Samurai Sculpt: What I’ve Learned So Far.

To date, there are sixteen live stream replays of the samurai sculpt up on the nickelcitypixels youtube channel.

Each video is about two hours, meaning that 32 hours of sculpting has been recorded. Estimating as much time was spent sculpting off-stream means that, to date, 64 hours has been dedicated to creating the samurai character. Here’s what I’ve learned so far.

I should’ve studied the subject more before I began. Samurai armor is complicated. I knew that going in, but I overestimated what I could learn from a handful of image search results. A more thorough search eventually elicited diagrams that revealed what was beneath the armor and how it was layered. Reading on the subject elicited additional information that couldn’t be gleaned from pictures. This information included the materials and processes used in armor creation and helped inform my own artistic process. It’s worth mentioning too that the illustrations and reference photos that accompanied the text were excellent. Perhaps surprisingly, searching text descriptions often revealed better image references than image searches!

Other people’s sculpts weren’t of much use. It seems reasonable to assume that the best reference would be three-dimensional. With a 3D model you can observe the subject from any angle: above, below, and behind. You can zoom in without loss of resolution and move around and beyond the obstructions typical of 2D imagery. There are plenty of 3D samurai models on sites like sketchfab.com. Unfortunately, most of the artists there should have studied the subject more before they began just as I should have. Ultimately though it seemed as if most of the sketchfab sculptors didn’t do any homework! Zooming in on the samurai models and moving around and beyond the typical obstructions, (say for example shoulder plates) didn’t reveal any answers. More than once the revelation was that the artist just left pieces floating. In all honesty, it’s been my experience that there simply isn’t enough publicly available reference to wholly replicate the in-real-life construction of samurai armor. But there’s a right way and a wrong way to go about cheating. And the right way is to have the 3D model, by and large, seemingly obey the laws of physics. Cheats should be obfuscated from any and every angle.

ABC: Always Be Cheating. Speaking of cheating the right way, I bought a bunch of samurai do-dads off of the artstation.com marketplace. None of these items were anything that I couldn’t create myself; the kits included various threads, knots, rivets, and plates. But because the most expensive kit was just $6, I couldn’t create those things in less time/money. And since it doesn’t make sense to spend more time/money than you have to, I c̶h̶e̶a̶t̶e̶d̶ did the reasonable thing. I bought and used the appropriate zTools for the job.

Do everything right the first time. This isn’t always possible. My samurai sculpt in particular is complicated by not having a fully-realized concept piece. As a result, I don’t always have a complete vision and I’m not always able to realize the incomplete vision completely. It’s two steps forward, one step back. But there were problems that I could have foreseen if I’d just put my mind’s eye to it. Take, for example, the armored sleeves. Whether or not I eventually decided to make a traditional sleeve or a sleeve made mostly of fantastical flourishes, I should’ve been able to anticipate its coverage. Or at the very least erred on the side of caution by overextending the coverage. Which I still haven’t done and still needs correcting. The lesson here is to take the time to work out the logistics before you make creative commitments.

If you can’t do it right the first time, come to grips with the time it’ll take. The last sculpt I did (THREE YEARS AGO!) took only twenty hours. In that instance I was able to obtain excellent reference live and in person and I only needed to sculpt the character’s torso. In comparison, I’ve spent three times as long on this sculpt and there isn’t a clear end in sight. It is clear to me, however, that spending only 2 hours a day will doom me to half-a-year’s worth of work. So, I’ve created a sculpting to-do list with approximations of how long it will take to do each thing. And understanding that efficiencies can be obtained by working with longer chunks of time, I’m ready to commit more hours each day. The two hours I had been spending is too stop-and-go; by the time you’re just getting warmed up you’re getting ready to leave. To ensure proper time management I’ll likely enact a variation on my index card strategy:

Moreover I’ve come to grips with the fact that since I’m not an artist, at least a quarter of the time will be spent learning tech and technique. Technique here refers to the spoils of artistic experimentation and tech learning is all about figuring which zbrush buttons to press.

Pro tip one: Save early and often! I’m getting more and more comfortable with a non-linear workflow. I save each time I make even minor progress. As a result I feel more comfortable experimenting with tech and technique knowing that I can always salvage an un-ruined zTool from a former file.

Pro tip two: Keep notes! Another thing I’ve found helpful is to keep notes. Among other things, I keep track of the brush sizes I apply to curve inserts. For example, I might want to have the same sized rope in multiple places. Visual continuity is good. Rather than getting all OCD as to whether or not the rope I’m creating is the same size as the previous one, I can simply consult my notes and dial in what I know is the right number. I also keep track of extraction numbers. Furthermore, certain insert meshes will have multiple variants; noting which you use where is a good idea.

These are the lessons I’ve learned from just over 60 hours of work on the samurai sculpt. Join me during the next live stream as I (hopefully) continue to live and learn, ultimately to create the video game magna opus 10,000 Dead Samurai!

Leave Comment

Your email address will not be published. Required fields are marked *