S2 Sparrow-2 signal panel

What the room says
before the words do

Sparrow-2 classifies every 80 ms of audio while it works out whose turn it is. This joins the room as a listener, shows the five verdicts it broadcasts, and pushes the room's condition back to the PAL so it can do something about it.

Send my mic Coach the PAL

↔ Speaking

You speaking PAL speaking

·) Momentary verdicts

—

Silence

—

waiting for speech

backchannel 0 · nonsense 0 · interruption 0

∿ CFT — turn taking, last 2s, every 1s

cft mean —

backchannel mean —

interruption mean —

no trace yet

≈ Background confidence — last 2s, every 1s

noise mean —

speakers mean —

no trace yet — needs SPARROW_TWO_BACKGROUND_LEVELS=1 on the PAL

What the room saysbefore the words do

What the room says
before the words do