Letting rain streams absorb the drops they touch
The site's rain streams were drawn on top of the drops instead of absorbing them. A probe found 123 of 1,013 drops inside drawn streams while the simulation counted 6; making each stream a lane the simulation owns brought that to 0.
In this piece
On 9 October the streams were still running straight over the other drops without absorbing them, and they still didn't look like the streams in the photographs. I had been treating a stream as a line drawn on the glass that drops happen to run along. The photographs show otherwise: a stream and the drop it touches are one body of water, and the drop should join it.
Before changing anything, I measured the old behaviour. Right after load, 123 of 1,013 drops sat inside drawn streams. The simulation, which decides what is physically touching, counted 6. The picture and the physics disagreed about where the water was, which is also why no amount of tuning had fixed it.
Giving the physics the lane
A stream became a lane: a stretch of glass the simulation owns, at the width it is drawn, for as long as it's drawn. A resting drop whose contact patch touches a lane drains into it. Rain that lands in a lane runs down it at once. A runner that swallows a bigger drop keeps its identity. The same probe afterwards counted 0 drops inside lanes.
Then came rules from the physics papers and the hand-traced streams. The running head eats every drop it touches. Height follows width, so beads are tall and necks thin. A fast head leaves pearls and a slow one a film that ages into a dotted line, and a path bends one width across, freezes, and the next runner follows it.
What is still wrong
In my reference clip the streams are a continuous rippled ribbon, and ours were still a chain of beads. I noted it the same day and it stayed on the list.
Sources and notes (7)
- Kaneda 1993, running water droplets — a running drop merges what it touches and leaves a trail
- Le Grand, Daerr & Limat 2005, shape of running drops — how a running drop turns oval, cornered, pearled
- Daerr et al., rivulets — rivulet meanders
- Stuppacher & Supan 2007, drops in real time — small drops stay idle until swept away
- Streams research — the notes this was written from; also records his reference clip E, as recorded in the repo
- Water session handoff — the notes this was written from
- Rain system design — the notes this was written from
Water, rain & windAll 11 parts →
