Well, I dunno how much more I’m going to post given that J seemingly isn’t anymore, but here’s one more for the road.
Well, I dunno how much more I’m going to post given that J seemingly isn’t anymore, but here’s one more for the road.
I’ve tried to write this post a couple times now, and it just hasn’t happened. This time I actually have something to write about!
This painting, which doesn’t have a convenient name since it’s original content, is part of a continued effort to better integrate figure and non-figure and the many layers of background to foreground, all of which has been a perpetual problem with no clear solution. This painting is also somewhat of a return to my roots in that even though the figures are in the foreground, the focus in effort is really on the non-figures in the background, and I think that is a first for a painting with both in it.
I’d been only thinking about it for a long time, but the day finally came when I decided to start using Photoshop, and that was that.
I’m using the “free” CS2, which may be 10 years old, but it’s still newer than both OC1.1 and PSP7 by another three. As my stance continues to be “software only gets worse”, I’m not going to go so far as to say the former is “better” yet, but it certainly has more features both of the latter combined. For the most part, the transition was relatively painless, much like the “transition” to pen – most of the graphic editing features in CS1 work the same way that they did in PSP7; they are just organized differently. I even learned about PSP features that I didn’t know existed (like layer blending modes).
The biggest change for me is switching from the OC brush engine to the PS brush engine. The OC brush engine actually has a couple sliders you can tweak, but the stock brush is pretty good at everything, so in my drive to Keep It Simple(, Stupid), I had for many years only used two brush settings: the stock one and something to do fills. I don’t think I can just use the stock brush settings in PS, but so far I’ve been able to get decent results by only changing brushes and manipulating opacity and leaving everything else alone.
So far I’ve done four paintings in CS2, and two have been pretty good and two have been pretty mediocre. Apparently I still don’t really have this painting thing down, as I can still sink a ton of time into a painting and have it turn out mediocre. Which sucks.
This is one of the two better ones. It’s Shinku from Rozen Maiden. The third season aired last season, and I was somewhat disappointed because while the relatively popular first two seasons were about the dolls, the producers inexplicably decided to make the third season about Jun.
This is the other of the two better ones. It’s Colette from Tales of Symphonia. This painting is actually somewhat of a remake of the same from 2005. I think the 2005 was made in PS as well, albeit 7 instead of CS2.
My continued efforts to find a solution to the cursor accuracy problem was a strong driver for trying a different graphics suite. PS does seem to eliminate the green delta, but the continued existence of the blue delta still makes drawing harder than I think it should be. All four of the paintings I did were digitally inked, but the linearts were still kind of rough, and I still definitely can’t comfortably draw from scratch on a tablet yet. More on this subject in a later post though!
The styles of Shinku and Colette reflect what I now consider my two major “branches” of coloring, something closer to paint and something closer to cell-shade, respectively. Much like how I kit-bashed several models together to make papercraft Titanic, I kit-bashed multiple tutorials to make these paintings. I really like Shinku’s background, which was done in PS from scratch, but I think Colette’s figure turned out better. Colette also has some places where I tried to color the lineart, which is something I’ve wanted to try for a while, but is apparently more difficult to do well than I had imagined.
If there’s anything I’ve learned (or relearned) in this process, it’s that everything still comes down to having a fantastic lineart, in which case the coloring doesn’t need to be spectacular. As such I’m going back to just drawing for a bit after this.
Well, as it goes: another month, another post…
I haven’t been quite as prolific with paintings in 2013 as I was in 2012, but I’ve been turning out a good number and variety nonetheless. I say time and time again that my constant struggle is one to simplify tooling and process, to make the finished product with as few tools and as few steps as possible. For the past couple years this has been a theme in my life outside of drawing as well as I’ve continuously tried to get rid of junk that I don’t need.
Anyway, this quest has led me to keep revisiting tried methods of painting, which is how this first one is done.
Pure pencil linearts have been done time and time again with varying degrees of success. Even this particular flavor of pure pencil lineart – more or less a sketch with little to no touchup – has been done enough for various Tales group paintings that you can see the evolution (The years are 2006, 2007, 2012, and 2013 respectively)! Overall I think I definitely did better on this than on the Graces 2012 painting with the same process, but it’s one of those cases where I can’t quite put my finger on why.
One of the reasons I want to simplify drawing process – other than for the sake of simplifying – is that I want to be able to draw detailed backgrounds without having to draw them twice. This means that I don’t want to have to draw it in pencil and then go over in ink, nor do I want to draw it in pencil and go over it again in digital ink. Going all pencil is one direction I could go, but with pencil I have that whole smearing problem, and increasingly more relevant is my increasingly passionate hatred of eraser crumbs.
Going all digital is a little more interesting because even after many false starts (almost as many as pen), I am still convinced that it is ultimately the way forward. That being said, I am increasingly convinced that the false starts are from beginning to hit the limits of my tooling, and that goes for things other than cursor accuracy now. My latest gripes are not being able to zoom out more than a certain amount in OpenCanvas and how the program renders everything without anti-aliasing at any zoom level other than full size. The former obviously makes it very difficult for me to look at a drawing in it’s entirety, which is especially important if I’m going to draw stuff digitally from scratch, and the latter makes it very difficult for me to visualize actual line quality when zoomed.
Of course, the accuracy problem is still more annoying especially for drawing, which is part of the reason I keep trying to let paint do the work. This next painting follows the same theory, but I am also trying to do some better background work as per my last post. Initially I intended to digital ink the figures and remove the background lineart, but I didn’t.
I thought the lighting here would be difficult to do, but I just did it like element by element, little by little, and I think it turned out pretty well. I did a tiny bit of erasing/touchup on the single-weight pencil sketch, but I didn’t do any color adjustments on the paint – for the most part I think I’m becoming decent at getting the colors that I want the first time around.
This painting was actually done back in July. It has a couple “first in a while” elements: the first Chikara painting since 2007, the first digital lineart since 2010, and the first cell-shade painting since, well, last year. Also of note is a from scratch, no lineart background that isn’t just a tree. Or a shadow. Or a rectangle.
Non-figure paintings are generally pretty few and far between, but I think I may be on the verge of changing that. Non-fanart paintings are also in the same boat. The GT1 is a car I designed for a project prompt that J proposed (more on that in the future if it goes through). It’s called the GT1 as I took design queues from the Pennsylvania Railroad’s GG1 and T1 locomotives. I didn’t intend to do a painting from the start, but this test has been in the pipeline and the GT1 became an appropriate subject at time.
The GT1 painting is new in that it is a “custom” painting with no final lineart. Like the ink comic from last time this painting was basically an inevitability since London Waterloo, which was my first actual painting without any final lineart despite being a copy. The roots go a little further back to the Pieces painting, but I definitely wasn’t ready for this back then. Nonetheless, as with ink drawings, I don’t think anything has fundamentally changed about my drawing; apparently I just needed to do these things to get there.
The process is the same as that of the Pieces painting: I paint in a sketch and remove the sketch afterward. Unfortunately it’s a mediocre painting that might still look better with the pencil sketch, and it’s made worse by the really crappy no-lineart figure. Like it’s really no better than just removing the lineart from the Estelle 2013 painting, and in that painting I never intended to remove the lineart. The one bright spot here is the shrubbery, which is done completely from scratch and doesn’t look half bad. Then again, I’d already done something similar for the tree in the ToG crossover painting from last year. So I’m becoming more convinced that no-lineart painting should be relegated to backgrounds, but I’ll hold that thought for a while longer.
While I was reading through old posts, I realized I hadn’t talked about computers or computing for what is probably a record long time now. I think the glut of project-specific posts over the last half a year have focused my attention and prevented me from letting loose and rambling, so I’ll take the liberty to do that now.
The fleet is actually down to just two laptops now, the x220 tablet and the t430. This is a low that hasn’t been reached since I got my D830 way back in 2007: I recently sold my second MBA because I still don’t have a use for it… just like the first time around! When I bought it I was still in school and thus taking it to class, and for that it was fantastic, but I now I no longer travel enough without wanting to have my tablet, so the niche was apparently gone a long time ago. Maybe this time I will learn.
I’m more and more dissatisfied with the x220 tablet as time goes on. I’ve been really wanting a thinner and lighter traditional convertible tablet with longer battery life and a 1080p screen since forever, but in this day and age of shitty “pure” tablets, no manufacturer is willing to give me what I want, so the x220 tablet stays. I’ve been considering getting a ThinkPad Helix because it is basically the closest thing out there to what I actually want, but the Helix is stupidly expensive (compared to an x220 or x230 tablet) because it is full of stupid features that I don’t want. So that sucks.
The t430 does what it needs to do as a desktop replacement. I’m generally satisfied with it. They even released a BIOS update that fixed some of the issues in the fan profile… wow!
On the other hand USB3 has a complete and utter shitshow. Back in 2011, I defined my current storage solution and bought those USB3 peripherals looking forward to the day that I would finally be freed from the limits of USB2. Well, first I found out that the ExpressCard USB3 adapter was a piece of crap, and after moving to the t430, I found that the USB3 to SATA adapters were also a piece of crap. Thus, I am still using old and slow USB2 hardware to do all my external file transfers. I can’t even replace my USB2 external hard drive because I can’t find a USB3 drive or enclosure that powers down when idling, and I am now super fond of this feature. So that sucks too.
My setup actually hasn’t changed for more than a year now… I think the only thing I’ve done is run a USB extension cable behind the monitors such that I have an easier way to connect USB stuff to the t430. I have considered replacing my speaker set as it occasionally seems to have issues, but I’m not really going to do it unless someone has extra speakers or the current ones die completely. They are making a new version of my lovely Corsair keyboard with backlighting and clicky switches, so I’m definitely going to buy one of those when it comes out. They claim in July but you never know with these things.
I even went back to the same desktop backgrounds. For a long time in high school and college I was changing desktop backgrounds two or three times a year, but somewhere along the way I got tired of that. I think my standards eventually increased to the point where I couldn’t find anything that sufficed, and that was that. I will force myself to find a higher-res wallpaper if I get a Helix though.
Well, I lied. I said there’d be another Lego post this time, but the hoverbarge I was building kind of got stuck, and it isn’t done yet.
For a lot of paintings I often have something in mind some time before execution, but this one kind of rushed its way into my queue. This is actual the fourth in what is now a fairly long string of Rita and Estelle paintings (the first one being 2008, most of which have unfortunately been rubbish. 2013 deviates from the “paint-heavy” direction I was going with Rita 2013 and Estelle 2013, and takes another stab at “hybrid cell-shading”.
Her full name is like “Estellise Sidos Heurassein”. I can never help but read that last part as Hussein. As in Saddam.
Estelle 2013 is more or less a linear development of the Rita 2013 painting and more or less a remake of the same from 2009. One of the concerns over Rita was whether or not the more fluid drawing and painting style would “scale” well to more static content and lighter colors, and Estelle tries to move in that direction. Another thought was whether or not a more fluid (and this is really a boilerplate word for the painting: I’m not actually sure how different it is) painting style would scale well to a more conventionally scaled and finished (ie inked) lineart, and there was an intermediate painting that tried to address that.
It was pretty decent, but obviously not okay enough for me to make a post (maybe more in that direction later).
I don’t really get much time to read these days, but I’m slowly working my way through Nassim Taleb’s Antifragile. In this book Taleb describes systems which he classifies as “fragile”, “robust”, or “anti-fragile”, depending on how well they react to the chaos of the world. This post’s painting is tenuously related. I’ll quote Wikipedia for a brief synopsis:
“Some things benefit from shocks; they thrive and grow when exposed to volatility, randomness, disorder, and stressors and love adventure, risk, and uncertainty. Yet, in spite of the ubiquity of the phenomenon, there is no word for the exact opposite of fragile. Let us call it antifragile. Antifragility is beyond resilience or robustness. The resilient resists shocks and stays the same; the antifragile gets better”.
I described a drawing “theory” way back in d266 stating that the big picture is more important than the small picture: if your entire image is about “right” it won’t be affected too much by individual “wrong” lines, which is suggestive of a robust system in Taleb’s world. I guess the motivation behind the following painting is to push that boundary.
Rita is the most broken character in Tales of Vesperia as she can do infinite combos by herself (admittedly by virtue of the 360 version being something of a beta). She actually has a number of paintings in the past, but they are mostly all rubbish.
Well this post is labelled for 2012 because the painting was supposed to be done last year. It’s labelled “R2” because of this rubbish painting from June.
Back in January I told J something like “I’m going to make three paintings that I actually like this year”. Had this actually been completed last year, it would have been the 13th out of 18. In 2011, it was basically 0 out of 5. In fact the total painting count for 2012 (and this is not counting “full” experimental and “sketch” paintings) was just 3 shy of my all-time high of 20 in 2006, which still surprises me considering the long-term trend.
Despite this achievement the biggest deal in drawing for 2012 was actually finally switching to pen. For a while I was thinking that it would never happen, especially given the numerous false starts I mentioned in d356. That being said, it’s been a couple months now, and I think I can safely say this isn’t a false start. For sure ink has solved more problems than it introduced, and some parts of this painting are definitely a product of the switch.
So let me list some of the “features”, which are new to this year:
Slightly hybrid “gritty” background. The “slightly hybrid” part refers to the fact that it’s not entirely detailed in pen. A couple small sections are paint-only, as seen in the raw lineart. “gritty” is J’s term, but to me it refers to not trying to be exact about the drawing and just going for the suggestion. It’s like the whole implicit/explicit thing I keep ranting about for painting.
Pen-tool based paint process. I talked about this in d358, and I even had those test paintings, but this is the first time I’m implementing it at scale in a traditional painting (as opposed to a from-reference a la last update). These are not strictly done with the pen tool by any stretch of the imagination, but using it really helps with precision in form and in color.
While I think the figures are actually pretty good, I’m still more happy with the background in this. It’s the first nice, non-abstract painting background that I’ve had for a really long time. Both the pen lineart and pen painting contributed immensely. I think I’m also starting to get a grip on painted clouds, but we’ll see with a few more samples. Copying stuff – or even modifying from reference – really helps in everything as well.
Overall I guess there is still a ton of volatility in painting quality. The rubbish June painting came right after the quite decent Philia painting; it was done basically the same way, so I don’t really understand why I think it’s that bad. I was trying to investigate a little by using the same characters and palette here, but I’m not really sure it helped.