My name is Daniel Briggs, I am 22 years old. I am a character artist currently residing in Orange County, CA.
My goal was to create a character that could be rendered in real-time; One that could be used in a video game engine. I modeled a high poly sculpt using Zbrush, topogun, and 3ds max. Which I then converted to a low poly model. The entire character shares 1 UV space and the final textures were 2048*2048. Final renders were done using the Marmoset Engine.
This character is based off of a concept from Diablo 3, I wanted to make a character that stayed true to the concept, and true to the Blizzard art style. I feel like I have accomplished that.
My textures are 100% hand painted. Poly paint played a huge role in this process. Check out my website for higher res images of the textures!
This is my final high poly sculpt. The proportions are much wider than the final low-poly version. I tried to make the character seem really strong and powerful in my sculpt.This is my High Poly sculpt rendered using several different mat caps, and some poly-paint. I comped everything in Photoshop and did some post work as well. This was a test to practice my hand painted textures.
This is really awesome. I’d love to see something like this rigged an animated. My only crit is some of the details on the armor looks a little soft, or the corners could be tighter. But that’s how they are in the concept art, so you can’t really be faulted on that.
The high res sculpt looks fantastic. I also love the polypainting on the highres face. My main criticism is about your low poly version. The detail seems to be very blurred and doesnt look very sharp. Also, the textures seem washed out and im not sure if anything is normal mapped because it looks pretty flat to me. I think your spec map needs a lot of work as well. The spec map should be a lot flatter in shading because you are really trying to separate pieces from each other to pop. Therefore stuff you want really shiny should be really light in value whereas pieces you want matte should be dark in value. you sort of have that going but then you have shading where almost all the pieces have values of light to extreme dark which sort of kills the spec separation between different materials. also, there are a few areas in your diffuse where the highlights are painted in really bright…i realize you are faking lighting but this will work against your spec when you paint it that bright in your diffuse. So, bottom line is i think the diffuse and spec need work to add more details and create a more believable separation between the different material that the character consists of. A lot of this is just experience working with a game engine so keep going!
You sir, are awesome for replying to my work. Thank you!
Thank you BojanStankovski! Comments like these always inspire me to create more!
Thank you neozerahan! That was my intent, I’m glad you liked it.
Thank you for the critique Sarakawa. The detail seems blurred especially in the first render because I used some depth of field in my renders. I tried to go for the diablo art style of having really clean hand-painted textures. I agree with you about the specular map, I plan on rigging and rendering a video of this guy so I will take your observations into account. I will keep going with it thank you for your kind words!
Yes I tried to stay as close to the concept as possible. I feel like I matched every detail in the concept to the model pretty well. Thank you for your comment!
Very cool character! This looks damn near identical to the concept The HP looks very good, in particular the cloth areas feel very accurate. The detail in the texturing is really good also but I do think that Sarakawa made a good point about the specular map. I think you should give that another revision and tweak it, along with maybe pushing the spec values in the shader to make the metals more convincing.
You should head over to Polycount.com, which specializes in game-art critique. People here, in general, may not know much about what makes a good model, but the people at polycount will be able to tear this apart with critique; and it’ll be a lot better as a result. My critique would be that in the concept, the helmet is one piece, rather clearly. And on your model, from the front, it looks awkwardly like 3 pieces, and it just doesn’t look plausible. Also, you say 11,500 polygons, but how many tris, because that is what matters. If your screenshots are mostly quads, I guess that’d be around 23,000 tris, which would be quite a high-poly model, even though it may be appropriate for realtime. I wouldn’t suggest labeling it as “Low-poly” on your website, as a result. Looks good though, just really, head over to polycount.