tts: fix metrics and TTSStoppedFrame frame in HTTP services

Fixes #1247

CHANGELOG.md

@@ -27,6 +27,9 @@ stt = DeepgramSTTService(..., live_options=LiveOptions(model="nova-2-general"))
 
 ### Fixed
 
+- Fixed an issue that was causing HTTP TTS services to push `TTSStoppedFrame`
+  more than once.
+
 - Fixed a `FishAudioTTSService` issue where `TTSStoppedFrame` was not being
   pushed.
 

src/pipecat/services/cartesia.py

@@ -364,9 +364,6 @@ class CartesiaHttpTTSService(TTSService):
     async def run_tts(self, text: str) -> AsyncGenerator[Frame, None]:
         logger.debug(f"Generating TTS: [{text}]")
 
-        await self.start_ttfb_metrics()
-        yield TTSStartedFrame()
-
         try:
             voice_controls = None
             if self._settings["speed"] or self._settings["emotion"]:
@@ -376,6 +373,8 @@ class CartesiaHttpTTSService(TTSService):
                 if self._settings["emotion"]:
                     voice_controls["emotion"] = self._settings["emotion"]
 
+            await self.start_ttfb_metrics()
+
             output = await self._client.tts.sse(
                 model_id=self._model_name,
                 transcript=text,
@@ -386,14 +385,17 @@ class CartesiaHttpTTSService(TTSService):
                 _experimental_voice_controls=voice_controls,
             )
 
+            await self.start_tts_usage_metrics(text)
+
+            yield TTSStartedFrame()
+
             frame = TTSAudioRawFrame(
                 audio=output["audio"], sample_rate=self.sample_rate, num_channels=1
             )
+
             yield frame
         except Exception as e:
             logger.error(f"{self} exception: {e}")
-
-        await self.start_tts_usage_metrics(text)
-
-        await self.stop_ttfb_metrics()
-        yield TTSStoppedFrame()
+        finally:
+            await self.stop_ttfb_metrics()
+            yield TTSStoppedFrame()

src/pipecat/services/elevenlabs.py

@@ -566,18 +566,16 @@ class ElevenLabsHttpTTSService(TTSService):
                     return
 
                 await self.start_tts_usage_metrics(text)
+
                 yield TTSStartedFrame()
 
                 async for chunk in response.content:
                     if chunk:
                         await self.stop_ttfb_metrics()
                         yield TTSAudioRawFrame(chunk, self.sample_rate, 1)
-
-                yield TTSStoppedFrame()
-
         except Exception as e:
             logger.error(f"Error in run_tts: {e}")
             yield ErrorFrame(error=str(e))
-
         finally:
+            await self.stop_ttfb_metrics()
             yield TTSStoppedFrame()

src/pipecat/services/playht.py

@@ -397,6 +397,7 @@ class PlayHTHttpTTSService(TTSService):
             await self.start_tts_usage_metrics(text)
 
             yield TTSStartedFrame()
+
             async for chunk in playht_gen:
                 # skip the RIFF header.
                 if in_header:
@@ -416,6 +417,8 @@ class PlayHTHttpTTSService(TTSService):
                         await self.stop_ttfb_metrics()
                         frame = TTSAudioRawFrame(chunk, self.sample_rate, 1)
                         yield frame
-            yield TTSStoppedFrame()
         except Exception as e:
             logger.error(f"{self} error generating TTS: {e}")
+        finally:
+            await self.stop_ttfb_metrics()
+            yield TTSStoppedFrame()

src/pipecat/services/rime.py

@@ -387,9 +387,6 @@ class RimeHttpTTSService(TTSService):
 
         try:
             await self.start_ttfb_metrics()
-            await self.start_tts_usage_metrics(text)
-
-            yield TTSStartedFrame()
 
             async with aiohttp.ClientSession() as session:
                 async with session.post(self._base_url, json=payload, headers=headers) as response:
@@ -399,6 +396,10 @@ class RimeHttpTTSService(TTSService):
                         yield ErrorFrame(error=error_message)
                         return
 
+                    await self.start_tts_usage_metrics(text)
+
+                    yield TTSStartedFrame()
+
                     # Process the streaming response
                     chunk_size = 8192
                     first_chunk = True
@@ -411,12 +412,9 @@ class RimeHttpTTSService(TTSService):
                         if chunk:
                             frame = TTSAudioRawFrame(chunk, self.sample_rate, 1)
                             yield frame
-
-            yield TTSStoppedFrame()
-
         except Exception as e:
             logger.exception(f"Error generating TTS: {e}")
             yield ErrorFrame(error=f"Rime TTS error: {str(e)}")
-
         finally:
+            await self.stop_ttfb_metrics()
             yield TTSStoppedFrame()