My second Tales of Graces painting features Cheria, who did not make it into the last one, primarily due to space constraints. The background is a hand-traced portion of the ToG logo. I don’t really believe there is anything particularly noteworthy about this painting except that the lineart is inked.
This painting has looked a little weird to me ever since I finished it, but I can’t really put my finger on why that is. It’s probably not the inked lineart, and some combination of colors, blending, background, etc, but it just doesn’t look like what I think my normal painting should look like. And then J asked me what [i]was[/i] normal these days.
Now that thought kind of piqued my curiosity because even disregarding the recent uptick in presumably significantly different test paintings and methodologies, the standard deviation of my paint quality has traditionally been much higher than that of my “regular” drawing quality. In comics, my methodology and quality really leveled out starting in 2010 (according to d266), but painting and “finished drawing” methodology and quality have been anything but consistent.
The main issue is that I obviously just haven’t found a “good” way to paint, or more fundamentally, color a drawing. In general, I’ve been just kind of throwing stuff out there to see what works and what doesn’t work. This accounts for some of the variance in paint quality, but certainly only some; there are a couple development “branches” going in parallel, and even within the branches the variance is high.
For example, the last really long running branch was “thin-lined digital lineart”, which effectively terminated in 2010:
I started using this “inking” method way back in 2007; in d160 I talk about experimenting with different drawing and paint methods as well, but development generally ran in series back then. Considering the number of times I switched method(s) back then, thin-lined digital lineart lasted for a really long time, during which it looks like I primarily worked on the actual painting part of each piece; I talk about reducing “splotchiness” in d346.
When I decided that lineart was more important to a painting than the actual painting itself, I gave up on digital inking because it was adding that extra layer of noise to each lineart. Conceivably there will be a revival sometime.
And so I went to “polished, pencil lineart”, which failed spectacularly after just a few paintings, the first of which was the very good Misaka 2010 and the last of which was the very not good Akemi 2011. Recently, the Tales of Graces PS3 2012 painting continues this development branch, though the lineart is nowhere near as “cleaned up” as its predecessors’.
And so development switched to what is effectively “clean sketch as lineart”, a lineart method in name, but which actually tried to affect painting by defining shaded areas at drawing time. This method actually produced a fair number of moderate and strong successes, some of which aren’t formal paintings and some of which aren’t posted yet. This method is technically what I would call “normal” for now as it has been going on until just recently and with some regularity:
One of the branches occurring in parallel in limited capacity is the “minimal lineart landscape” branch:
This is self explanatory, but just very difficult for me to do, hence the limited progress.
There is also the colored pencil branch, and the now resurfacing ink-lineart branch, but more on ink linearts in the next update.
For the most part, after putting together these little sequences, my conclusion is that, when it comes to painting, actual drawing quality (like at draw time, not necessarily ink or polish time) seems to vary much more than drawing quality in sketch or comic. Why this happens I am still unsure. This variance adds up when I have more layers of abstraction between the original drawing and the finished work (digital ink being further away than sketch as lineart), and I think this is why the hit rate for the sketch-as-lineart method was fairly high.
But that’s it for now.