ver 1.9.2

app.py

@@ -2,6 +2,7 @@ import os
 import re
 import io
 import torch
+import librosa
 import zipfile
 import requests
 import torchaudio
@@ -10,7 +11,6 @@ import gradio as gr
 from uroman import uroman
 import concurrent.futures
 from pydub import AudioSegment
-from pydub.effects import speedup
 from datasets import load_dataset
 from IPython.display import Audio
 from scipy.signal import butter, lfilter
@@ -188,10 +188,8 @@ def generate_audio_with_pause(srt_file_path):
         # Adjust audio speed by speedup
         if desired_duration < current_duration:
             speedup_factor = current_duration / desired_duration
-            audio_data_list = audio_data.tolist()
-            audio_data_list = speedup(audio_data_list, playback_speed=speedup_factor)
-            audio_data = np.array(audio_data_list)
-        
+            audio_data = librosa.effects.time_stretch(audio_data, speedup_factor)
+            
         audio_clips.append(audio_data)
 
         # Add pause

requirements.txt

@@ -9,4 +9,5 @@ librosa==0.10.0
 pydub==0.25.1
 speechbrain==0.5.16
 moviepy
-IPython
+IPython
+librosa