with your selection still active, go to your little color selector and click on the foreground color. In the Color Picker dialogue, you want to set your foreground color to the CMYK value 60/40/40/100, which is known as "rich black." It's a black color that is not only 100% K (black), but also some of the other three inks in a four color process: Cyan (C), Magenta (M), and Yellow (Y). It gives you a denser, fuller black.

With your foreground color set to that rich black, turn off visibility to the Background layer and select the ART layer at the top. Go to Edit > Fill and then select Foreground Color, to fill your selection with that sexy rich black. You should end up with just black lineart with no background, and your selection "marching ants" still trooping around all over the place. Set the Layer Style (in the Layers palette, the little dropdown thingie) to "Multiply."

Now turn off visibility to ART, and select TRAP. With your selection still active, go to Select > Modify > Contract, and when the little Contract Selection box pops up, enter "1" in the box. "2" is okay too, especially if you're working at a higher than 600dpi resolution. This is essentially what's happening (I've turned ART back on so you can still see the original lines):

What we're going to do is create a trap layer--basically a slightly shrunken version of the lineart--the different between our shrunken version and the "real" art will create an area where the colors will sneak under the lineart, so if there's a plate shift during the printing process, we have a little buffer. And it's easy enough to do that I just do it all the time, even though most of my own printing I do myself on desktop printers--when I send these off to a "real" printer, whether it's digital or offset or whatever, the files are nice and tight and sharp and I don't have to go back and redo everything. Why do the work twice?

Okay anyway. So now we have our contracted selection. With ART invisible, and TRAP selected, go back to Edit > Fill and select "Black" from the dropdown. This layer isn't going to be seen in the final product, so it's not important to make it some super-black, and I like having a slight difference in the two blacks so it's easy to remember which layer I'm looking at when I'm switching back and forth later.

Now turn visibility for TRAP off, and select WHITE. Go to Edit > Fill and choose White from the dropdown. You'll end up with a whole layer of white, which is going to be our background, which we'll keep separate and not really mess with again, most likely.

Once you're done with that, drag your Background layer to the trashcan--we don't need that anymore. Now you should have 6 total layers:

1) An ART layer, which is the actual lineart, all the black pixels in a more dense black, sitting on top of everything.

2) A TRAP layer, a contracted version which we'll use as our working blacks, to all the colors to sneak under the lineart for trapping purposes.

3) A PLAN layer, which we'll use to test out what we actually want to do, and then have on hand as reference throughout the actual coloring process.

4) A FGRND layer, where we'll do our more fine work, things in the foreground that necessarily need to sit "above" the simpler or less important background shapes.

5) A BGRND layer, which will most likely (in my work anyway) be larger swaths of color, sitting under the figures in the foreground.

6) And lastly a WHITE layer, which just gives us a nice simple ground to set everything against.

Now let's do some coloring. For this I'm going to use the Robocop print I just did, which turned out pretty well. I have a fairly limited understanding of color and color theory, so for this one I got my girlfriend to select a palette for me--she's an interior designer, so she works with color all day and is super comfortable with it. This is what she came up with:

Pretty sweet palette, am I right? So using those colors, I went in and made a really rough splash around, just testing out where to put which colors. I'm not going to get too deep into theory here, mainly because I know almost nothing about it, but my two goals when using color are 1) to move the eye through the strip in the way I want it to move; and 2) to create balance if possible, both on the page and aesthetically. For instance, Robocop is a total nutballs over-the-top violent movie that BARELY avoided an X rating when it came out in 1986. So using a friendly, bright, simple color scheme makes me laugh a little bit, and makes things a little less intense. I'm not very intense.

Here's what I came up with in my plan:

If you look at the final image, you can tell I changed a bunch of stuff, but this gave me a good starting point, and kept me from throwing up my hands halfway through when it got too complex. A lot of people won't even need a PLAN layer because they're better at dealing with color and holding information in their heads; but those are not talents I possess.

Now that all that administration is out of the way, we can actually color. I use a Cintiq, which is pretty great, although they're super expensive and now that I'm flat broke it's hard to justify to myself over say, just a regular non-display Wacom tablet.

We're going to be primarily using three tools, which we'll toggle a good bit between: Pencil, Eraser, and Paint Bucket. Make sure Anti-Aliasing is turned off on all of them, so you maintain those nice crisp lines. With your Paint Bucket tool selected, also make sure "Contiguous" and "All Layers" are checked. I'll show you why in a second.

Okay so we'll focus on one panel: here we see poor Officer Alex Murphy, Robocop Himself. I have my colors more or less picked out as you can see, now I just need to go in and do the sharper work. I also make sure I have my color palette added to my Swatches palette, so I can get to them really quick without a lot of confused hunting.

Turn all your layers visible EXCEPT "ART". I like to start with the simplest shapes in the back, so select BGRND. You should see your slightly thinner TRAP lines, and then be coloring under them on the BGRND layer. What we're going to do is essentially use the Paint Bucket tool to fill large shapes for us--if you have "All Layers" selected and click in an area, it will fill everything until it bumps up against filled pixels of any kind, on any layer. Go on and try it, you'll see. Unless you're really good at creating tight, closed shapes, you'll get a little more than you want. So go in and "close" the shapes by filling those gaps with color so Photoshop reads them as closed, like below:

Once you have any little gaps filled, you can fill the rest of the area with your Paint Bucket tool:

Then go in and lather, rinse, repeat with other areas/colors, working mainly with simpler shapes:

Until you're done with the background layer more or less:

Once the background is done, you can select your FGRND layer and put in little touches on top. For more complex coloring, you might have more layers, for instance if you need to group certain colors or objects so you can easily change them all at once. Maybe letters or a flock of birds or something. But whenever possible I try to stay simple, and this particular image lent itself to that (fortunately). I added little touches to Robocop's face, including those precious little cheeks of his, some highlights on his body, and so forth. Then I turned the ART layer visible and whoa! It's like you've been running with weights!

A few more:

And there we go! Like I say, this is how I do it--with coloring, there are literally four billion different ways, with gradients, channels, Manga Studio, and of course just straight up coloring on the paper, like with actual paint (?!?!). I say go for it. Get in there and splash around. Next installment I'll go into printing a little bit, see you back here for that stuff!

:: Comment

Content © 2018 by Dustin Harbin | Site design by Harbin and implemented by adult