Pulls in zebra-report 9215eaa: JitterBufferProcessor now drops the
queue down to targetSamples (not 1.5×target) on a shrinking retarget,
eliminating the residual 6%-cap speed-up phase fox heard as "janky
slow" during listener-to-speaker promotion. The processor reports
the dropped sample count back to JS so installJitterBuffer's port
handler zeros every paired video receiver's jitterBufferTarget +
playoutDelayHint — video frames drop in lockstep with the audio
skip so lip-sync survives the jump.
Fox 2026-06-06: "if we skip ahead from whatever listener is at to
speaker speed, we need to make sure the video skips ahead the same
amount or rate to keep the lips synced".
Behavior change is isolated to the role-change moment (4s→0.5s on
promote). Steady-state audio and listener-only sessions unchanged.
russell@unturf.com
e45db971aa