End recording when the user stops, not the music

Silero VAD answers "is this voice-like?", which music, TV and podcasts
satisfy just as well as the user does. With Spotify playing through the
device's own speakers the VAD never went quiet, so record_utterance never
saw the end of the command and ran to the 20 s cap on master and slave
alike -- and music alone was enough to defeat the "no speech during the
grace window" early return.

Wrap the VAD in a SpeechEndpointer that tracks the background loudness of
the recent window and only reports speech while a frame is clearly louder
than it. Both wake-word loops feed their idle audio through observe(), so
the gate is already calibrated when a turn starts, and only non-voice
frames update the background so a long sentence cannot raise the gate onto
itself. Onset/release hysteresis keeps trailing quiet syllables.

The master owns the knob (wake_word.voice_gate_ratio, 0 = off) and pushes
it to satellites in the wake-word block.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

config.example.yaml

@@ -56,6 +56,10 @@ wake_word:
   vad_threshold: 0.5         # silero-vad probability threshold
   vad_gate: 0.0              # >0: wake only when openWakeWord's own VAD agrees (fewer false wakes from TV/music)
   silence_ms: 800            # end-of-utterance silence
+  voice_gate_ratio: 3.0      # how much louder than the background a frame must be to count as
+                             # the user talking — lets recording end while your own speakers play
+                             # music (Silero rates music as speech). 0 = off, raise if commands
+                             # get cut off, lower if music keeps the recorder running.
 
 mcp:
   servers_config: "mcp_servers.json"