video-dubbing / app.py
artificialguybr's picture
Create app.py
bea60dd
raw
history blame
2.48 kB
import gradio as gr
import subprocess
import whisper
from googletrans import Translator
import asyncio
import edge_tts
import os
# Extract and Transcribe Audio
def extract_and_transcribe_audio(video_path):
ffmpeg_command = f"ffmpeg -i '{video_path}' -acodec pcm_s24le -ar 48000 -q:a 0 -map a -y 'output_audio.wav'"
subprocess.run(ffmpeg_command, shell=True)
model = whisper.load_model("base")
result = model.transcribe("output_audio.wav")
return result["text"], result['language']
# Translate Text
def translate_text(whisper_text, whisper_language, target_language):
language_mapping = {
'English': 'en',
'Spanish': 'es',
# ... (other mappings)
}
target_language_code = language_mapping[target_language]
translator = Translator()
translated_text = translator.translate(whisper_text, src=whisper_language, dest=target_language_code).text
return translated_text
# Generate Voice
async def generate_voice(translated_text, target_language):
VOICE_MAPPING = {
'English': 'en-GB-SoniaNeural',
'Spanish': 'es-ES-PabloNeural',
# ... (other mappings)
}
voice = VOICE_MAPPING[target_language]
communicate = edge_tts.Communicate(translated_text, voice)
await communicate.save("output_synth.wav")
return "output_synth.wav"
# Generate Lip-synced Video (Placeholder)
def generate_lip_synced_video(video_path, output_audio_path):
# Your lip-synced video generation code here
# ...
return "output_high_qual.mp4"
# Main function to be called by Gradio
def process_video(video, target_language):
video_path = "uploaded_video.mp4"
with open(video_path, "wb") as f:
f.write(video.read())
# Step 1: Extract and Transcribe Audio
whisper_text, whisper_language = extract_and_transcribe_audio(video_path)
# Step 2: Translate Text
translated_text = translate_text(whisper_text, whisper_language, target_language)
# Step 3: Generate Voice
loop = asyncio.get_event_loop()
output_audio_path = loop.run_until_complete(generate_voice(translated_text, target_language))
# Step 4: Generate Lip-synced Video
output_video_path = generate_lip_synced_video(video_path, output_audio_path)
return output_video_path
# Gradio Interface
iface = gr.Interface(
fn=process_video,
inputs=["file", gr.Interface.Component(type="dropdown", choices=["English", "Spanish"])],
outputs="file",
live=False
)
iface.launch()