SpeedTree® Cinema’s v7 animated growth simulation has proven especially popular among our customers, with Noah and two commercials already featuring the new capabilities. Digital Domain used v7 growth to capture the progress of a South African city, and, as described below by Creative Director Grant Miller, Ingenuity Engine of Los Angeles used the feature to promote water conservation in a vibrant, 30-second spot for Southern Nevada Water Authority:
“During the pitch process for this project we completed a short test to win the job and confirm our technique. For the test we used existing flower assets and scaled them to simulate their growth. While this worked fine as a proof of concept, when we got into the storyboards we realized we’d need much more detail in both the plants and their growth animation. This was compounded by the fact that the client requested specific water smart plants native to Southern Nevada.
With one month to complete the project this meant we’d need to build roughly 20 unique plants and flowers with animated growth in just a few weeks. We turned to SpeedTree, which had conveniently just added plant growth in the 7.0 release. SpeedTree has always served us well for creating specific trees and plants; the procedural generation combined with manual finishing touches are a great balance and very easy to use. The new controls in 7.0 provided the same experience for animating growth: speeds can be offset, tangents can be smoothed, and individual elements can be randomly delayed for realism. After creating each plant asset in SpeedTree, we exported the final growth sequences using Alembic in order to assemble the scene in 3dsMax.
For the growing vines that frame each word we created a straight vine in SpeedTree, exported its growth sequence, then used path deformers in 3dsMax to allow specific placement to the camera. The overall growth of the plants as well as the general environment was handled by Forest Pack which added support for animated growth maps in the 4.2 beta. We also hand-placed dozens of plants and flowers to allow for precise timing and artistic control. To keep the scene light, hand-placed plants were swapped out for delayed-load Alembic caches once each section was finalized. The completed scene consisted of over 30 million non-instanced polygons and was rendered in Maxwell in roughly 2 hours per frame with in-camera motion blur, depth of field, and single-sided sub-surface scattering on the leaves and petals.
Generating the large number of assets quickly in SpeedTree allowed us to focus on artfully assembling the final scene and afforded us the breathing room we needed to make revisions and final touches. I can honestly say that without the advancements from SpeedTree and handful of other vendors that we would not have been able to complete this great project on time.”