This repository contains a Google Colab Python script designed for creative audio processing and sampling. The script, inspired by work from Colugo Music, is a versatile tool for anyone interested in generating unique sound samples from YouTube videos.
The script performs several key functions:
- Random Search Phrase Generation: It generates random search phrases from a list of words and uses these phrases to search for and download audio from YouTube.
- Audio Processing: Downloaded audio is then processed to create two types of samples: one-shot and looped.
- File Management and Versioning: Processed audio files are organized into specific directories, and a combined loop file is created without overwriting previous versions.
- Requirements: The script auto-installs several Python packages, including
pydub
,yt-dlp
,glob2
andnltk
- Google Drive Integration: The script is designed to run in a Google Colab environment and saves files directly to Google Drive. It will ask to mount your Google Drive and will create the base directory to
dodgylegally
- Folder Structure: The script automatically creates necessary directories for storing raw, processed, and combined audio files.
- Word List: A file named
birdwater.txt
is auto-generated containing a list of randomly generated dictionary words for generating search phrases. The script checks for this file and creates it if it doesn't exist.
To use the script, simply run it in a Google Colab notebook. It will prompt for the number of samples to process and then execute the following steps:
- Generates random search phrases in
birdwater.txt
and downloads corresponding audio from YouTube. - Processes the audio to create one-shot
dodgylegally/wavs/oneshot/
and loopdodgylegally/wavs/loop/
samples - Combines looped samples into a single audio file that repeats each sample 3-4 times, incrementing the version number to avoid overwriting. You can find it in
dodgylegally/wavs/processed/combined/
You can easily customize various aspects of the script, such as the batch size for downloads, the base directory in Google Drive, and the structure of the output directories.
This script was hacked together by Daniel Raffel, based on the original work shared by Colugo Music. It is intended for creative and experimental sound processing projects.
This project is open-source and based on a script by Colugo Music that has been released to the public doman. Feel free to use, modify, and distribute as desired there is no license.