feat: ack tone, duck-before-TTS, instant ducking on voice/unmute

- Add ascending two-tone chime (880Hz/1320Hz) before TTS playback
  as audible acknowledgment that the voice trigger was recognized
- Signal music ducking 1.5s before TTS starts so music is already
  lowered when audio begins playing
- Snap duck volume to floor instantly on voice packet or user unmute
  via pymumble callback, eliminating the 1s poll delay
- Register USERUPDATED callback to preemptively duck when a user
  unmutes (they're about to speak)
- Strip leading punctuation from trigger remainder (Whisper artifacts)

plugins/voice.py

@@ -11,6 +11,8 @@ import asyncio
 import io
 import json
 import logging
+import math
+import struct
 import threading
 import time
 import urllib.request
@@ -83,6 +85,50 @@ def _pcm_to_wav(pcm: bytes) -> bytes:
     return buf.getvalue()
 
 
+# -- Acknowledge tone --------------------------------------------------------
+
+_ACK_FREQ = (880, 1320)    # A5 -> E6 ascending
+_ACK_NOTE_DUR = 0.15        # seconds per note
+_ACK_AMP = 12000            # gentle amplitude
+_ACK_FRAME = 960            # 20ms at 48kHz, matches Mumble native
+
+
+async def _ack_tone(bot) -> None:
+    """Play a short two-tone ascending chime via pymumble sound_output."""
+    mu = getattr(bot, "_mumble", None)
+    if mu is None:
+        return
+    so = mu.sound_output
+    if so is None:
+        return
+
+    # Unmute if self-muted (stream_audio handles re-mute later)
+    if getattr(bot, "_self_mute", False):
+        if bot._mute_task and not bot._mute_task.done():
+            bot._mute_task.cancel()
+            bot._mute_task = None
+        try:
+            mu.users.myself.unmute()
+        except Exception:
+            pass
+
+    frames_per_note = int(_ACK_NOTE_DUR / 0.02)  # 0.02s per frame
+    for freq in _ACK_FREQ:
+        for i in range(frames_per_note):
+            samples = []
+            for j in range(_ACK_FRAME):
+                t = (i * _ACK_FRAME + j) / _SAMPLE_RATE
+                samples.append(int(_ACK_AMP * math.sin(2 * math.pi * freq * t)))
+            pcm = struct.pack(f"<{_ACK_FRAME}h", *samples)
+            so.add_sound(pcm)
+            while so.get_buffer_size() > 0.5:
+                await asyncio.sleep(0.02)
+
+    # Wait for tone to finish
+    while so.get_buffer_size() > 0:
+        await asyncio.sleep(0.05)
+
+
 # -- STT: Sound listener (pymumble thread) ----------------------------------
 
 
@@ -170,7 +216,7 @@ async def _flush_monitor(bot):
 
                 trigger = ps["trigger"]
                 if trigger and text.lower().startswith(trigger.lower()):
-                    remainder = text[len(trigger):].strip()
+                    remainder = text[len(trigger):].strip().lstrip(",.;:!?")
                     if remainder:
                         log.info("voice: trigger from %s: %s", name, remainder)
                         bot._spawn(
@@ -243,8 +289,10 @@ async def _tts_play(bot, text: str):
     if wav_path is None:
         return
     try:
-        # Signal music plugin to duck while TTS is playing
+        # Signal music plugin to duck, wait for it to take effect
         bot.registry._tts_active = True
+        await asyncio.sleep(1.5)
+        await _ack_tone(bot)
         done = asyncio.Event()
         await bot.stream_audio(str(wav_path), volume=1.0, on_done=done)
         await done.wait()