Physical input channel that a word/turn was attributed to.
A channel name declared in StreamingTranscriberParams.channels (e.g. "mic", "system").
"unknown": no channel was clearly dominant during the word's time window (silent
or all channels evenly active under our threshold).
This is independent of AssemblyAI's diarization speaker_label / words[i].speaker,
which identifies voices by acoustic characteristics. A given speaker_label can map
to any physical channel; the two dimensions can disagree.
Physical input channel that a word/turn was attributed to.
StreamingTranscriberParams.channels(e.g."mic","system")."unknown": no channel was clearly dominant during the word's time window (silent or all channels evenly active under our threshold).This is independent of AssemblyAI's diarization
speaker_label/words[i].speaker, which identifies voices by acoustic characteristics. A given speaker_label can map to any physical channel; the two dimensions can disagree.