Update app.py

app.py

@@ -141,9 +141,6 @@ def generate_speech(text, voice1, voice2, temperature, top_p, repetition_penalty
         return None
     
     try:
-        music = AudioSegment.from_mp3("Maiko-intro-outro.mp3")
-        music = music.set_frame_rate(24000)
-        
         progress(0.1, "Processing text...")
         lines = text.split('\n')
         audio_samples = []
@@ -179,24 +176,15 @@ def generate_speech(text, voice1, voice2, temperature, top_p, repetition_penalty
         
         final_audio = np.concatenate(audio_samples)
         
-        speech_audio = AudioSegment(
-            final_audio.tobytes(),
-            frame_rate=24000,
-            sample_width=final_audio.dtype.itemsize,
-            channels=1
-        )
-        
-        combined_audio = music + speech_audio + music
-        
-        combined_numpy = np.array(combined_audio.get_array_of_samples(), dtype=np.float32)
-        
-        combined_numpy = np.int16(combined_numpy / np.max(np.abs(combined_numpy)) * 32767)
+        # Normalize the audio
+        final_audio = np.int16(final_audio / np.max(np.abs(final_audio)) * 32767)
         
+        # Trim to 15 seconds if longer
         max_samples = 24000 * 15
-        if len(combined_numpy) > max_samples:
-            combined_numpy = combined_numpy[:max_samples]
+        if len(final_audio) > max_samples:
+            final_audio = final_audio[:max_samples]
         
-        return (24000, combined_numpy)
+        return (24000, final_audio)
     except Exception as e:
         print(f"Error generating speech: {e}")
         return None