Merge pull request #1213 from pipecat-ai/aleix/base-transport-output-bot-vad-stop-secs

BaseOutputTransport: use specific VAD stop secs for the bot

CHANGELOG.md

@@ -76,6 +76,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- Fixed an issue that was causing `BotStoppedSpeakingFrame` to be generated too
+  late. This could then cause issues unblocking `STTMuteFilter` later than
+  desired.
+
 - Fixed an issue that was causing `AudioBufferProcessor` to not record
   synchronized audio.
 

src/pipecat/transports/base_output.py

@@ -14,7 +14,6 @@ from loguru import logger
 from PIL import Image
 
 from pipecat.audio.utils import create_default_resampler
-from pipecat.audio.vad.vad_analyzer import VAD_STOP_SECS
 from pipecat.frames.frames import (
     BotSpeakingFrame,
     BotStartedSpeakingFrame,
@@ -38,6 +37,8 @@ from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
 from pipecat.transports.base_transport import TransportParams
 from pipecat.utils.time import nanoseconds_to_seconds
 
+BOT_VAD_STOP_SECS = 0.3
+
 
 class BaseOutputTransport(FrameProcessor):
     def __init__(self, params: TransportParams, **kwargs):
@@ -321,15 +322,10 @@ class BaseOutputTransport(FrameProcessor):
                     )
                     yield frame
 
-        vad_stop_secs = (
-            self._params.vad_analyzer.params.stop_secs
-            if self._params.vad_analyzer
-            else VAD_STOP_SECS
-        )
         if self._params.audio_out_mixer:
-            return with_mixer(vad_stop_secs)
+            return with_mixer(BOT_VAD_STOP_SECS)
         else:
-            return without_mixer(vad_stop_secs)
+            return without_mixer(BOT_VAD_STOP_SECS)
 
     async def _sink_task_handler(self):
         async for frame in self._next_frame():