Add TTS_SPEED config for speech rate control (atempo filter in ffmpeg)

tts_engine.py

@@ -92,12 +92,19 @@ def _generate_tts_audio(text, lang, mp3_path):
         return False
 
 
-def _convert_to_wav(mp3_path, wav_path, volume_db=0):
+def _convert_to_wav(mp3_path, wav_path, volume_db=0, speed=1.0):
-    _cmd = ['ffmpeg', '-y', '-i', mp3_path,
+    speed = max(0.5, min(2.0, speed))
-            '-ar', '8000', '-ac', '1', '-sample_fmt', 's16']
+    _filters = []
+    if speed != 1.0:
+        _filters.append('atempo={:.2f}'.format(speed))
+        logger.info('(TTS) Aplicando velocidad: x%.2f', speed)
     if volume_db != 0:
-        _cmd += ['-af', 'volume={}dB'.format(volume_db)]
+        _filters.append('volume={}dB'.format(volume_db))
         logger.info('(TTS) Aplicando ajuste de volumen: %ddB', volume_db)
+    _cmd = ['ffmpeg', '-y', '-i', mp3_path,
+            '-ar', '8000', '-ac', '1', '-sample_fmt', 's16']
+    if _filters:
+        _cmd += ['-af', ','.join(_filters)]
     _cmd += ['-f', 'wav', wav_path]
     try:
         result = subprocess.run(_cmd, capture_output=True, timeout=60)
@@ -298,7 +305,7 @@ def _encode_ambe_ambeserver(wav_path, ambe_path, host, port):
     return True
 
 
-def text_to_ambe(txt_path, ambe_path, language, vocoder_cmd, ambeserver_host='', ambeserver_port=2460, volume_db=0):
+def text_to_ambe(txt_path, ambe_path, language, vocoder_cmd, ambeserver_host='', ambeserver_port=2460, volume_db=0, speed=1.0):
     if not os.path.isfile(txt_path):
         logger.warning('(TTS) Archivo de texto no encontrado: %s', txt_path)
         return False
@@ -332,7 +339,7 @@ def text_to_ambe(txt_path, ambe_path, language, vocoder_cmd, ambeserver_host='',
     if not _generate_tts_audio(text, _tts_lang, _mp3_path):
         return False
 
-    if not _convert_to_wav(_mp3_path, _wav_path, volume_db):
+    if not _convert_to_wav(_mp3_path, _wav_path, volume_db, speed):
         _cleanup([_mp3_path])
         return False
 
@@ -385,6 +392,7 @@ def ensure_tts_ambe(config, tts_num):
     _ambeserver_host = config['GLOBAL'].get('TTS_AMBESERVER_HOST', '')
     _ambeserver_port = config['GLOBAL'].get('TTS_AMBESERVER_PORT', 2460)
     _volume_db = config['GLOBAL'].get('TTS_VOLUME', -3)
+    _speed = config['GLOBAL'].get('TTS_SPEED', 1.0)
 
     _txt_path = './Audio/{}/ondemand/{}.txt'.format(_lang, _file)
     _ambe_path = './Audio/{}/ondemand/{}.ambe'.format(_lang, _file)
@@ -403,7 +411,7 @@ def ensure_tts_ambe(config, tts_num):
         logger.warning('(TTS-%d) Archivo de texto no encontrado: %s', tts_num, _txt_path)
         return None
 
-    if text_to_ambe(_txt_path, _ambe_path, _lang, _vocoder_cmd, _ambeserver_host, _ambeserver_port, _volume_db):
+    if text_to_ambe(_txt_path, _ambe_path, _lang, _vocoder_cmd, _ambeserver_host, _ambeserver_port, _volume_db, _speed):
         return _ambe_path
     else:
         if os.path.isfile(_ambe_path):