Update app.py

app.py

@@ -10,6 +10,9 @@ from pydub import AudioSegment
 import librosa
 import soundfile as sf
 import numpy as np
+from pydub import AudioSegment
+from pydub.playback import play
+
 
 def get_silence(duration_ms=1000):
     # Create silent audio segment with specified parameters
@@ -110,15 +113,14 @@ async def generate_audio_with_voice_prefix(text_segment, default_voice, rate, pi
                     if speed_factor > 0:
                         if speed_factor < 1.0:
                             speed_factor = 1.0
-                        y, sr = librosa.load(audio_path, sr=None)
+                        #y, sr = librosa.load(audio_path, sr=None)
                         
-                        # Apply phase vocoder to stretch the audio
-                        hop_length = 512  # Hop length, you can experiment with this
-                        y_stretched = librosa.phase_vocoder(y, rate=speed_factor, hop_length=hop_length)
-                        # Reconstruct the audio using ISTFT
-                        y_reconstructed = librosa.istft(y_stretched, hop_length=hop_length)
-                        # Save the stretched audio back to a file
-                        sf.write(audio_path, y_reconstructed, sr)
+                        # Load audio file
+                        audio = AudioSegment.from_file(audio_path)
+                        # Apply time-stretching
+                        audio_stretched = audio.speedup(playback_speed=speed_factor)
+                        # Save the stretched audio
+                        audio_stretched.export(audio_path, format="mp3")
 
                         #y_stretched = librosa.effects.time_stretch(y, rate=speed_factor)
                         #sf.write(audio_path, y_stretched, sr)