Here are a couple images of a digital sculpture that I did for a ‘recently released pirate movie’. At the time the film was in early preproduction and this piece was more speculative design than anything concrete for the film, though it is loosely tied to an event in the script. The piece also has thinly veiled references to Damien Hirst’s infamous pickled shark sculpture. For more details on how I created the sculpture keep reading below.
The digital sculpture was executed to a very tight schedule, going from start to final rendered image in two days! Funny, but it seems some of my best work comes under this sort of tight deadline – less time to indulge every option and fiddle with the details I suppose. For people interesting in technical details on how I created the digital sculpture (3d geeks – you know who you are!), here is a little background on the process:
The piece was created primarily in Zbrush, using Zspheres as a starting point for the body. The pose and gesture of the body was established immediately and sculpted in place. By sculpting “in-pose” like this I do waste a little time not using symmetry, but often the result is more natural when both sides are created independently. Also when the body is posed, each side of the body is only vaguely similar in form.
Now, once the body and face were developed to 80% (meaning the volumes, pose, and proportions looked good, but the details weren’t finished) I exported a high-resolution version of the mesh to a piece of software called 3d-Coat. There I used its incredible voxel sculpting tools to add the hair and fins to the existing sculpture. I didn’t spend too much time detailing the fine curls of the hair in 3d-Coat, knowing I would send it back to Zbrush for finishing. The last step in 3d-Coat was to auto-remesh the voxel sculpture to a medium resolution and export this along with a high resolution version that retained all the original details.
Back in Zbrush I subdivided the medium resolution mesh a couple times until it was roughly 6 million polygons, then I projected the details from the high-res mesh onto this subdivided model. This gave me a model with a few levels of subdivision, which I find very useful because I generally don’t sculpt solely at the highest subdivision level of a model. I often step down to lower subdivisions to use the low-res mesh as a cage for modifying large forms, volumes, and contours of the model (this is why I created the medium resolution mesh in 3d-Coat).
In Zbrush I finished the last 20% of the model – I brought the face up to an acceptable level, detailed the curls and twists in the locks of the hair, and refined the contact of the body with the glass box.
The final sculpture was decimated using Zbrush’s inimitable Decimation Master plugin and exported to Maya for rendering. The final renders were created with Mental Ray. There were no real tricks in the rendering, I just paid attention to general good practice in lighting and rendering – area lights (MR portal lights in this case), energy conserving materials, lens shaders, gamma, etc.
Thanks for reading, I hope this little behind-the-scenes was broadly informative. For anyone interested in learning more about the tools and techniques of my process, please have a look at the Digital Figure Sculpture course where I go in depth into everything talked about above.