High-Quality Audio Extraction from YouTube with Python

A practical guide to extracting high-quality audio from YouTube videos with Python libraries and automating the process with ffmpeg.

Today I will start to write 3 articles (script pyhton: Download , Audio to text and Dashboard ) — Tutorials with step-by-step to build a YouTube Audio Downloader & Transcriber. The first part, I will make an tutorial in python to automate the download of audio from youtube.

Let’s to describe the library that we need to do task:

Summary of libraries in my code

• sys: For command-line argument handling.
• pytubefix: For fetching and downloading YouTube audio streams.
• ffmpeg: For converting and processing multimedia files (audio conversion).
• time: For introducing delays in the execution of the script.
• datetime: For generating timestamps and working with dates/times.

Below the full code to make the download:

import sys
import pytubefix
import ffmpeg
import time
from datetime import datetime

# Get the YouTube video URL from command-line arguments
youtube_url = sys.argv[1]

# Specify the output file name for the audio
filename = f"audios/"
timestamp = datetime.now().strftime('%Y%m%d%H%M%S')
filename += f"audio_{timestamp}.wav"

time.sleep(2)

# Create a YouTube object and fetch the stream URL
print('Downloading audio from youtube...')
yt = pytubefix.YouTube(youtube_url)

print('Fetching audio stream...')
stream = yt.streams[0].url  # Get the first available stream (usually audio)

print('Use ffmpeg to convert the audio stream to a .wav file...')
# Use ffmpeg to convert the audio stream to a .wav file
ffmpeg.input(stream).output(filename, format='wav', loglevel="error").run()

# Save the filename
text_file = open(f"filename_audio.txt", "w")
text_file.write(f"{filename}")
text_file.close()

print(f"Audio downloaded and saved as {filename}")

As you read above, we have a simple script to download the audio, now it’s time to do the download!

To run…

# Terminal:
python downloadytaudio.py https://www.youtube.com/watch?v=lotOsFYpmDc&pp=ygUQZW5nbGlzaCBsZWFybmluZw%3D%3D

Good job!

If you want to download a lot of audios, you can create a list with the links of the videos and pass to the script to do the job.

Here a simple script to do the task:

import time
import subprocess

# Define the path to your text file
file_path = 'yt_links/links.txt'

# script master to run ALL the steps of the project. 
def main():

    # Open the file and read each line
    with open(file_path, 'r') as file:
        youtube_links = file.readlines()

    # Strip whitespace from each line (e.g., newline characters) and print the links
    youtube_links = [link.strip() for link in youtube_links]

    for youtube_url in youtube_links:
        start_time = time.time()
        print(f'>>> youtube_url: {youtube_url}')

        print('Downloading audio from youtube...')
        subprocess.run(['python3', 'downloadytaudio.py', youtube_url])

if __name__ == "__main__":
    main()

An example to show how it’s work, a list of youtube videos:

# your text file yt_links/links.txt:

https://www.youtube.com/watch?v=zXE_uVuNvhw
https://www.youtube.com/watch?v=_TIw_vfFXgQ
https://www.youtube.com/watch?v=elO6nmqj2Ng

and the files you will be storage on the folder:

Good! We did it!

Thanks for take the time to read the article, I will continue to write. The next part, we will transcribe the audio to text, this can be used to do Sentiment Analysis in Deep Learning or to get the main points from video using an AI.