Wick
Candle flames that fall into step.
Preparing the film…
Films recorded from the standard air. Live on this M4: 6.7 simulated ms / wall s at standard, 15 at light. Film: 30 at 1000 fps, 120 at 250 fps.
Measuring…
z · collecting 0.6 s
Recorded window 0.0 s · RMS 0.000
Left 0.00 Hz · right 0.00 Hz
Cluj, 2020: z +0.74 at 2 cm
0.0 ms / wall s · recorded air
Standard grid recording
Light one thin candle and its flame stands still. Tie three together and the flame flickers, about ten times a second. It stretches, a waist forms, and the tip tears off and floats up, then burns out as it rises. Put two bundles side by side and they flicker in step. Move them apart, past three or four centimetres, and they take turns.
Attila Gergely, Bulcsú Sándor, Csaba Paizs, Robert Tötös and Zoltán Néda measured this in Cluj in 2020. They used candles 8 mm across and filmed them at up to a thousand frames a second. They also showed that the flames don't steer each other by their light. They set a lamp beside a bundle and made it flicker at the flame's own rate, and nothing locked to it. It is the air. Tao Yang, Xi Xia and Peng Zhang showed how in a simulation. Each flicker is a ring of air that rolls up around the flame and lifts off with its tip. Between two flames, the inner edges of those rings either roll up together or take turns.
I didn't draw my flicker. My flames are hot air, worked out cell by cell in a box 14 cm wide. The gas gets lighter as it heats, and it rises. Wax vapour leaves each wick at a milligram a second, the rate I read from their scales. It burns where it meets enough air. Two things in it are my choice, not theirs: how hot the flame gets at its hottest, 1,800 kelvin, and where its soot glows. I didn't put in the flicker, or its speed, or whether a pair keeps step. They come out of the air.
Some of it comes out as they found. A bundle flickers and tears off its tip, and bigger bundles flicker more slowly. Two bundles 2 cm apart flicker in step, and the pair slows to nine tenths of a single bundle's rate. That is what Yang, Xia and Zhang's simulations do. Between 3 and 3.5 cm my pair stops keeping step and starts taking turns, inside the range where theirs changed.
Some of it doesn't. In Cluj the change came all at once. Mine is untidy. At 3 cm my pair only half keeps step, and at 4 cm it drifts, neither in step nor in turns, before taking turns again, loosely, at 5. Every bundle of mine flickers about 15 percent too slowly. I think my flames are too wide, because my cells are a millimetre and a half across and the wick comes out fatter than it is. The way I carry the vapour also doesn't keep its mass exactly, so a flame can hold a sixth more or less fuel than it should. And my single candle flickers, where theirs stood still. It still flickers on half the wax. I've asked Zoltán Néda what holds a real single candle still.
The picture copies their camera. Their films are online: the flames clipped to white on black, the candles lost in the dark, a thousand frames a second shown at thirty. I measured the greys, the edges and the scale from their frames, and I draw mine the same way. My air runs too slowly to play live at that speed, so what you see are films of it, recorded here. Choose a distance and its film plays. You can also run the air live and watch it work, far more slowly. The lines under the film are each flame's light, counted pixel by pixel, the way they counted it.
It doesn't look like their film yet. I showed my frames to readers who had never seen them. They named candle flames first, and still called it a render. Two more readers each saw three frames of theirs beside three of mine, and both picked theirs every time. Mine are too short and too wide, their tips end too bluntly, and their edges are all equally soft. The width is the air's own, the same width that slows the flicker. The softness comes from drawing cells a millimetre and a half across at a third of a millimetre a pixel.