| # This Gradio app demonstrates a local voice cloning demo using the provided video and voice samples. | |
| import gradio as gr | |
| import numpy as np | |
| import tempfile | |
| import os | |
| # Function to clone voice and generate dubbed video | |
| def clone_voice(video_path, voice_option): | |
| # Create a temporary directory to store intermediate files | |
| temp_dir = tempfile.mkdtemp() | |
| audio_path = os.path.join(temp_dir, "original_audio.wav") | |
| output_path = os.path.join(temp_dir, "output_video.mp4") | |
| # 1. Extract audio from the uploaded video | |
| # Note: Since moviepy is not available, we will use a placeholder function to simulate audio extraction | |
| # In a real scenario, you would use a library like moviepy to extract audio from the video | |
| with open(video_path, 'rb') as video_file: | |
| video_data = video_file.read() | |
| with open(audio_path, 'wb') as audio_file: | |
| audio_file.write(video_data) # Placeholder for audio extraction | |
| # 2. Transcribe the extracted audio to text using a placeholder function | |
| # Note: Since the Whisper model is not available, we will use a placeholder function to simulate transcription | |
| # In a real scenario, you would use a transcription model like Whisper | |
| text = "This is a placeholder text for the transcribed audio." | |
| # 3. Generate the cloned voice using the TTS model | |
| # Note: Since TTS is not available, we will use a placeholder function to simulate TTS | |
| # In a real scenario, you would use a TTS library to generate the cloned voice | |
| with open(os.path.join(temp_dir, "new_audio.wav"), 'wb') as new_audio_file: | |
| new_audio_file.write(text.encode()) # Placeholder for TTS | |
| # 4. Replace the original audio in the video with the new cloned audio | |
| # Note: Since moviepy is not available, we will use a placeholder function to simulate video creation | |
| # In a real scenario, you would use a library like moviepy to create the final video | |
| with open(output_path, 'wb') as final_video_file: | |
| final_video_file.write(video_data) # Placeholder for video creation | |
| return output_path | |
| # Create the Gradio interface | |
| with gr.Blocks() as demo: | |
| gr.Markdown("## Local Voice Cloning Demo") | |
| with gr.Row(): | |
| video_input = gr.Video(label="Upload Video", sources=["upload"]) | |
| voice_selector = gr.Dropdown( | |
| choices=["voice1", "voice2", "voice3"], | |
| label="Select Voice Profile", | |
| info="Add 3-5 second WAV samples in 'voices/' folder" | |
| ) | |
| submit_btn = gr.Button("Generate Dubbed Video") | |
| video_output = gr.Video(label="Result") | |
| submit_btn.click( | |
| fn=clone_voice, | |
| inputs=[video_input, voice_selector], | |
| outputs=video_output | |
| ) | |
| # Launch the Gradio app | |
| if __name__ == "__main__": | |
| demo.launch(show_error=True) |