The Making of 'Why is a Raven like a Writing desk'
Nicolas Brunet explains the process he used to create his artwork Why is a Raven like a Writing Desk and shows we can survive after creating a hair and fur scene...
This artwork was initially created for a 2010 CG contest made by the website 3DVF about the theme: ‘Alice in your Wonderland'. The contest ran for several months, and I worked on the project in my spare time. It's on that particular project that I really tamed the hair and fur process.
Lewis Carroll and the real Alice are sailing in a boat, their reflections representing the Mad Hatter and Alice respectively. Water also acts as a mirror in which the elements appears from the universe created by Lewis Carroll.
Due to the visual constraints created from the swirling of the water I was afraid that the reflection was not sufficiently visible and difficult to read. So my second idea was that Alice and the Mad Hatter are lying in a boat, along with the Cheshire Cat, and sailing in the moat of the Queen of Heart castle among the decapitated heads.
Initially I thought of using a realistic style for the characters but I quickly abandoned the idea because it didn't suit the mood of the project.
The process was the same for almost all the elements of the scene; an Editable Poly object modified using visual references, and for some models, a jump into ZBrush to add details.
Cloth simulators were used to give good shapes to the clothes, which were then refined in sculpt mode. In order to put the characters in place, I created a quick rig and skinned the bodies. The easiest character to work on was the cat, his body was modeled in the final position and skinned along a spline. The decapitated heads are from the Mad Hatter model with a few variations.
Setup and lighting
With the key elements created, it was time to put it all together. I had to set the camera up and see if the paper version worked in 3D (the decapitated heads populated moat as particles).
Visually, a light filtering through branches got me interested so I started several tests to get some sort of light/shadows look. I used cups containing candles that could add magic to the final render.
The challenge of the project. The great idea I had on this project was to start with the most difficult character: the Cheshire Cat. Surprisingly it was the easiest model to generate hair and fur on!
The head was separated from the body, and then different hair and fur modifiers were added to the model for separating the different parts to brush: Beard, mustache, nose/forehead, eyes, ears and the back of the head. For each of these modifiers, I played with the multi-strand parameters, textures and noises bitmaps and controlled the intensity and length of hair. The body is made with a single hair and fur modifier.
For Alice's hair, I initially made a modern haircut. Unfortunately a crash corrupted my scene and I had to start the styling from scratch, this time starting with Alice lying on a cushion which served as interactive geometry during brushing. Then a few strands based on splines to cover the forehead to make them look a bit tamed.
For the Mad Hatter, I used the same method, separating the main strands before the brush process.
The part I like the most today: the creation of textures and shaders. With the little knowledge I had at the time, the textures were summarily created. I was still trying to learn how shaders worked.
After trial and errors for a few days, I launched a plethora of tests to get realistic mugs, half transparent dress, a plastic side for the Mad Hatter clothes, an acceptable wood for boat etc…
Muddy water nightmare
The last challenge of the project was recreating the muddy waters of the moat. I started with basic tests; I even filled my bathtub with water and put some soil inside as reference.
The complex part was the volumetric aspect of water mixing with refraction appearance and gradual dispersion. I then discovered the volumetric shader from mental ray which created exactly this type of effect but it was not compatible with my type of lighting and rendering.
In the end I rendered several passes and composited it together in post, with the help of a ZDepth pass.
My biggest regret is not having done enough research in advance to help the legibility of the final image. However it seems that I became a master of hair and fur in the space of two months which is significant. This project was the initial nudge to create other projects that followed since 2010.
A big THANKS to my friend l'Andalou who pushed me to participate in the contest, even if the artwork was not finished on the podium – it was nevertheless published in some magazines and is now part of the 3dtotal gallery.