After watching a series of videos on CGCookie.com in which the expert artist took concept art of a sci-fi pistol and modeled the living hell out of it, I said that yeah, OK…I understood everything he did, why he did it, and had even managed to predict his decisions a few times. I felt I was ready to take on a similar project. And for my first project, I choose…this:
That went about as well as you can imagine. I managed to block out part of the grip (a few times) which resulted in some weird Blender acrobatics going down, and decided that maybe I should tone my subject down a bit. I figured what could be more “toned down” than a sword? So I found this reference image and went to town:
Oh boy. Who knew that a blade on a stick would be such a pain in the ass? I’m sure that experience artists could whip this out in less than 30 minutes, but I went to bed and dreamt about modeling this moronic thing, which should be an indicator of how long I’d been focusing on it yesterday afternoon.
All in all, I guess it’s got some good technique points. There’s the tip, which is rounded here (for today’s safety-conscious sword buyer, I suppose) and tapered. There’s the fuller, aka the “blood groove” which I learned has nothing to do with blood but everything to do with reducing the weight of the blade. The hand-guard is craptastically designed as a chunk of metal that’s been poorly bent and is therefor stupidly hard to manage for a novice who wants to model a great looking weapon. The hilt is basically a tapered cylinder, and the pommel a sideways cylinder with some personality. On the face of it all, a sword seems like a pretty simple task: extrude a cube and taper it for a blade, and inset the fuller. Extrude, rotate, and scale the hand-guard. Elongate a cylinder, scale it down, and taper the handle. Scale another cylinder, insert and extrude a few times for the pommel. Easy!
Well, the blade was a pain because a cube doesn’t round well without loop cuts, and adding those — while also helpful for the fuller — caused headaches for the blade overall. Extruding, rotating, and scaling the hand-guard just looked like ass. The handle ultimately ended up too wide (which I didn’t notice until the render was done), and the pommel was…kind of OK. Nothing spectacular, but I don’t expect that a workhorse Medieval longsword was big on fancy adornments anyway.
I spent some time trying to understand UV unwrapping. I marked seams, moved and scaled and packed islands for the “proper texel density”, but that made everything worse. So, back to “smart UV unwrap I went”.
After wrestling with a few normals issues, I got a working version into Substance Painter and immediately proceeded to violate Substance Painter Mandate Number 001: don’t just drag stuff from the shelf and call it done. I know it’s not done, but I wanted to see what a rendered version could look like, so I used some battered steel, some cloth, threw a HDRI in there that I thought might provide some decent lighting, and waited for IRay to finish it’s grinding:
And there we have it. My first free-form model. It’s not anything to write home about (although I did write a blog post about it), and I believe that more than anything I just needed something that I understood in Substance Painter because my 30 day trial is running out and I have to decide if this [gestures widely at 3D modeling] is something I want to continue with…can continue with…so I can maybe consider thinking about possibly taking on the Substance suite subscription.