Update main.py

main.py

@@ -296,18 +296,6 @@ emotion_settings = {
     "shame": {"pitch": 0.8, "speed": 0.85},
 }
 
-import librosa
-import soundfile as sf
-
-def adjust_audio_speed(audio_path, speed_factor):
-    y, sr = librosa.load(audio_path)
-    y_speeded = librosa.effects.time_stretch(y, speed_factor)
-    sf.write(audio_path, y_speeded, sr)
-
-def adjust_audio_pitch(audio_path, pitch_factor):
-    y, sr = librosa.load(audio_path)
-    y_shifted = librosa.effects.pitch_shift(y, sr, n_steps=pitch_factor)
-    sf.write(audio_path, y_shifted, sr)
 
 def emotion_aware_tts_pipeline(input_text=None, file_input=None):
     try:
@@ -331,11 +319,10 @@ def emotion_aware_tts_pipeline(input_text=None, file_input=None):
             audio_path = "output.wav"
             tts_model.tts_to_file(text=input_text, file_path=audio_path)
 
-            # Adjust pitch and speed using librosa
-            if pitch != 1.0:
-                adjust_audio_pitch(audio_path, pitch)
-            if speed != 1.0:
-                adjust_audio_speed(audio_path, speed)
+            # Adjust pitch
+            pitch_factor = (pitch - 1.0) * 12  # Convert to semitones for librosa
+            adjust_pitch(audio_path, pitch_factor)
+
 
             return f"Detected Emotion: {emotion} (Confidence: {confidence:.2f})", audio_path
         else:
@@ -344,6 +331,7 @@ def emotion_aware_tts_pipeline(input_text=None, file_input=None):
         return f"Error: {str(e)}", None
 
 
+
 # Define Gradio interface
 iface = gr.Interface(
     fn=emotion_aware_tts_pipeline,