Update app.py

app.py

@@ -170,10 +170,10 @@ def translate_local(text_to_translate, target_lang='ta-IN', device=None):
     
     return translated_text
 
-def synthesize_speech(synth_text, target_lang, gender="Male", pace="normal", output_path="temp_audio_synthesized.wav", device="cpu"):
-
-    ref_audio_path = str('reference/'+target_lang.upper()+'/'+gender.upper()+'_'+pace.upper()+'.wav')
-    ref_text_path = str('reference/'+target_lang.upper()+'/'+gender.upper()+'_'+pace.upper()+'.txt')
+def synthesize_speech(synth_text, target_lang, pace="normal", output_path="temp_audio_synthesized.wav", device="cpu"):
+    
+    ref_audio_path = str('reference/TAMIL/MALE_'+pace.upper()+'.wav')
+    ref_text_path = str('reference/TAMIL/MALE_'+pace.upper()+'.txt')
 
     ref_audio_path = ref_audio_path
     with open(ref_text_path, encoding='utf-8') as f:
@@ -260,7 +260,7 @@ def main_run(video_path,target_lang,user_transcript=None, user_translation=None)
     else:
         translated_text = translate_local(original_text,target_lang)
         print(f"Translated Text: {translated_text}")
-    translated_audio = synthesize_speech(translated_text, target_lang, "Male", pace)
+    translated_audio = synthesize_speech(translated_text, target_lang, pace)
     synced_translated_audio = match_audio_duration(original_audio_file, translated_audio) 
     cloned_synced_translated_audio = clone_voice(synced_translated_audio, original_audio_file)
     final_video_nobgm = merge_audio_video(video_path, cloned_synced_translated_audio)
@@ -280,7 +280,7 @@ def audio_pipeline_run(audio_path,target_lang,user_transcript=None, user_transla
     else:
         translated_text = translate_local(original_text,target_lang)
         print(f"Translated Text: {translated_text}")
-    translated_audio = synthesize_speech(translated_text, target_lang, "Male", pace)
+    translated_audio = synthesize_speech(translated_text, target_lang, pace)
     synced_translated_audio = match_audio_duration(audio_path, translated_audio) 
     cloned_synced_translated_audio = clone_voice(synced_translated_audio, audio_path)
     print(f"✅ Pipeline finished")