Talker

idle

prebuffer 5 frames volume auto level decode in browser pitch 0 st timbre 0 % brightness 0 dB/oct media element output

output

pitch (log scale, 50–500 Hz) · level (dB re speaker reference) · phone labels at boundaries · last 10 s

link

bit/s, last second–

bit/s, 10 s average–

bits per frame–

frames/s received–

packets/s–

packet jitter (rms)–

bytes received–

timing

server capture → send–

server inference–

client queue–

playout–

underruns–

frames rendered–

synth load–

audio

audio context–

output loudness–

client level lift–

codec monitor stream–

browser decoder–

speech

pitch–

level–

symbols in vocabulary–

recognizer–

source–

WebTransport: –
certificate: –