|
import gradio as gr |
|
import matplotlib.pyplot as plt |
|
import numpy as np |
|
import os |
|
import soundfile as sf |
|
import requests |
|
import librosa.display |
|
|
|
|
|
def download_file(url): |
|
file_id = url.split('/')[-2] |
|
download_url = f'https://docs.google.com/uc?export=download&id={file_id}' |
|
response = requests.get(download_url, allow_redirects=True) |
|
local_filename = url.split('/')[-1] + '.wav' |
|
open(local_filename, 'wb').write(response.content) |
|
return local_filename |
|
|
|
|
|
def main(): |
|
with gr.Blocks() as app: |
|
gr.Markdown( |
|
""" |
|
<h1><center>Audio Analyzer by Ilaria</center></h1>\n |
|
<h3><center>Help me on <a href="https://ko-fi.com/ilariaowo/shop">Ko-Fi</a>!</center></h3>\n |
|
## Special thanks to Alex Murkoff for helping me code it! |
|
#### Need help with AI? Join [AI Hub](https://discord.gg/aihub)!\n |
|
**Note**: Try to keep the audio length under **2 minutes**, |
|
since long audio files dont work well with a static spectrogram |
|
""" |
|
) |
|
|
|
with gr.Row(): |
|
image_output = gr.Image(type='filepath', interactive=False) |
|
|
|
with gr.Row(): |
|
with gr.Column(): |
|
audio_input = gr.Audio(type='filepath') |
|
create_spec_butt = gr.Button(value='Create Spectrogram And Get Info', variant='primary') |
|
|
|
with gr.Column(): |
|
output_markdown = gr.Markdown(value="", visible=True) |
|
|
|
with gr.Accordion('Audio Downloader', open=False): |
|
url_input = gr.Textbox(value='', label='Google Drive Audio URL') |
|
download_butt = gr.Button(value='Download audio', variant='primary') |
|
|
|
download_butt.click(fn=download_file, inputs=[url_input], outputs=[audio_input]) |
|
create_spec_butt.click(fn=create_spectrogram_and_get_info, inputs=[audio_input], |
|
outputs=[output_markdown, image_output]) |
|
|
|
download_butt.click(fn=download_file, inputs=[url_input], outputs=[audio_input]) |
|
create_spec_butt.click(fn=create_spectrogram_and_get_info, inputs=[audio_input], |
|
outputs=[output_markdown, image_output]) |
|
|
|
app.queue(max_size=1022).launch() |
|
|
|
def create_spectrogram_and_get_info(audio_file): |
|
plt.clf() |
|
|
|
y, sr = librosa.load(audio_file, sr=None) |
|
S = librosa.feature.melspectrogram(y=y, sr=sr, n_mels=256) |
|
log_S = librosa.amplitude_to_db(S, ref=np.max, top_db=256) |
|
plt.figure(figsize=(12, 5.5)) |
|
librosa.display.specshow(log_S, sr=sr, x_axis='time') |
|
plt.colorbar(format='%+2.0f dB', pad=0.01) |
|
plt.tight_layout(pad=0.5) |
|
plt.savefig('spectrogram.png', dpi=500) |
|
audio_info = sf.info(audio_file) |
|
bit_depth = {'PCM_16': 16, 'FLOAT': 32}.get(audio_info.subtype, 0) |
|
minutes, seconds = divmod(audio_info.duration, 60) |
|
seconds, milliseconds = divmod(seconds, 1) |
|
milliseconds *= 1000 |
|
|
|
|
|
speed_in_kbps = audio_info.samplerate * bit_depth / 1000 |
|
filename_without_extension, _ = os.path.splitext(os.path.basename(audio_file)) |
|
info_table = f""" |
|
|
|
|
|
| Information | Value | |
|
| :---: | :---: | |
|
| File Name | {filename_without_extension} | |
|
| Duration | {int(minutes)} minutes - {int(seconds)} seconds - {int(milliseconds)} milliseconds | |
|
| Bitrate | {speed_in_kbps} kbp/s | |
|
| Audio Channels | {audio_info.channels} | |
|
| Samples per second | {audio_info.samplerate} Hz | |
|
| Bit per second | {audio_info.samplerate * audio_info.channels * bit_depth} bit/s | |
|
|
|
""" |
|
|
|
|
|
return info_table, 'spectrogram.png' |
|
|
|
|
|
main() |
