Azure TTS fixed by clearing the audio queue before synthesizing the next text

CHANGELOG.md

@@ -72,6 +72,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- Fixed an issue where audio would get stuck in the queue when an interrupt occurs 
+  during Azure TTS synthesis.
+
 - Fixed a race condition that occurs in Python 3.10+ where the task could miss
   the `CancelledError` and continue running indefinitely, freezing the pipeline.
 

src/pipecat/services/azure/tts.py

@@ -282,6 +282,13 @@ class AzureTTSService(AzureBaseTTSService):
         """
         logger.debug(f"{self}: Generating TTS [{text}]")
 
+        # Clear the audio queue in case there's still audio in it, causing the next audio response
+        # to be cut off by the 'None' element returned at the end of the previous audio synthesis.
+        # Empty the audio queue before processing the new text
+        while not self._audio_queue.empty():
+            self._audio_queue.get_nowait()
+            self._audio_queue.task_done()
+
         try:
             if self._speech_synthesizer is None:
                 error_msg = "Speech synthesizer not initialized."