# Convection (the FIRE_FIXTURE driver): no wind response in the temp
# channel (the manual 0.375 m/s sits below the wind_factor floor, so
# the field reads as calm), the brazier deposits heat at its flame
# envelope and the gas lamps a faint thread at the mantle. The channel
# equilibrates in ~8 sim-s (the mist band - days of calm change nothing
# past frame ~480): the brazier column holds buoyancy ~55 heat*px ->
# the derived rise ~31 px/s (~0.5 m/s, the retired 2D channel's
# reference calibration), the lamp's thread stays faint. The run ends
# before the downwind orb ignites (frame 1611 in this scene): the
# merged cluster's deposit spread is a different regime (fire-merge's
# lane). No input: the scene idles.
# frames: 1560
