If you have been following some of the recent posts in the drawing section, you will know that centaurs are the order of the day. Here is a preliminary study that takes the rough, rough, rough! study from further down this page and refines it, fixing the the anatomy and proportions. Click on the images for larger versions. Stay tuned for more…

Copyright 2008 © Scott Eaton

…with excerpts from William Rimmer’s Art Anatomy, 1877, and rare female anatomy plates from Dr. Paul Richer’s Nouvelle Anatomie du Corps Humain, Vol. II , 1920. Both are classics of artistic anatomy and great learning resources.

Copyright © Pixar Animation Studios

This sketch was developed for Pixar’s RenderMan Courseware and demonstrates using procedural displacement mapping to achieve an interesting visual effect. The setup, covered in fine detail in the Courseware, uses on a particles system that writes primitive variables (primvars) onto the table top each time there is a collision event. The table top has a custom displacement shader that uses these privars to locate the ripples both spatially and temporally. From there, the rest of the displacement shader handles the pattern generation for the concentric rings. A movie of just the ripple pattern is shown below. This was output as an AOV (arbitrary output variable), a framebuffer separate from the final render, and was subsequently used to composite the final effect in Shake.

Copyright © Pixar Animation Studios

Copyright © Scott Eaton

This sequence shows the use of corrective displacement mapping to achieve high resolution facial expression on top of a low-resolution animation mesh. The character on the left is rendered with his standard displacement map (sculpted in Zbrush) as well as a number of per-expression corrective maps, also sculpted in Zbrush. The same character on the right is rendered without the base displacement and shows only the corrective displacement mapping.

The entire pipeline for creating and rigging the facial setup was the topic of my Alias 3December workshop in 2005, titled appropriately “Corrective Displacement Maps for Facial Expressions”. Even though it is old, you can download the slides from the presentation here.

Custom Gather

Copyright © Pixar Animation Studios

The is an example of using the gather() construct in an unusual fashion to create an interesting visual effect. The gather() call fires a hemisphere of rays above each shading point on the ground plane, and instead of tracing the usual diffuse or specular reflections (what raytracing is most often used for), the rays return their length (i.e. the distance to) each hit object, in this case the spheres. From there it is a simple case of normalizing the average distance and mapping it into a colored ramp. I put all the nitty gritty technical details in the RenderMan Courseware, so check there for a step-by-step on how it is done.

Visualizing Ray Depth

Copyright © Pixar Animation Studios

This example, also from the Courseware, is a visualization of ray depth. Each ray is traced into the scene and is reflected/refracted a number of times before it terminates on the bouncing sphere. The “camera” ray is always at trace depth 0, this is when the sphere is in plain view to the camera. As it falls behind the transparent screen the trace depth increases by one. Reflections off the floor and through the screen increase the depth a step further. The technical details of the implementation are in the Courseware.

Copyright © Pixar Animation Studios

Here is another video developed for Pixar’s RenderMan Courseware, this one demonstrates an interesting techniques that can be achieved with a fairly simple custom displacement shader that relies mostly on an understanding of the fundamentals of computer graphics. Specifically, the technique uses a sequence of depth maps (shadow maps) rendered from beneath the surface capturing the distance from the shadow camera to each point on the displacement-mapped spheres. Each point can then be transformed…

…into world space and used as an exact displacement amount for an object pressing into a surface. The effect of using a single depth map for displacement is show below:

Copyright © Pixar Animation Studios

And of course, to achieve persistence, you need to accumulate the depth maps per-frame and use the running total to displacement the current frame. The Courseware goes into detail about this process showing how to do this by either writing an external Python script or using It and Iceman, RenderMan Studio’s internal compositor and scripting language.

Copyright © Pixar Animation Studios