✨ (config.py): add new configuration options for lip-sync optimization, context adaptation, and additional context to enhance translation accuracy

♻️ (stt.py): increase default max buffer size from 5MB to 20MB to accommodate larger audio data
♻️ (stt.py): simplify audio sending logic by removing chunking and sending the entire buffered audio at once for improved performance

src/pipecat/services/gladia/config.py

@@ -74,11 +74,17 @@ class TranslationConfig(BaseModel):
         target_languages: List of target language codes for translation
         model: Translation model to use ("base" or "enhanced")
         match_original_utterances: Whether to align translations with original utterances
+        lipsync: Whether to enable lip-sync optimization for translations
+        context_adaptation: Whether to enable context-aware translation adaptation
+        context: Additional context to help with translation accuracy
     """
 
     target_languages: Optional[List[str]] = None
     model: Optional[str] = None
     match_original_utterances: Optional[bool] = None
+    lipsync: Optional[bool] = None
+    context_adaptation: Optional[bool] = None
+    context: Optional[str] = None
 
 
 class RealtimeProcessingConfig(BaseModel):

src/pipecat/services/gladia/stt.py

@@ -197,7 +197,7 @@ class GladiaSTTService(STTService):
         params: Optional[GladiaInputParams] = None,
         max_reconnection_attempts: int = 5,
         reconnection_delay: float = 1.0,
-        max_buffer_size: int = 1024 * 1024 * 5,  # 5MB default buffer
+        max_buffer_size: int = 1024 * 1024 * 20,  # 20MB default buffer
         **kwargs,
     ):
         """Initialize the Gladia STT service.
@@ -207,8 +207,7 @@ class GladiaSTTService(STTService):
             url: Gladia API URL
             confidence: Minimum confidence threshold for transcriptions
             sample_rate: Audio sample rate in Hz
-            model: Model to use ("solaria-1", "solaria-mini-1", "fast",
+            model: Model to use ("solaria-1")
-                or "accurate")
             params: Additional configuration parameters
             max_reconnection_attempts: Maximum number of reconnection attempts
             reconnection_delay: Initial delay between reconnection attempts (exponential backoff)
@@ -507,16 +506,9 @@ class GladiaSTTService(STTService):
     async def _send_buffered_audio(self):
         """Send any buffered audio after reconnection."""
         async with self._buffer_lock:
-            if self._bytes_sent < len(self._audio_buffer):
+            if self._audio_buffer:
-                buffered_data = self._audio_buffer[self._bytes_sent :]
+                logger.info(f"Sending {len(self._audio_buffer)} bytes of buffered audio")
-                if buffered_data:
+                await self._send_audio(bytes(self._audio_buffer))
-                    logger.info(f"Sending {len(buffered_data)} bytes of buffered audio")
-                    # Send in chunks to avoid overwhelming the connection
-                    chunk_size = 16384  # 16KB chunks
-                    for i in range(0, len(buffered_data), chunk_size):
-                        chunk = buffered_data[i : i + chunk_size]
-                        await self._send_audio(bytes(chunk))
-                        await asyncio.sleep(0.01)  # Small delay between chunks
 
     async def _send_stop_recording(self):
         if self._websocket and not self._websocket.closed: