The Heathentown pitch reel was a great opportunity to work with Gabe Hardman, a very talented guy and comic author, who has drawn the boards for many of PLF's previs projects, including Spiderman III, Superman Returns, Logan's Run, and some others.

One challenge facing the previs artist is how to inject character, emotion and subtle performance nuances into previs, without undertaking and elaborate and time consuming facial rigging and performance process, which everyone knows is extremely hard to do well. Storyboards are great for this, but don't always provide the technical information which a previs process seeks to deliver. Board-o-matics in After Effects are one alternative - cutting up the boards and sliding the parts around to tell the story better and time things out. This can work pretty well, but isn't that easy to do well, and still lacks the 3D underpinnings which allow valid technical planning to be don via previs.

I had long wanted an opportunity to develop the drawing based methodology for previs, as a way to put emotion and character into a simple piece, and luckily for me Gabe came along with a cool idea, and the needed willingness to do the needed drawings. The approach I settled on was to start with old school 3D previs, in a simple 3D environment, with simply rigged 3D characters and correctly set up cameras. Gabe was involved in composing the shots, setting up the layouts, selecting lenses etc. We cut these 3D shots together, made adjustments, and had a good idea of what each shot was and exactly what drawings were needed. Gabe then mad these drawings and I took them into XSI, hid the 3D characters, and put the drawings in their place/ Sometimes I had to give him wireframe images for him to draw on top of, and camera-projected these onto the environment. I was able to do some small camera moves on these 3D assemblies of drawings, and preserve the 3 dimensionality of the cameras, lenses, and camera shake, motion blur, etc, Depth of field was faked in After effects.
For shots which had characters in multiple expressions, Gabe did multiple drawings, and I used the FX Tree in XSI to dissolve between them, resulting in mostly single pass renders for shots.

The conquistador posed a challenge to this technique, as the amount of movement was too much to dissolve between. For these shots, I gave Gabe a series of still images of the 3D conquistador in various poses, and gabe drew on top of these. I camera projected these images onto my 3D conquistador, and again got single pass renders of this character's mesh moving the drawing around in a 3 dimensional way.

I think this approach ("Draw-vis") could be used effectively in future projects which are simple in scope and depend on strong characters and emotions, as long as there is an artist on hand willing to draw the specific drawings needed.

You can see how this all worked by studying the "Making Of" reel, which has the shots rendered from Overview cameras showing the 3D scene and camera which created each shot.