Auto-Charcuterie

0 fps1 slices0 particles

Arrangement

The critical case. Does it drape over a dome, or tent across it like a circus canopy?

Solver

Grid resolution8Particles per side. Cost is quadratic.Slices1Perf target: 60fps at 8.Substeps20More substeps, stiffer and more stable.Iterations16Constraint passes per substep. Forced even.Structural stiffness1.00How much the weave resists stretching.Shear stiffness0.85Resists in-plane racking.Bend stiffness0.350The floppiness dial. High = holds a fold instead of pooling.Damping0.985Velocity retained per substep.Friction0.55Stops slices sliding off domes.Relaxation1.40Jacobi over-relaxation. Too high oscillates.Gravity9.8Particle spacing0.062Rest length. Sets overall slice size.Thickness0.014Collision offset. Raise it if slices look like paper.Slice separation0.030Gap held between different slices. Reads as slice thickness.Slice-vs-slice0.80Push-apart strength. 0 disables the spatial hash entirely.Wind0.00Breaks unnatural symmetry on flat settles.Max step0.020Displacement clamp per substep. Anti-explosion guard.

Settling

XPBD never reaches exact equilibrium, so a stack shimmers forever unless slices are explicitly put to sleep. A frozen slice stops simulating but still collides, and wakes if something lands on it.

Tint sleeping slices blueSleep delay45Quiet frames before a slice freezes. 0 disables sleeping.Sleep threshold2.0e-8Peak v² below which a slice counts as still. Raise to settle sooner.

Exit criteria

All four must hold before Phase 1 starts. If they can't, we fall back to rigid slices with baked drape shapes.

Drapes over the mound rather than tenting across itEdges fold and curl rather than staying flatSettles without jitter60fps with 8 slices simulating