Files
Kelsi 4f45b9f83a fix(water): wake froth read as painted discs, and the V collapsed to a line
The blotches were solid because the cellular breakup was inverted.
cellularFoam returns distance to the nearest cell point, so thresholding it
at 0.11 marked only the handful of pixels sitting on a point — isolated
specks over most of the patch. The interior was then filled to fully opaque
by mixing coverage to 1.0 wherever the foam was dense, which is every pixel
near a wake point. The result was a hard white disc with a 25% rim.

Threshold across the middle of the cell-distance range instead: most of the
patch becomes foam and the gaps between cells are the texture, which is
what aerated water looks like. The discs also lose their plateau — falloff
now runs from the centre out — and cap at 60% rather than near-opaque, so
the froth lightens the water instead of painting over it.

The V collapsed because the lateral drift was damped at 1.6/s, which bled
it away inside half a lifetime and let both arms fall back onto the path.
A wake's arms are straight because separation keeps growing with distance
behind, so the drag is now nominal. Emission spacing drops to 0.40/0.60 so
patches overlap into a continuous trail rather than a row of blobs, with
lifetimes chosen to keep a swimmer's trail inside the 32-point budget.
2026-07-31 12:28:32 -07:00
..