Create Amazing Character Art
Learn how Abraão Segundo used hair and skin shaders to help him create a realistic piece of art based on Conan the Barbarian
In this tutorial we will learn a bit about the workflow for characters and movies, as well as tips that we can use inside of maya and zbrush to improve our development as a character artist. the main steps will require a lot of attention to avoid loss of quality. the hair was made by my friend, thales Simonato. We worked side by side so i could give him directions and make final adjustments.
Proportions and silhouettes Source a lot of references for the character, weapons and materials, and put them all into a reference plan inside of Photoshop. We need to be careful with the proportions and silhouette of our character, the size of arms, legs, torso and head. For the head size you need to count how many heads there are from the head to the feet – eight is a good average for realistic proportions. Look at any concept that you are following and at anatomy.
Model the head the head is one of the most difficult parts. For primary shapes, find references, check the proportions and the distance between the eyes. Line the ears with the eyes, the base of nose with base of ears, and the mouth more closely to the nose than the chin. For secondary shapes, check all views and break the symmetry: look at zygomatic shapes, chin volumes, dark circles, signs, jaw and other instantly recognisable shapes.
Body – main shapes now we need to focus on the secondary shapes, which are the most visible and recognisable when looking at a model. in this case, these are the muscles, shapes and curvature of each part of the body, arms, legs, torso and hands. after we develop the proportions we will add volume with a brush that you like, Clay or another, with the intensity set to low. always look at a reference with the same pose as you want for your character. this will be easier than trying to imagine it. that way you decrease the chance of making mistakes.
Topology, Maya and Zbrush there are different workflows for modelling. For this process we have to do the maximum amount of details possible on the topology because if we need some close-ups on some part of your character during the movie or trailer then those parts will be there without any dependence on displacement or normal maps.
We can make clothes inside of marvelous Designer, export to zbrush, apply the zremesh tool, fix the flow of topology and finish inside maya. For pieces and details, like the sword guard, create a Dynamesh inside zbrush, export decimation to maya, use the modelling tool kit for topology and then return to zbrush for final details. a clean and correct topology is better for all processes such as UVS, masks and poses.
We can solve most of these using our equipment inside zbrush, with zmodeler. at this stage it is necessary that you make the model exactly the same as the reference.
Separate files after all parts of your character are in the correct position, we need separate ztl files. as you don’t have a lot of files, you can put similar materials in the same file. For example: file 01, all skin Subtools, file 2, all metals and so on. each weapon or shield can be in a different file too. remember that each ztl file can’t be more than 2gb, and each Subtool needs to be less than 50 million polys. a project like this needs good planning if you don’t have a powerful machine.
Work on UVS after we separate the files, it’s time for UVS. We can choose between normal UVS, or Udim in our projects for cinematics. For the best quality for our textures we chose Udim. make polygroups in the topology, respecting the cut lines that you imagine for your project. We will open all that we can inside UV master from zbrush. after we export to maya, to organise it better for Udim UV, we can optimise our textures and organise them by hand or use the Layout tool.
For the head, the best choice is to open inside your 3D software or with Uvlayout.
Add details Divide the Subtool until it’s at a maximum of 50 millions polys. it’s very important to know that the secondary shapes are necessary for achieving natural details. Create separate layers, pores, scratches, veins, wrinkles and skin signs. Use alphas and brushes to reach the best result. Follow the tutorials on the website: texturing.xyz.
For the skin we use the Dam_standard brush with some adjustments. go to Stroke>mouse avg Set 1, this setting will make the brush more sensitive to your movements. For scars we can use the Blob brush or Standard plus Clay Build-up and Veins Standard brush.
Export displacement and normal maps to export the details we can use this configuration: Plugin> multi map exporter> Select Displacement> choose the size (in our case 8192 if you use just one UV set, if you use Udim it can be 4096). Flip V>export Options, and for the Displacement map set as Subdiv Level 1, Dpsubpix 4, mid 0, and set as 32-bit, exr, Scale as 1 and intensity as 0. We can just use displacement for skin because this map brings us all of the details that we need from zbrush. We sometimes use the normal map for assets and small parts.
Apply displacement with Arnold Render export all meshes to maya, organise your scene and let’s start to link all Displacement maps. First, check your Udim files to see if the names are correct. if not rename them in the correct form, for example: name.1004, name.1003. Select the mesh and start the process. apply an a is ta nd a rd surface and open the hypers had er. Select the surface node and link your texture and change the UV tiling mode for Udim (mari). Select the mesh again and open the arnold tab on the attribute editor. Find the two tabs Subdivision and Displacement attributes, open both. Set type to Catclark, iterations to 3 or 4, UV Smoothing to Linear and Bounds Padding to 1.
Texture in the Conan project we used mari for the skin texture on the face. Paint in 8k if you can or, at minimum, 4k. We can use the Project tool in our first layer and a good tip is to use a depth mask for projection. then create a procedural layer with small pigmentation of yellow and red colours. next, create a layer for red areas, as ears, month, nose and scars. Use dark colours for more variation. We can use Substance for all assets and create some masks for mixer shaders inside the maya with arnold render. inside Substance you can customise your material and save for use in similar assets.
Arnold basics Choose the arnold render, set the file format, create a camera and choose the size. We will start the lighting process, so select the arnold renderer and set the sampling. Set the camera configuration. Select the camera that you created, go into attribute editor, open the arnold tab and make your configuration. the last step here is to open the arnold render View, turn on the Progressive refinement and click on update Full Scene.
Use displacement for skin because this map brings us all of the details that we need from zbrush
Light to break the black shadows up, create an environment Light (hdri) and reflection. go into the arnold > light > skydome Light menu. Select the light, go to attribute editor > colour and link your image (you can find awesome images on hdrihaven.com). Create one light at a time, and find the best angle to show volume shapes for your model. Be careful with frontal lights as sometimes you can hide the shapes.
For this project we just needed to use four lights: key lights, on the top, the back light and the fill light on the side. try to represent the same environment of hdri.
For the planes, we use a break direct light and create variations on the shadows.
Arnold shader For shaders, use ai standard surface with the ai mixshader. this way we can work with specular and rough. in the first shader we use a map with a lot of variation to break the unreal roughness. in the second material we apply the maps, Diffuse or albedo, reflectivity, roughness and normal map. Pay attention to the reflectivity and roughness inside arnold render. to reach a better result with these maps you need to link an ai ranger between the image map and the shader.
remember that if you are working with an Udim shader, you may need to rename the file, especially if you export displacement from zbrush. For example: name.1001, name.1002. Change the set when you link the image.
Skin shader For the skin shader, basically we used a mix between two standard shaders: one just for reflection and the other for the skin.
We used six maps, Diffuse, Specular, rough, Deep, Deep mask and SSS (subderm). remember that it’s very important to use the real size inside maya, for the shader to deliver the best result. Sometimes we can adjust in the Scale Shader to 0.9. For the dirt over the skin, it’s easier to make a mix between the skin and other shader, this way you have more control. to do this we can create a mask, inside Substance Painter, use aimixshader to join the two shaders. Connect the mask map in the mix Weight and, mix mode in Blend.
Groom hair here we used Xgen from maya. Create a low region for the hair. On Conan’s hair, the challenge was to achieve dry, messy-looking hair while maintaining the layering and volume at the same time. We ended up using 13 descriptions, including peach fuzz, facial hair, hands, body hair, boots and the actual hair. the workflow for the hair is very similar to the other processes. We start blocking out the main shapes and once it has the right volume we break it up into smaller clumps and add fine details, such as frizz, strays and sub-clumps. For me, this project was very important because it was my first time experimenting with realistic hair grooming.
We ended up using 13 descriptions, including peach fuzz, facial hair, hands, body hair, boots and the actual hair
Render now it’s time to render our project. Let’s set all the configurations. Set the file format as exr, and the size and aa samples if you use depth or hair in your image. make sure you set the Sample above 6, Diffuse at 2, Specular at 2 and SSS at 3. render the character separately from the background, select the Skydome and set the Visibility Camera to 0. now in your final render we’ll come with the alpha channel. after your render is done, you can hide the character and render just the background.
Post-production Open the file inside Photoshop, go to image > mode > change to 16 bits. With a new window open, change the method to exposure and gamma. after this, you can duplicate the background layer, apply a Curve adjustment or a Level adjustment. Create a layer for glows, find the highlight point in your image and choose a soft brush to create small glows. Find some fog images or smoke on the internet, try to find black-and-white images and you can create an environment fog with small particles too using the Screen Blend mode if you want. Be careful with the main areas of the image, don’t put anything in front of your character’s face. Be careful with the adjustments and equalise well, with no black shadows and no intense highlights.