introduce PipelineParams audio input/output sample rates

src/pipecat/transports/network/fastapi_websocket.py

@@ -69,6 +69,7 @@ class FastAPIWebsocketInputTransport(BaseInputTransport):
 
     async def start(self, frame: StartFrame):
         await super().start(frame)
+        await self._params.serializer.setup(frame)
         if self._params.session_timeout:
             self._monitor_websocket_task = self.create_task(self._monitor_websocket())
         await self._callbacks.on_client_connected(self._websocket)
@@ -118,9 +119,19 @@ class FastAPIWebsocketOutputTransport(BaseOutputTransport):
         self._websocket = websocket
         self._params = params
 
-        self._send_interval = (self._audio_chunk_size / self._params.audio_out_sample_rate) / 2
+        # write_raw_audio_frames() is called quickly, as soon as we get audio
+        # (e.g. from the TTS), and since this is just a network connection we
+        # would be sending it to quickly. Instead, we want to block to emulate
+        # an audio device, this is what the send interval is. It will be
+        # computed on StartFrame.
+        self._send_interval = 0
         self._next_send_time = 0
 
+    async def start(self, frame: StartFrame):
+        await super().start(frame)
+        await self._params.serializer.setup(frame)
+        self._send_interval = (self._audio_chunk_size / self.sample_rate) / 2
+
     async def process_frame(self, frame: Frame, direction: FrameDirection):
         await super().process_frame(frame, direction)
 
@@ -136,7 +147,7 @@ class FastAPIWebsocketOutputTransport(BaseOutputTransport):
 
         frame = OutputAudioRawFrame(
             audio=frames,
-            sample_rate=self._params.audio_out_sample_rate,
+            sample_rate=self.sample_rate,
             num_channels=self._params.audio_out_channels,
         )