added timing logs for cartesia, deepgram, elevenlabs

src/pipecat/services/cartesia.py

@@ -3,6 +3,7 @@
 #
 # SPDX-License-Identifier: BSD 2-Clause License
 #
+import time
 
 from cartesia.tts import AsyncCartesiaTTS
 
@@ -40,6 +41,8 @@ class CartesiaTTSService(TTSService):
             logger.error(f"Cartesia initialization error: {e}")
 
     async def run_tts(self, text: str) -> AsyncGenerator[Frame, None]:
+        start_time = time.time()
+        ttfb = None
         logger.debug(f"Generating TTS: [{text}]")
 
         try:
@@ -52,6 +55,9 @@ class CartesiaTTSService(TTSService):
             )
 
             async for chunk in chunk_generator:
+                if ttfb is None:
+                    ttfb = time.time() - start_time
+                    logger.debug(f"TTS ttfb: {ttfb}")
                 yield AudioRawFrame(chunk["audio"], chunk["sampling_rate"], 1)
         except Exception as e:
             logger.error(f"Cartesia exception: {e}")

src/pipecat/services/deepgram.py

@@ -5,6 +5,7 @@
 #
 
 import aiohttp
+import time
 
 from typing import AsyncGenerator
 
@@ -30,6 +31,8 @@ class DeepgramTTSService(TTSService):
         self._aiohttp_session = aiohttp_session
 
     async def run_tts(self, text: str) -> AsyncGenerator[Frame, None]:
+        start_time = time.time()
+        ttfb = None
         logger.debug(f"Generating TTS: [{text}]")
 
         base_url = "https://api.deepgram.com/v1/speak"
@@ -46,6 +49,9 @@ class DeepgramTTSService(TTSService):
                     return
 
                 async for data in r.content:
+                    if ttfb is None:
+                        ttfb = time.time() - start_time
+                        logger.debug(f"TTS ttfb: {ttfb}")
                     frame = AudioRawFrame(audio=data, sample_rate=16000, num_channels=1)
                     yield frame
         except Exception as e:

src/pipecat/services/elevenlabs.py

@@ -5,6 +5,7 @@
 #
 
 import aiohttp
+import time
 
 from typing import AsyncGenerator
 
@@ -32,6 +33,8 @@ class ElevenLabsTTSService(TTSService):
         self._model = model
 
     async def run_tts(self, text: str) -> AsyncGenerator[Frame, None]:
+        start_time = time.time()
+        ttfb = None
         logger.debug(f"Generating TTS: [{text}]")
 
         url = f"https://api.elevenlabs.io/v1/text-to-speech/{self._voice_id}/stream"
@@ -56,5 +59,8 @@ class ElevenLabsTTSService(TTSService):
 
             async for chunk in r.content:
                 if len(chunk) > 0:
+                    if ttfb is None:
+                        ttfb = time.time() - start_time
+                        logger.debug(f"TTS ttfb: {ttfb}")
                     frame = AudioRawFrame(chunk, 16000, 1)
                     yield frame