Camelot From YouTube

Music analysis tool for detecting key, BPM, structure, and energy at the phrase level.

Repo: https://github.com/chrisaacson69/camelot_from_youtube Links: Two Libraries Out of a Grown Tool — the extraction and its lessons (2026-10-22), DJ Set 1, Order Playlist, Set Mastering Pipeline, CyborgDJ, Multi-Genre & Wide-BPM Mixing, Well-Structured Python at Scale — this repo is its first specimen (CLI pile → GUI; restructure next), dj-score — the shared library this repo’s analysis code is migrating into; camelot stays the UI and has run on it since 2026-10-21

Overview

Python tool suite for DJ audio analysis. Started as a CLI key detector using chroma/Krumhansl profiles, expanded into a full GUI (audio_ui.py) with BPM detection, beat grid, key timeline, structural event detection (drops, builds, fills, transitions), stem separation, and loudness analysis.

Components

• camelot_from_youtube.py — CLI: single-window, consensus, and timeline key analysis
• audio_ui.py — tkinter/matplotlib GUI with playback, zoom, per-chart state, event editing
• key_detect.py — Camelot wheel mapping and key detection
• rekordbox_export.py — Export analysis data to Rekordbox XML (BPM, key, beat grid, cue points)

Key Features

• BPM detection with beat grid alignment
• Key detection per phrase with Camelot wheel mapping
• Structural event detection (energy changes, percussive fills, melodic drops)
• Stem separation (drums, bass, vocals, other)
• Rekordbox XML export with hot cues and memory cues
• Independent zoom/pan per chart view with auto-follow during playback

Dependencies

• librosa, numpy, scipy — audio analysis
• matplotlib, tkinter — GUI
• pygame — audio playback
• pyrekordbox — Rekordbox XML export
• Foreign tool: yt-dlp (YouTube audio download, separate repo)

Tags

music, python, ai