About Author: Boink Boink

I’ve created a test environment to set up the compositing pipeline in Nuke that I will be using at the end of the project for compositing the renders. I also tested several material techniques to achieve different realistic material looks. I made a really simple cliff in ZBrush where I polypainted a texture and also an alpha texture for growing grass on the cliff. I also made an ocean with internal sub-surface scattering (particles light bounces off) to give the water a realistic look. It didn’t take much longer then 4 hours to put this scene together, so I am very confident about the speed in wish the final environment will be created for my final project.

When I render out images with VRay I always save them as floating-point 32-bit OpenEXR (.exr) image files. The advantage with using OpenEXR is that all the render passes gets saved inside one EXR image file, which makes it so much easier to work with the compositing. When importing EXR files into Nuke; Nuke automatically reads all the render passes so you can extract and composite the render as you like with ease. The pipeline you see in the image above was the first pipeline I put together quickly.

I wasn’t entirely happy with the setup the previous pipeline had, so I created a new and improved one that also allows for a bit more control. In this pipeline the lighting and global illumination passes are divided by the diffuse filter in order to extract the raw passes (so I don’t need to actually render the raw passes out). The raw passes are the lighting and global illumination unaffected by materials, so you can realistically change the lighting without breaking the realism.

Compositing allows you to manipulate elements of the render without needing to render again and again. The entire purpose of compositing is to greatly enhance and improve the visual quality of your renders. It also saves you a LOT of time in the big scheme of things.

These two renders took 17 minutes each to render out. Which is pretty fast for still shots. If I where to animate the environment, I would change the displacement with normal mapping instead which would greatly decrease the render time. The visual quality would still look the same as long as you don’t directly pause the animation and look closely at the terrain. For still frames using displacement is definitely a plus since it adds an extra level of immersion in detail.

VRayLightMtl (VRay light material) is a material I will definitely use in my project. It is a material that illuminates the scene through the global illumination pass and can create very interesting environment lighting when the material is masked to give certain effects. The advantage of using light materials in VRay is how fast you can achieve noiseless render quality in animation and still frames. Magical effects, glowing eyes, and so on can look really good with this material as it will emit light into environment.

This is a render I created some time ago which directly benefited from the usage of VRayLightMtl. The scene also uses VRayEnvironmentFog so that the light from the lava scatters into the atmosphere.

When creating the clothing for the character I knew that they had to be physical because of the “gills” on the jacket and the pants. This is not a problem for ZBrush to handle since when a surface is very thin, the brushes ZBrush has sculpts on both sides so it keeps the volume of the mesh right . I created all the base meshes for the clothes in 3Ds Max. I did not focus on keeping them from not intersecting in 3Ds Max because all that will be fixed when modelling in ZBrush. I also created the base mesh for the claws as I modelled the other elements.

Once the base meshes are imported into ZBrush I start sculpting. Most of the elements will exceed 2 million polygons. The armour vest ended up on 4 million polygons for example. Some of these elements will need retopologizing and some are just fine as they are already.

I’m a little over half way done with the clothing and then I’ll retopologize the meshes and the body and start fleshing out the real details, specially the characters skin and also get the personality in the face right.

Here you can see all the elements the character is made of right now.

I’ve been using a method in ZBrush which is about modelling complicated parts of the body separate from the main form and then retopologize and project the new detail onto the body. This method gives you a lot of freedom to experiment with the form your looking for without out worry about the topological aspect of it. Only at the end do you merge the different objects together to create one seamless model.

I created a simply box in 3Ds Max where the ears are suppose to be and imported it into zbrush and then sculpted it to intersect with the main body. I can easily move the placement of the ears without worrying about masking parts of the body off to not ruin other parts of the model.

The modelling of the spikes is a bit of a challenge. I knew I couldn’t create them from the body so they are made as a separate object. The spikes are a bunch of cylindrical spikes that has been copied, pasted and deformed to fit their place, one after another. This becomes the base mesh for the sculpting of the spikes. I modelled the spikes all the way down to the shoulders, but I am not sure if I should use those small ones and maybe just model those smaller spikes out of the skin of  the main body of the character instead.

Just a quick post to show this aspect of the planning.

The first thing I started with when modelling the characters body is creating a simple mesh with ZSpheres in ZBrush that converts into polygons for sculpting. The topology is not important at this stage. It is all about form and getting the shape right. The appropriate topology will be applied after the model is about 80% done by using 3Ds Max’s Graphite Modelling Tools to create polygons on the surface of the model. After that I import the new topology and then use projection in ZBrush to transfer all the detail from the old topology model to the new one.

ZBrush’s ZSphere have helped my work flow a lot and really empowered me (and all the other users of ZBrush) to easily create form and figures from scratch without concern about topology. ZSpheres increases the productivity of character production pipelines, so everyone who wants to create characters should really try ZBrush.

When I got the ZSphere mesh generated I start sculpting it until it the model takes the form of the character. I subdivide the mesh step by step as I establish more and more general forms. When the body is properly sculpted and retopologized, I will have to do a modelling pass where I refine everything on the model to make it look really good.

When most of the overall form is established I will start on the ears and the spikes on his head.