transcribe music

Transcribe will chop the music up into time slices and detect the pitch of the raw audio data of each slice using the McLeod pitch method.

(transcribe-venv) sevagh:transcribe $ ./transcribe.py \
	/home/sevagh/repos/transcribe/'Guitar Tuning Standard EADGBE-bKS_m7JObxg.m4a'

System dependencies

This project uses Conda for development. On Fedora: sudo dnf install cairo-devel libffi-devel python3-tkinter ffmpeg ffmpeg-devel

xar

Additional system dependencies for playing around with https://github.com/facebookincubator/xar: sudo dnf install squashfs-tools squashfuse

Additional reading

• Speeding up real-time pitch detection with FFT autocorrelation
• Speeding up numpy-based pitch detection with numba and scipy
• Packaging Python code with XAR