NSect's picture
Duplicate from balacoon/voice_conversion_service
d673948
"""
Copyright 2023 Balacoon
Voice Conversion service interactive demo
"""
import glob
import logging
import os
import gradio as gr
from vc_service_request import vc_service_request
script_dir = os.path.dirname(os.path.abspath(__file__))
def main():
logging.basicConfig(level=logging.INFO)
badges = """
<div style="display: flex">
<span style="margin-right: 5px">
[<img src="https://play.google.com/intl/en_us/badges/static/images/badges/en_badge_web_generic.png" width="200" height="77">](https://play.google.com/store/apps/details?id=com.app.vc)
</span>
</div>
"""
with gr.Blocks() as demo:
gr.Markdown(
"""
<h1 align="center">Balacoon🦝 Voice Conversion</h1>
Welcome to the live demo of Balacoon's Voice Conversion service.
Check out our [website](https://balacoon.com/demo/#voice-conversion)
to learn more.
Voice Conversion allows you to transform your own voice
into the voice of another person using just a single sample.
For optimal results, we recommend using clean audio files in English.
Here's how it works:
1. Begin by recording your voice.
2. Select an audio sample that represents the target voice you want to convert to.
3. Click the "Convert" button and listen to the result!
If providing your own audio files, please use WAVE PCM.
Service works with 16kHz, 16 bit, mono audio.
If you are interested to plug in Voice Conversion
service into your own application, don't hesitate to get in touch with us at
[contact@balacoon.com](mailto:contact@balacoon.com)
"""
)
gr.Markdown(badges)
with gr.Row():
with gr.Column(variant="panel"):
src_audio_mic = gr.Audio(source="microphone", label="Record your voice")
src_audio_file = gr.Audio(
source="upload", label="Or upload audio to convert"
)
with gr.Column(variant="panel"):
tgt_audio_file = gr.Audio(
source="upload", label="Select audio with target voice"
)
tgt_examples_paths = glob.glob(
os.path.join(script_dir, "references", "*.wav")
)
gr.Examples(
tgt_examples_paths,
inputs=[tgt_audio_file],
)
with gr.Row():
convert_btn = gr.Button("Convert")
with gr.Row():
result_audio = gr.Audio()
def voice_conversion(src_from_mic_, src_from_file_, tgt_from_file_):
"""
helper function which checks where source come from
"""
src_ = None
if src_from_mic_:
src_ = src_from_mic_
elif src_from_file_:
src_ = src_from_file_
tgt_ = tgt_from_file_
if not src_ or not tgt_:
logging.warning("source or target are not provided")
return
return vc_service_request(src_, tgt_)
convert_btn.click(
voice_conversion,
inputs=[src_audio_mic, src_audio_file, tgt_audio_file],
outputs=result_audio,
)
demo.queue(concurrency_count=1).launch()
if __name__ == "__main__":
main()