We decided pretty early in Mask of the Rose pre-production that we wanted the characters to appear solid and painterly. There were three main reasons for this:
-We wanted to distinguish the look of the game from other visual novels, which tend towards line drawing and flat shading.
-We were aiming for a visual style that felt like an evolution of Fallen London, which usually favours a combination of painterly rendering and mild exaggeration for its characters:
-Honestly, it’s just a style that I really enjoy drawing, and If you’re going to spend a year or more making something, you might as well have a good time.
This decision has created some interesting challenges. Visual novel characters have to be expressive and malleable, which means multiple poses and expressions. They have to make visual sense against multiple backgrounds. And, because you never really know how they will function until you get them into a scene, they have to be straightforwardly adjustable. All of these things are, of course, much easier and faster to achieve with line drawings and flat colour, but we were determined to make the rendered route work.
I’m going to use Griz to illustrate some of our approaches to these problems – she was the first character I designed and has served as a test subject for most of our experiments.
The first problem was how to change poses. At first we experimented with a static torso and a range of arm positions, but the results were stiff and unconvincing – people don’t move their arms in isolation, after all.
Full body movement is better at conveying emotion, but we couldn’t commit to drawing anything up to a dozen full body poses for each character, and it would potentially have looked strange anyway; large pose and expression changes in quick succession can come across less as “acting” and more as an interpretative dance.
So what we’ve settled on – for the moment – is a small set of full-body poses that can convey several attitudes each, supported by a larger array of facial expressions. So for example, the hands-on-hips pose can convey both surprise and aggression, depending on which head you attach:
Those facial expressions, meanwhile, are constructed from a much larger range of eye, brow and mouth assets, all carefully blended onto a single head model, which in turn tilts a bit to sell various attitudes.
Finally, let’s look at how we get Griz to blend nicely with a scene background. This is largely achieved through lighting. Unity has some excellent tools for lighting 2D sprites, and we’re making extensive use of them. First off we give a general colour treatment to whole scene, using global light settings, post processing effects and some fog particles to tie it together
Next, we light the character sprites to match the scenes, much as you would on a film or stage set. Note that in the images below Griz’s left (our right!) is now blueish to match the cool tones on that side of the background.
The great thing about this lighting is that it’s all dynamic, rather than drawn onto the characters. That means we can change it over the course of a scene to highlight important moments, apply the same lighting to multiple characters to blend them into a scene, and animate it if the situation demands:
As usual, this is all *very much* work in progress!
I expect we’ll do more iteration on the characters as we go along, because there are a lot of them, and we’ll keep finding new and better ways to represent them on screen.