fix audio output thru mixer

Author: chenosaurus

examples/local-audio/full_duplex.py

@@ -1,19 +1,24 @@
 import os
 import asyncio
 import logging
+from dotenv import load_dotenv, find_dotenv
 
 from livekit import rtc
 
 
 async def main() -> None:
     logging.basicConfig(level=logging.INFO)
 
+    # Load environment variables from a .env file if present
+    load_dotenv(find_dotenv())
+    
     url = os.getenv("LIVEKIT_URL")
     token = os.getenv("LIVEKIT_TOKEN")
     if not url or not token:
         raise RuntimeError("LIVEKIT_URL and LIVEKIT_TOKEN must be set in env")
 
     room = rtc.Room()
+    
     devices = rtc.MediaDevices()
 
     # Open microphone with AEC and prepare a player for remote audio feeding AEC reverse stream
@@ -43,7 +48,7 @@ async def _remove_stream(stream: rtc.AudioStream, participant_sid: str | None =
         if pub_sid is not None:
             streams_by_pub.pop(pub_sid, None)
 
-    async def on_track_subscribed(track: rtc.Track, publication: rtc.RemoteTrackPublication, participant: rtc.RemoteParticipant):
+    def on_track_subscribed(track: rtc.Track, publication: rtc.RemoteTrackPublication, participant: rtc.RemoteParticipant):
         if track.kind == rtc.TrackKind.KIND_AUDIO:
             stream = rtc.AudioStream(track, sample_rate=48000, num_channels=1)
             streams_by_pub[publication.sid] = stream

examples/local-audio/publish_mic.py

@@ -4,10 +4,7 @@
 import logging
 from dotenv import load_dotenv, find_dotenv
 
-sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "..", "..", "livekit-rtc")))
-
 from livekit import rtc
-from livekit.rtc import MediaDevices
 
 async def main() -> None:
     logging.basicConfig(level=logging.INFO)
@@ -23,7 +20,7 @@ async def main() -> None:
     room = rtc.Room()
 
     # Create media devices helper and open default microphone with AEC enabled
-    devices = MediaDevices()
+    devices = rtc.MediaDevices()
     mic = devices.open_microphone(enable_aec=True)
 
     try:

livekit-rtc/livekit/rtc/audio_mixer.py

@@ -184,6 +184,10 @@ async def _get_contribution(
             except StopAsyncIteration:
                 exhausted = True
                 break
+            # AudioStream may yield either AudioFrame or AudioFrameEvent; unwrap if needed
+            if hasattr(frame, "frame"):
+                frame = frame.frame  # type: ignore[assignment]
+
             new_data = np.frombuffer(frame.data.tobytes(), dtype=np.int16).reshape(
                 -1, self._num_channels
             )

livekit-rtc/livekit/rtc/media_devices.py

@@ -181,7 +181,7 @@ def _callback(outdata: np.ndarray, frame_count: int, time_info: Any, status: Any
                         break
                     render_chunk = outdata[start:end, 0]
                     render_frame = AudioFrame(
-                        render_chunk.tobytes(), FRAME_SAMPLES, 1, FRAME_SAMPLES
+                        render_chunk.tobytes(), self._sample_rate, 1, FRAME_SAMPLES
                     )
                     try:
                         self._apm.process_reverse_stream(render_frame)