After compositing my renders, I decided to create two final images. One basic beauty shot, with just a plain black background and no blurring or grain filters, intended to display the model and textures without any distractions.
The second final image is the composition I was working towards according to my concept. It features the background from earlier, as well as more stylized saturated colours and a film grain filter.
I am pretty satisfied with my image and the work I put in. I think it turned out relatively well, and that I improved upon many aspects of 3D work that I would consider weaknesses of mine. Like texturing, lighting and rendering.
Texturing is really mostly a matter of putting in the time to work on it, as opposed to trying to get it done as quickly and painlessly as possible. I’d say the texturing I did on this project isn’t great(the character is decent, the environment and the demonic arms aren’t so good), but quite a bit better than I have done previously. It’s definitely something for me to build on.
I’ve always thought that texturing is all about photo-manipulation, which I have had no trouble with in regards to environments, but I am still not sure how to effectively work that into characters, as their UV’s are too complicated to just slap a sampled image on them.
I spent quite some time researching lighting and rendering. I read Essential Lighting Techniques With 3DS Max by Darren Brooker, which helped me out a lot. I think my understanding of light and shadow improved immensely during this project.
I am particularly happy about learning more about render passes and compositing. I really love the amount of control I have over my final images with this approach, and it’s definitely something I’m going to use in my workflow from now on.
The UVW issue I ran into was quite regrettable. Fortunately I was able to resolve it and move on, but I was stuck for over 3 weeks on that one, trying to work around it as best I could while looking for a solution. As much as I love Zbrush, I find that in my pipeline, the transition back and forth between Max and Zbrush is the place I most often run into problems that destroy both my schedule and my momentum. It seems every time I feel like I am aware of, and taking care to avoid, all potential problems, a new one pops up to get me.
That said, aside from the UVW issue, I think my workflow was pretty effective this project. I managed that time well and had no trouble with the deadline.
I got off to a good start by spending some time properly visualising the image with my concepts and mockups before starting 3D work proper. It was nice to work with a clear objective in mind rather than making it up as I went along. The modeling was quick and painless, and I had it done by the beginning of December, leaving me plenty of time to work on my textures and read about lighting and rendering.
I’m happy with the resulting image. I believe I got a decent result, considering I decided to go for low poly with normal maps, as opposed to high poly with displacement maps. That is a character that could conceivably be put directly into a game engine.
I am a games artist and I wanted to work somewhat within the confines of game art, so I decided to go relatively low poly and rely on my normal maps and standard shaders. Kind of challenge myself to see what I could do with it, so the characters and environment in the image is about 24000 polygons in total.
As for the lighting, I wanted to set it as a dark sunset shot, so I chose to have one key light set from the top to simulate the sky, with a high key to filler ratio to create a bit of contrast and shadows. I also put about 6 filler lights with a blue shade to soften the shadows a bit and light up the places that were too dark, while still keeping the illusion of darkness.
I still thought it didn’t look very nice. The lighting seemed off somehow, and the background seemed a bit “busy” and didn’t blend well with the character in the foreground. I decided to remove the background and just go for a plain black background, and then create a composed image in Photoshop later where I would include the background. I also put a volume light on the key light in order to create a bit of a glow around the character.
I decided to try my hand at render compositing. Rather than spend a day rendering out one image, I rendered one simple beauty shot, and then used the scanline render elements rollout to create a lighting, specular, z-depth, and diffuse pass. I also did an ambient occlusion render pass with mental ray.
I then took all my render passes to photoshop and composited them. The ambient occlusion render pass in particular was quite the eye opener. It added a lot of depth to my image that made all the difference in the world. With a little work with the dodge tool in photoshop, the specular pass was really useful as well.
For the specular and lighting passes, I played with the hue a little bit, to change them into more of an orange-y colour that would make the light and highlights warmer, to match my background image that was more of a sunset image than moonlight.
I created a background image for my character. I didn’t want to spend a lot of time on the background, so I found a nice nature image of the horizon and sky I liked, and used it as a base for a paintover to create an environment that fit with my intended style.
It turned out quite well, however I wasn’t sure if I was going to use it, as a lot of character pictures actually work better with a plain black background and a spotlight, with maybe some volume lighting. Backgrounds also may take the attention away from the character if you’re not careful.
I ran into a rather large problem during my UVW mapping. After finishing the UV’s for all my meshes, I exported them back to zbrush, only to find that the UV’s were broken. When I tried to render out my normal maps, I was met with this issue:
As you can see, there is a lot of faceting going on there. On a properly working UV mapped model, this would be completely smooth, only broken up by the seams I had put in place.
Now this faceting did not cause any issues with textures created in photoshop, but they completely broke any maps generated from zbrush. Naturally, this turn of events was fairly catastrophic.
This was a new problem I had never seen before, but I had recently upgraded from Zbrush 3.5 to Zbrush4 and from Max 2009 to Max 2010, so I assumed the problem would be caused by one of those.
I did a lot of research on the problem, and located forum threads where people reported similar issues on sites like Zbrushcentral, but none of these people had found any solutions, so it was up to me to figure it out for myself.
After a lot of testing and changing of software, I managed to solve the problem by rolling back to Max 2009, merging in the assets from the Max 2010 file, clearing all UVW information and redoing the UVW mapping, then exporting with Max 2009. Once in Zbrush4, instead of replacing the subdivision1 model with the UVW mapped base mesh, I imported the new meshes separately and projected the model information on the new meshes. That eventually solved my UVW faceting issue.
Unfortunately, this whole process delayed me by almost 3 weeks.
I never found out exactly what caused the issue or where it occurred, but as far as I can tell, the problem has something to do with the export process from Max to Zbrush. I found a way that works, and I will stick to that from now. I’ve also changed my workflow to include regular UV tests when I start the UVW mapping phase from now on, so I can catch any potential issues like this much earlier.
For my diffuse texturing, I created base textures in Zbrush using polypainting, and further refined them in photoshop.
I played with the idea of using mental ray materials and relying on their properties for armor pieces and skin. However I decided against it based on the fact that I am a games artist, and I want to work within the confines of what I would usually do for a game. So I decided that I would rely on diffuse, specular and normal maps for my texturing.
I decided to do the pose in Zbrush with the transpose tool, as it’s pretty easy to work with, and it saves me having to go back and forth between max and zbrush yet again.
For the pose, I wanted a strong pose, that shows tension and resistance. As I imagined it, Spartacus would be restrained by these arms reaching out of the dark and grabbing him, and he would be fighting it, trying to break loose. So the challenge would be to create that feeling of tension in the pose.
Thanks to my mockups, pose tests and concept work, I had a very good idea of what my pose would be before I started using the transpose tool on my character model in zbrush. So finding the basic pose was very quick, however finding the tension required quite some time of subtle tweaking. A small bend in the knees, change the angle of the arms and so on.
Once I was satisfied with the pose, I went back to sculpting and changed his facial expression, changed some of the muscles, so they would stretch or contract appropriately, as well as adding some skin wrinkles caused by bending joints.
After I finished the pose and sculpt of the Spartacus character, I created his weapons. I gave him a fairly simple gladius and shield with a snake motif. I also made a shield and some spears that would be displayed on the base.
After seeing many battles, the weapons are pretty beat up and scratched. For the skull and snake motiv on Spartacus’ shield, as well as the designs on the gladius guard, I used alpha masking in Zbrush with the inflate feature in the deformation rollout.
After finishing the weapons, I got started on the environment. I wanted to do it pretty simple, a pile base that would consist of the pile of skulls, and the arms reaching out of the base.
At first, I decided to try sculpting the skulls into the base manually, one by one. After getting pretty far into the process, I decided to scrap it, as it got kind of messy and required several retopology sessions to get the proper topology.
Instead, I decided to sculpt the base as a living creature rather than a pile of inanimate skulls, twisting and writhing, with outlines of skulls pressing against the skin from inside, trying to escape. In addition, I added some skulls and skeletal parts as separate meshes on top, where the character stands. This approach was much more manageable and time efficient. Had I gone with this from the beginning, I could have saved myself a bit of work.
I wanted the arms to be strong, big and powerful, yet decaying. So rather than going for skeletal arms, or thin tentacle type arms, I looked at some artwork of Venom from Spider-Man. Venom is large and muscular, but his body composition typically stretches and morphs as he moves. I wanted the arms to kind of do that.
Additionally, I wanted them to have a decaying look. I decided to sculpt in torn skin and open wounds with rotting flesh in the arms, to both achieve a more decaying look and make them look that much more grotesque.
I have pretty much finished modeling on the character. I’ve done everything except the fine details such as skin pores, scars, fabric patterns etc. I’m holding off on that for two reasons.
Once I pose the character there will be some resculpting required. Fixing deformation caused by the posing process, as well as changes to the body based on the pose, such as muscles stretching and contracting.
The second reason is because I want to do the UV’s first. Which I am going to do after posing, as opposed to when he is in a T pose, because there may be stretching caused by the pose. During animation this would usually not matter because no one will notice, but in a still image like this, it may be very apparent. So I am posing the character first, then doing the UV’s. And then I am adding the fabric details etc, because in many cases, just using a texture as an alpha for creating fabric patters will let you create things in moments that would otherwise take hours upon hours to sculpt manually.
The hair is not in place yet. It will be done with alpha mapped planes after posing.
I have pretty much finished creating and adding detail work to the weapons to be featured in the scene.
I have also started work on the base for the character. I’m doing it as a typical model base, so it could be suitable for 3d printing as well as a beauty shot.
Originally I was planning to repurpose an old skeleton mesh I had created a few years ago for the base, just use the skull and various skeleton parts and copy it out to form the base. This actually worked pretty well, but ended up being very polygon heavy as I used over 50 skulls, and left me in a UVW hell. I could either just UV map one skull and copy it out, making all the skulls look the same, or I could make them all part of the same object and map every one of them. Neither option was very attractive.
So I chose to just sculpt the whole thing. I created basically just a big lump of clay, and decided to use the skulls I had copied around as a general guideline for my sculpt.
Rather than just a pile of skulls, I have decided to make the base more of an organic being. Kind of a character in itself. It’s going to be a large blob with skulls and heads and arms writhing under and poking through it’s “skin”.
Best reference I can think of off the top of my head is Envy from Full Metal Alchemist.
I must admit I’m kind of wondering what I have gotten myself into with this base. Just blocking in all the skulls took forever. Thank the gods for the Decimation Master plugin for Zbrush. Every time I build up a skull, the polygons get stretched, but then I can just decimate and have a lower res mesh that keeps most of, if not all the detail, with a nice and even spread of polys. Saves me a lot of time that otherwise would have been spent retopologizing. In any case, I didn’t want to spend too much time on the base, but this looks like a bit of a project. But hopefully I’ll get a nice result in the end.
I started modeling the character. I’m not doing a lot of modeling in Max, since for this project I’m just doing an image, and polycount is not an issue, so I will be importing a higher subdivision level into max for the final render. This means that I can just focus on creating the different mesh pieces with enough geometry in right places for the sculpt. The actual sculpting phase will create all the detail geometry for me, which I will then import back. No need to spend time fiddling around in max to model it there.
The model ended up being about 7400 polygons unoptimized, didn’t delete the insides of eqipment etc. I never do this until after the sculpt is done, because I believe it is important that everything has thickness in Zbrush. Usually I would do the sculpt and the UV’s, then optimize the mesh before rigging. No need to optimize for this image though.
I’ve finished building the base mesh for the character itself. Didn’t start weapons and environment just yet. Will finish basic detail sculpting and preliminary posing of the character before I do that.
I’m not too far into the sculpting phase yet. I did most of the body sculpt (didn’t add small details yet, skin pores, wrinkles, scars etc), and I’m about to start the equipment.
Quick note about the environment. I’m thinking I’m just going to model the pile of bones he is standing on, not the background. The background can just be a plane with an image on it. I’ll do the pile of bones like a base, so it’s suitable for a turnaround, or even a 3d print.