Digital Painting Advice

This will be slightly off-topic for this blog’s normal theme, not that it has one, so feel free to ignore. Here are some generic tips for digital painting that no-one asked for – as bullshit-free as I can get them.

1) Get your tutorials from CG Society, not DeviantART

You’d think that the idea of a tutorial would be to teach a technique to someone, and to raise their ability a certain degree. Oh, my sweet summer child how wrong you are! The real aim of a tutorial – and I do mean this with the utmost offence to those people who do it – is for the artist in question to show off how good they are! At least, this appears to be the point if you flip through DA’s tutorial section. Because, of course, “then add the shading” and “now I scribble in the detail” followed by “I hope you found this useful” means I should have the right to stab you in the eyeball with a stylus. Check out CG Society instead, it is literally meant for professionals.

2) Learn colour relativity

Colour theory is one thing artists like to bang on about because it makes it look as if they’ve actually learned to some something complicated and that it might just be science, or at least difficult. But it really is useful – no sarcasm, it really is. If you put down something that’s grey, and it looks really dark, it won’t look dark once you’ve put in every other area that’s dark – in fact, it’ll look like some washed out middling grey. The same goes for colours, things that will look blue might actually be purple, or a bit red, or green, or really de-saturated (particularly true with eyes). There’s no real trick to solving this, just be aware of it and correct as you go.

See that dark-brown square in the middle of the top face and the light-yellow square in the middle of the front face? Look again, closely…

3) The paint-in-black-and-white-first trick is bollocks, sort-of…

There’s a technique in digital painting where you paint entirely in monochrome first, and then you pop in another layer, set the blend mode to “colour” and then colourise it. The theory is that this way you can establish the value – the range of black to white – first and get it right. This is important because if the value is off, it just won’t look right. In my experience this sort of works, but does effectively double your workload and really screws with your choice of colour, keeping it a little flat as you end up colourising one whole area with one tone. To compensate for this, change the colour frequently to get a more varied hue… but then you end up effectively painting it all again anyway, so it’s arguably pointless. If it sounds like too much work, simply pick a program that lets your display the image in black and white monochrome, and use that to check the value.

4) Photoshop is kinda overrated, too

There’s no need to hand your hard earned cash to Adobe. Or even torrent the thing, to be honest. Unless you’re doing high-concept and slightly abstract art that requires the textured brushes, there’s no real need for Photoshop. Simpler programs like MyPaint don’t have textured or shaped brushes, but they’re really not essential – and even if you do, Krita is free/open-source and has them. But to be frank, once you play with something as stripped back as MyPaint you might just find all the excess tools Photoshop and other image-manipulation programs come with are just distractions that get in the way. Stripping it back to just the simple tools actually teaches you to paint.

Free, and with an installer under 10Mb – which gives you a lot of bang for your byte.

5) Flip the image

Seriously, mirror the image. There are two reasons for this. One, you can check that it looks okay. If you’re drawing a face, for example, it might look okay and then suddenly, in a mirror image, you scream “holy shit what is that monstrosity!” Drawing inverted refreshes your perspective and lets you check symmetry, and the final product will be much better for it. Secondly, it helps you draw curves better. If you’re, for example, right handed, you’ll draw a curve the way your wrist moves (the top left arc of a circle) easier than the opposite. Flipping the image lets you more easily work on curves. This isn’t entirely as lazy and absurd as it sounds, when drawing on paper people do this instinctively.

I'm spotting a few things I didn't catch first time on the flipped image...

I’m spotting a few things I didn’t catch first time on the flipped image…

6) You don’t really need expensive kit

I’ve basically used entry-level graphics tablets since I started digital painting. There’s really no need for expensive ones – ignore the people saying otherwise. This is especially true if you’re not getting paid daily for this. You need half decent pressure sensitivity and that’s it. You don’t need a large tablet area, your hand-eye coordination and muscle memory will adjust pretty quickly to all sizes, the touch and feel (whether the stylus is more scratchy or slippy) you’ll also adjust to, and you certainly don’t need to donate your life savings to Wacom. These things are tools, and while a better tool might make you feel a little better about yourself (and I won’t deny that I actually like my Bamboo), they won’t actually contribute any significant improvement.

7) Don’t be afraid to trace

Seriously, you’re working digital, you can add layers, just put your source image in the background and scribble over it. Purists might say it’s cheating, but I say there’s no point in wasting your time if it transpires that your actual aim is to replicate what you see.

8) Layers are not always your friend

Layers? Yes or no? To be frank they are a double-edged sword. On the one hand, they’re great for keeping elements separate. On the other, they can get confusing – and if you’re not paying attention you will simply end up drawing into Layer A what you meant to put into Layer B and vice versa, and five minutes of that without noticing will leave you with a horrid mess and you have to go through to clean up. Generally I tend to keep to 3-4 layers these days: a background, just to keep it isolated from the subject as edges get tricky sometimes; the subject, for the same reason; any effects on top; and a sketchy outline on top of all that. Hair I tend to do in multiple layers but that’s for good reasons – see the CG Society tutorial on realistic hair for that.

9) Watch people

If I’m in a room with a strong light source, I have this weird tendency to watch how the light actually moves around the contours of their face. How it goes in a bit around the chin and lips, or over a slight ridge over their eyebrow… I’m not sure if it actually helps, and it might just make me come across as creepy when I’m really just doing research, but it’s certainly worth being aware of exactly what goes on with shadows and highlights and how they contour around people.

File:Low key Nina.jpg

Low-key lighting is good for picking out contours. Understanding that does require getting to grips with the concept of a three-light setup – do some 3D modelling or photography for that.

10) If you want realism, think shadows

The human mind and eye takes into account dozens of separate cues to build up a picture of the world around us and infer their 3D structure. Without that ability, we’d be pretty much incapable of surviving in the wild. As a result, if you want to do anything resembling a “trompe-l’œil” you don’t need pixel-perfect brush strokes nor an atomic level of detail, you need to figure out how things cast shadows and interact with each other. Suppose you’re drawing someone in a baseball cap, it might look staggeringly brilliant, but until you pen in the shadow the brim will cast across their face, there will be something wrong with it – and you might just not be able to figure out what unless you’re looking for it!

What we have here is a series of leather straps holding objects against a wall. Take a close look at where things are stuffed into the straps and how the shadow bows downward a bit where the strap is pulled from the wall. A less subtle version of this point can be found here.

11) Sample carefully

You’re working digitally, and you have an image that you want to use as a reference, clearly the best way to get the right colour is to take your colour picker and… well, yes and no. Certainly, it will be the fastest and most efficient way to get the colour you want to use, but be aware of the following two technical limitations:

  • Compression – if your sample image is a JPEG file, or any file with a “lossy” compression method, the colour you sample might not be the right one. When you zoom out of an image, you might see the right colour, but when you zoom in to the point where you can see individual pixels, you might see a grey-ish one there, a blue one here, a green one there, and then when you zoom out it’s orange. This is perfectly normal for image compression (and also as part of colour-relativity). There are a few ways around this. You could smudge/blend your source image to get rid of the compression and get the colour you want. You could sample more frequently. Or you could use the colour-picker to get an idea of what the colour is. I.e., select it, and drag the picker over the image in question, and watch the coordinates on your colour wheel or colour triangle (or which ever tool you have available) jump about. Then you’ll see what shades and tones are on the image, and can select accordingly.
  • If you have a brush tool that has a non-100% or pressure-sensitive opacity, or a blend function built into it, then the colour you put down will depend on the sample you picked and the existing colour underneath it. If that underlying colour is black or white, then the colour you draw will look washed out (see below). As a result, you might want to pick the colour, then from your colour wheel/tool add a slight and tasteful boost to the saturation of the colour in question.

Also keep in mind that colours are often the result of underlying texture – especially in skin and hair – so any samples you do take won’t look right until you put in the detail/texture. Sorry, there is no way around that, you’ll just have to put in the work!

What looks like one colour far away will look like something different close up - so you probably can't get away with just sampling single pixels with a tool to get the tone you want. It may require actual work.

What looks like one colour far away will look like something different close up – so you probably can’t get away with just sampling single pixels with a tool to get the tone you want. It may require actual work.

12) Don’t highlight with just white

If you’re using a digital brush with a slight blend or blur option built in (a mixer brush), or something with a slightly lower opacity, the way it mixes with the underlying colour becomes quite important. Because the software has to interpolate between RGB or HSV values to mix your new “paint” with the “paint” below it, white will have a strong tendency to just blur out any of the hue. This is especially true if you’re highlighting over a dark object – the dark colour will have low saturation, and the software doesn’t know you want it to highlight straight to, say, red. It’ll just end up treating it as you highlighting something very nearly black with something that is simply very very light grey. So it’ll produce grey. Instead, if you need to, highlight sequentially through progressively lighter colours if you need to highlight from a much darker colour.

13) Just practice

Yeah, really, just do it more. Everyone has a set number of really crap drawings inside them, and the faster you get them out of the way the faster you can get to the good ones.

3 thoughts on “Digital Painting Advice

  1. Pingback: Digital Painting Advice – Lydia S. Padgett

  2. Yep, I’d concur with about 95% of what you say here. Except on tracing – yes, tracing is bloody useful, particularly if you’re aiming for an accurate portrait. But artists should also be able to draw accurately from life as well; if you can only draw from photos then you’re severely limited to the poses available in stock photos (unless you make your own stock images). You’d be stuck doing Greg Land stuff, only hopefully not using porn magazines for source material….

    • I’d like to think that caveat was implied, though.

      It’s more that any “artistic” knee-jerk reaction against it isn’t something worth paying attention to. If your aim doesn’t require pulling a pose out of your own head, skip it. I’m a great believer in efficiency that way – the less time you spend on the stuff you don’t absolutely need to do, the more time you have to focus on less trivial things. And that goes for many different aspects; while it’s nice to get things right first time, I don’t think there’s any shame in using post-production in Photoshop/GIMP/Krita to fix things later on.


Go on, derp away...

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s