Incredible, I love this so much. I'm glad someone out there decides to go make older games like this in newer engines. Great work!
Firstly,Amazing work !! But a doubt..for the background thing ..you mentioned of using a mesh with zero edges which helps out in covering up the repetition process..what is it?..any detailed description please.
Quite fantastic. I am a friend of Grayson Wixom and have an entertainment publication thehollywoodtimes.today and am trying to get one of my journalists to interview you.
Nikolay Sladkov broke down his awesome Barbarian character created with 3ds Max & ZBrush: working on hair & fur, sculpting texture details, animating.
Hi everyone! My name is Nikolay Sladkov, I am from Russia. I was born in Irkutsk and spent my youth there. It is a central part of Eastern Siberia. This land is beautiful but harsh: the winter is long and cold, while the summer is short and hot.
Today I can describe myself as a 3D Generalist with a narrow specialization in character creation. My career in the game development industry started in 2010 when I moved to St.Petersburg and began working for my first gamedev company Saber Interactive as a Texture Artist. I participated in such projects as Inversion and Battle: Los Angeles. Then I worked for 3 years in PlayKot as a 3D Artist, took part in all open projects. I was responsible for a full cycle of 3D production: from modeling and texturing to animation. That work was easier than the previous one, plus I obtained a solid understanding of the full process. After that, I moved to Moscow where I had been working on a few unannounced projects at Game Insight slightly less than two years. Then I decided to take a break, reevaluate goals and set my mind on the portfolio. Now it consists of just two artworks which I spent plenty of time on but they perfectly demonstrate all my skills.
Being a part of my plan to level up as an artist, The Barbarian is a character presentation reel I made from the beginning (base sculpting) to the end (final smooth video sequence). My main challenge was to raise the bar of quality of each step as high as I could, step out of my comfort zone whenever I had a chance. I can’t say that the result exactly matched what I initially planned but the experience I got was incredible plus now I have a newly completed artwork in my portfolio. Also, I wanted all my theoretical knowledge to be put into practice, like using multi-tile textures, 32-bit displacement, mimic morphs, corrective morphs, hair dynamics (unfortunately I did not obtain the desired result here), etc.
I tried to make the whole project as challenging as possible, and the modeling part wasn’t an exception. It started with a sphere for the head and Z-spheres for the body. Any character should be started from a sketch to feel a general shape, silhouette, location, size of the additional details. However, if the concept dictates a lot of bare skin, and you have not prepared the body beforehand (I didn’t do it deliberately) the best way is to start with the body because it will be the base for everything else. It really doesn’t matter which part is first, but I prefer to start with a head: sculpt a general shape, improve similarity with the concept without details and switch to the body. When the rough body is ready it can be attached to the head and I continue working on the whole character. At this stage, it is very important to be aware of the anatomy and skeleton structure, understand how and where muscles are connected to the bones and have a sufficient collection of references. When the first version of a character is ready, match it with the concept once again and consider whether it should be changed. In my case, I needed to slightly overbuild the muscles. The final result was far from being ideal but I did my best to make it look similar to the concept. The task was complicated by the fact that the character is shown in the aggressive pose and from the unusual angle, with an expressive face, while I needed a neutral pose for the animation.
After the body was ready, I checked my character again, from head to feet, to adjust details, likeness, proportions, and anatomy. At that stage, the character relatively resembled the final version but it still was a DynaMesh sculpt without details. On the other hand, it was ready for a final topology and UVs. Before proceeding to retopology I sculpted rough versions of all the outfit pieces to see how it would look altogether, chose the right places for details, changed the body shape to match it with new things. In other words, I checked if the overall look fit the concept.
Face, if it is open or actually exists, is probably the main part of the character. And the eyes are above all, as it is in our human nature to notice them before everything else. Of course, head and face could be considered as a compound object but you should always keep in mind that they are inseparable. In the concept, the face is contorted with emotion which gives the character a certain charisma and shown from the angle. On the one hand, it is great, on the other – it is an obstacle for copying and achieving likeness because the model should have a calm face. That is why I decided to look for similar faces and the most interesting concept-like person happened to be an actor Jeff Bridges. I did not intend to copy him but I took many features from that reference. When modeling a face it is important not to detail it more than other parts. I was adding details to the whole character simultaneously, in a few sessions each time adding a new portion of changes. It is never late to fix general shapes and details. If something was done wrong and you clearly see it – fix it even if it requires changing other parts. Visible mistakes should be fixed, as it is a key to success.
When all parts were ready and in right positions, had proper sizes, general shape and properly detailed, I decided to ease the texturing and sculpted micro details typical to the planned materials. For leather parts, there were three levels of details. First is general shape: large pieces, big cracks, and cuts, ornament, etc. The second level included light scratches, scuffs, large leather texture. And the third one is micro details: various leather textures and micro deformations. To unite all of that I used not only Alphas but also standard brushes like Dam Standard, Trim Dynamic, Clay Buildup. With such a highly detailed sculpt (and a texture as a result), it’s much easier and fun to create composite materials.
Hair & Fur
I am almost sure that my approach to hair and fur was wrong to some extent. It was more convenient for me to build up and comb it in ZBrush with Fibermesh. It allows quickly assign and comb separate groups of hair and fur. However, it was not hair creation that caused the biggest trouble, but hair animation. The hardest part was the beard since it was divided into separate groups, fastened at the ends. These tied ends must not only determine dynamics but also be animated themselves according to the physical laws. I had been struggling with this problem a lot of days and could not find anything better then animate Fibermesh strips with 3ds Max Cloth. I hope someday this process will become easier and friendlier for the artist, relieving him/her from the technical aspects and letting concentrate on pure art instead. In the end, I made the hair with Cloth and Ornatrix and fur with Hair and Fur modifier. I know that this approach to the fur creation in Ornatrix is incorrect. When I was rendering the final version, Ephere announced a new feature of their plugin: improved interactive dynamics based not on PhysX. I made me green with envy but decided not to change my project.
Continuing Working with Textures
As I mentioned before, I decided to sculpt all micro details. All skin defects and scars were made right in ZBrush. The main difficulty there (and I knew about it from the beginning) was the fact that the right side of the character reveals open anatomy from the head down to the neck, chest, back, and arm. To support the details in close-ups I made six UDIMs for the body, where the face was the main part and its scale in UV was two times bigger than other parts. When exporting from ZBrush I used textures with 8k resolution aiming not only at rendering for 4k video close-ups but also for printing in 300 dpi with no loss in quality. I made all the textures in 3DCoat. It has a perfect PBR editor with the possibility to change the workflow during texture export – an amazing classic way of texturing following modern trends. I talk about the advantages of that software for ages. For the skin, I generated additional depth maps SSS, layers Color, Shallow, Mid and Deep for SSS, and also two reflection maps. All of that was used in V-Ray skin material. The final version of the skin was improved by rather small post-processing and local Glow effect in After Effects.
I still haven’t sorted out the hair and fur topic to the full, especially their materials and animation. I painted them by simple color maps: one map for the hair base, another one for the top part. Each hair takes data from the UV of the object it grows from.
Animation and rigging is a large field where I’m not an expert to talk about it with confidence. However, I can briefly describe what I did. First of all, I made a base skeleton CAT on the base of the final version of the character. Then I assigned all main meshes to the bones by Skin modifier, except for the parts that were to be animated by calculated dynamics. Then I made a simplified proxy model and began animating. Firstly I adjusted a parametric walk and built a trajectory in CAT. Initially, the character walked upstairs, stopped on the platform, flew up and fell to the floor. So the video actually starts from the middle, continues by the last part and ends with the middle part again. When the animation was ready I set cameras and animated them. Then I created an animatic and set a camera sequence and combined it with music. When I was satisfied with the animatic, I moved to the dynamics of the moving parts which were going to swing and interact with other parts according to the physical laws. These were leather strips along the legs, chains, belts, cloths, frontal leather rags, a ring in the big axe and a small detail on the sword.
I used Cloth to calculate the dynamics of the elastic parts and also made simplified versions of the final models to accelerate the rendering. When the dynamics of the “fabric” were ready, I used SkinWrap to combine those parts with the main ones. Then I made dynamics for solid parts, chain, and ring. I did it with PhysX tools and a few new simplified objects for dynamics. Here I faced a problem – abrupt movements of the character tore the chain. To fix it I applied Ball&Socket constraint to each piece of the left and right chains. It helped. For the ring in the axe, I used Hinge Constraint. When all physical-based dynamics were calculated, I saved the animation with Point Cache to avoid any surprises during the rendering. As for the muscles, for more believable animation of such a muscular man, I used corrective morphs in all extreme positions of the limbs and main bends. It compensates for the lack of the correcting bones, smoothes minor mistakes in rigging and provides the right transformations where the body bends. To create correcting morphs I bent the limbs till the extreme positions, exported them to .obj and started changing the mesh in ZBrush until I got the desired result. I ended up with twelve correcting and ten mimic morphs.
Rendering took a lot of time, enough for me to have a good rest and reconsider what I had done. One of my conclusions was the following: next time I would make an animation with RT Render. This time all my models weren’t appropriate for RT an I had to use V-Ray. I decided to render the character separately from the environment for the more convenient assembling of the final video. My final picture was in OpenEXR, Half Float. I saved a few additional passes for the character: zdepth, reflection, object id and alpha. Also, I rendered the cast shadow separately. The biggest challenge with rendering was to get a good final quality without noise and keep the calculations within 10 minutes per frame. The process was slowed down by 32-bit displacement. Probably it was better to replace it with a normal bump somewhere because sometimes calculations took about 40 minutes for close-ups, and it is too much.
I prepared a special lighting configuration for each act (video consists of three acts). Each time I slightly adjusted cameras to emphasize metal with patches of light or to add volume. To my mind, to make just one configuration for all cases was not enough.
Well, I spent 15 months on this project, and it is a lot of time. My plan was to finish it earlier. My key motivation was probably not to let myself stop halfway because it would have meant a total waste of time I had already spent. I had no choice but to go on, not lose too much time for details, explore unexplored topics, and control quality to get the acceptable result.