The delay one SIDE of one output needs so its arrival at the microphone matches the REFERENCE
loudspeaker's.
A LOUDSPEAKER IS A SIDE, not a destination (#845, per-leg spec amendment 2026-09-09): MAIN's
two boxes are the L and R PORTS of one leg, they stand at two distances, and the correction
belongs on the side of the box that measured early. The reference side is not touched.
A DELTA on the line already dialled in, for the same reason §4's is: the out-route delay is
downstream of the analyser tap, so it is already inside arrivalMs. An output measuring
12 ms while holding 4 ms of delay is 8 ms of flight; asked to match a reference at 15 ms it
wants 4 + 3 = 7 ms, not 3.
arrivalMs and referenceArrivalMs are both "time from that bus's own lane tap to the mic's
lane tap" — comparable because every lane's rings are filled in one process callback.
The delay one SIDE of one output needs so its arrival at the microphone matches the REFERENCE loudspeaker's.
A LOUDSPEAKER IS A SIDE, not a destination (#845, per-leg spec amendment 2026-09-09): MAIN's two boxes are the L and R PORTS of one leg, they stand at two distances, and the correction belongs on the side of the box that measured early. The reference side is not touched.
A DELTA on the line already dialled in, for the same reason §4's is: the out-route delay is downstream of the analyser tap, so it is already inside
arrivalMs. An output measuring 12 ms while holding 4 ms of delay is 8 ms of flight; asked to match a reference at 15 ms it wants 4 + 3 = 7 ms, not 3.arrivalMsandreferenceArrivalMsare both "time from that bus's own lane tap to the mic's lane tap" — comparable because every lane's rings are filled in one process callback.