Goth Dance Analyzer

Analyze and compare dance styles using MediaPipe pose estimation and FFT rhythm detection.

What It Does

• Records dancers performing a setlist
• Extracts skeleton data from video
• Calculates movement metrics (velocity, range, symmetry, etc.)
• Detects rhythm patterns using frequency domain analysis
• Generates summary scores and dancer fingerprints
• Compares dancers across songs

Setlist

Default setlist (goth classics, slow to fast):

Song	Artist	BPM
Bela Lugosi's Dead	Bauhaus	~75
Lucretia My Reflection	Sisters of Mercy	~115
Cities in Dust	Siouxsie	~120
Gallowdance	Lebanon Hanover	~138
Love Will Tear Us Apart	Joy Division	~147

Use any setlist you want — just add your music files to setlist/ and update the song list in scripts/capture.py.

Requirements

• Python 3.14+
• Webcam

Install

pip install -r requirements.txt

Scripts

1. Test Skeleton Tracking

Verify your camera and lighting work before recording.

python scripts/test_skeleton.py

• Shows live skeleton overlay
• Displays detection rate
• Press Q to quit

Aim for >90% detection rate. If lower, improve lighting or adjust camera angle.

2. Record Dance Sessions

Record dancers performing the setlist.

python scripts/capture.py

• Enter dancer name
• Select song from list
• Press SPACE to start/stop recording
• Press S to save and exit
• Press Q to quit without saving

Saves to: recordings/<dancer_name>/<dancer_name>_<song>.mp4

3. Extract Skeleton Data

Extract pose landmarks from recorded videos.

# Single video
python scripts/extract_skeleton.py recordings/alice/alice_gallowdance.mp4

# All videos in recordings folder
python scripts/extract_skeleton.py --batch recordings

Saves to: skeleton_data/<dancer_name>/<video>_skeleton.json

4. Analyze and Compare

Calculate metrics and generate comparison charts.

# Analyze all skeleton files
python scripts/analyze.py --dir skeleton_data

# Analyze single file
python scripts/analyze.py skeleton_data/alice/alice_gallowdance_skeleton.json

Outputs:

• analysis/metrics.json — raw and normalized metrics
• analysis/charts/summary.pdf — fingerprints + per-song summary scores
• analysis/charts/detailed_metrics.pdf — per-song metric breakdown
• analysis/charts/rhythm_spectrums.pdf — FFT frequency analysis

5. Export Video with Skeleton Overlay

Render skeleton and live metrics onto original footage.

# Single video
python scripts/export_video.py recordings/alice/alice_gallowdance.mp4

# All videos
python scripts/export_video.py --batch recordings

# Without metrics panel
python scripts/export_video.py --no-metrics recordings/alice/alice_gallowdance.mp4

Saves to: exports/<video_name>_skeleton.mp4

Metrics

Raw Metrics

Metric	Description
Arm Velocity	How fast arms move
Movement Range	How big/expansive gestures are
Vertical Motion	Amount of jumping/bouncing
Symmetry	Left/right mirror movement
Stillness Ratio	How often dancer pauses
Upper Body Focus	Arms vs legs emphasis
Movement BPM	Dominant rhythm frequency (via FFT)
Rhythm Strength	How pronounced the rhythm is
Rhythm Consistency	How stable the rhythm is over time

Summary Scores

Score	Formula
Energy	(Arm Velocity + Movement Range + Vertical Motion) / 3
Control	(Symmetry + Rhythm Consistency) / 2
Groove	(Rhythm Strength + Rhythm Consistency) / 2
Flow	1 - Stillness Ratio

Dancer Archetypes

Based on highest summary score:

• Energetic — high movement intensity
• Precise — controlled, symmetric movement
• Groovy — strong rhythmic patterns
• Fluid — continuous, flowing motion

Workflow

1. python scripts/test_skeleton.py — verify setup
2. python scripts/capture.py — record each dancer
3. python scripts/extract_skeleton.py --batch recordings — extract all
4. python scripts/analyze.py --dir skeleton_data — compare everyone
5. python scripts/export_video.py --batch recordings — export shareable videos

License

MIT