Week 11
From Fracturing to Adding Debris To Simulation
This week marked the beginning of the debris simulation phase, transitioning from just fracturing the environment to layering in dynamic secondary motion — in this case, impact-driven debris. The focus was to implement a debris emitter system that could replicate the explosive energy seen in the reference shot, particularly in how large chunks of substructure erupt from beneath the tiles, resembling broken concrete slabs.
To simulate this behavior accurately, I broke the task into two core components:
1. Procedural CUSTOM Velocity Tool
The first step was to create a custom velocity shaping tool that could generate a controllable explosive burst from the ground. Built procedurally, this setup allows for easy tweaking and reuse in future FX shots. It helped visualize how the initial velocity impacts debris spread, directional control, and timing, offering a high level of art-directability.
2. POP-Driven Packed Debris Simulation
Next, I implemented a POP source emitter that injected packed debris into the simulation. Rather than letting debris float above the destruction, the particles were emitted within the RBD sim, ensuring interactivity with the fractured tiles. This helped the debris feel physically grounded and responsive as tiles moved and fractured.
The debris itself was built from a custom concrete slab debris library, chosen to match the reference — where the debris seems to erupt directly from under the tile layer, not fall onto it.
Flipbook Testing & Timing Analysis
I ran multiple flipbook render tests to dial in the visual behavior, coverage, and emitter timing. After observing the reference closely, I noted that the debris begins emerging at around frame 23 — a crucial cue that was missing in the early tests.
This early attempt had some visual appeal, but also led to valuable feedback and observations that shaped the refinements in the later week:
Emitter needs impulse animation — a timed blast, not a constant emission.
Emitter should remain still — avoid animated movement that breaks the realism.
Velocity needed refinement — current debris had unnatural speed spikes.
Debris was blocked — ground cracking pieces interfered with velocity direction.
Too much debris coverage — the shot doesn’t have debris falling toward camera.
Debris timing needs to sync more closely with the bulge/crack event (~frame 23).
The FX seemed to be missing the dome-like cracking structure, which pushes tiles upward before debris emission — this was tackled in the following weeks.
One out of the multiple test versions for the debris emitter
Looking Ahead
This setup formed the first iteration of the debris system and revealed key areas for improvement — especially the need for event sequencing, timing sync, and clear separation between the cracking and explosion stages.
Next week, I will be looking into redesigning the debris system to be physically reactive and driven by fractured tile motion, resulting in a much more grounded and impactful simulation.