carlthome / audioscrape

Scrape audio from YouTube and SoundCloud with a simple command-line interface.

Welcome to my GitHub.com profile! 👋

My name is Carl Thomé. I'm a computer scientist focused on machine learning and music technology. My primary interest in life is music and I'm a music information retrieval (MIR) enthusiast. I love digital signal processing (DSP) but I'm particularly interested in associative memory for artificial neural networks (ANNs). My current love is differentiable programming and self-supervised learning. I mistakingly think I'd like to learn category theory. Automatic differentiation go brrr.

🤔 Some personal beliefs

• Scientific knowledge should be accessible to everyone.
• Engineering is an artistic craft that deserves practice.
• Everyone can learn mathematics if it's taught right.
• Free software alternatives are important for a society.
• Collaborating beats competing.
• Prestige kills creativity.
• Formalisms are fun but secondary to intuition.
• Following interests is more important than having a career.

🍭 A little bit of gamification

audioscrape's Issues

download audio as a particular file format

It'd be great to add a tag to download the pafy audio stream as a certain file format. (e.g. audioscrape "barking dog" -ff wav) instead of the file format extension returned by the pafy stream.

Thanks for providing this tool!

Add "minimum quality" flags

Lossless only, minimum bitrate, etc.

Add spectrogram clustering of downloaded audio for anomaly detection

Add to PyPI

Add output directory flag

Add more providers

https://www.freesound.org/, for example.