music-video-creator-showcase

Music Video Creator

Transform audio into stunning visualizations

---

Overview

Music Video Creator is a Python framework that transforms WAV audio files into professional music visualization videos. Using STFT (Short-Time Fourier Transform) analysis and GPU-accelerated rendering, it extracts audio features and maps them to dynamic visual effects.

Key Capabilities:

• 19 visual styles (9 Matplotlib + 10 GPU/ModernGL including 3D raymarching)
• Automatic scene detection (intro, verse, build, drop, breakdown, outro)
• Beat-synchronized effects (zoom, flash, shake)
• Genre presets (Electronic, Rock, Pop, Ambient, Blues)
• YouTube VP9 optimization (1440p upscaling)
• GIF output with platform presets (Discord, Twitter, etc.)
• Whisper-powered lyrics transcription and subtitle burn-in

---

Visual Styles

GPU-Accelerated (Fast)

Style	Description
sphere_3d_gpu	3D audio-reactive sphere with noise displacement
terrain_3d_gpu	Synthwave terrain flyover
tunnel_3d_gpu	Infinite tunnel with bass-reactive walls
reactive_disk_gpu	Bass-reactive spinning disk
particles_gpu	GPU particle system
frequency_bars_gpu	GPU spectrum analyzer bars
circular_gpu	GPU radial spectrum
waveform_gpu	GPU waveform display
spectrum_texture_gpu	128-bin spectrum texture mapping
sphere_3d_calm_gpu	Calm sphere (reduced reactivity)

Matplotlib (Detailed)

Style	Description
waveform_spectrogram	Combined waveform + spectrogram
spectrogram_only	Spectrogram with playhead
waveform_only	Waveform with playhead
frequency_bars	Classic spectrum analyzer
particles	Audio-reactive particle system
geometric	Pulsing polygons and circles
circular_spectrum	Radial spectrum analyzer
beat_pulse	Rhythm-synced pulsing effects
beat_rings	Expanding rings on beats

---

Sample Output

3D Sphere with Lyrics

Terrain Flyover

Reactive Disk

Style Mashup

---

Architecture

WAV Input
    │
    ▼
┌─────────────────┐
│  Audio Loader   │  Load, normalize, mono conversion
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  STFT Analysis  │  Short-Time Fourier Transform
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│    Features     │  RMS, bass/mid/high energy,
│   Extraction    │  spectral centroid, onset strength
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│ Beat Detection  │  Onset detection + peak picking
│ Scene Detection │  Auto-segment: intro→verse→drop→outro
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Visualization  │  19 styles (Matplotlib or GPU/ModernGL)
│    + Effects    │  Zoom, flash, shake on beats
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│   Rendering     │  FFmpeg pipe (memory-efficient)
│                 │  Optional: 1440p upscale for YouTube VP9
└────────┬────────┘
         │
         ▼
    MP4/GIF Output

---

CLI Examples

# Basic rendering with GPU style
python main.py song.wav -o output.mp4 --style sphere_3d_gpu --pipe

# YouTube-optimized (1440p upscale forces VP9 codec)
python main.py song.wav -o output.mp4 --youtube --style reactive_disk_gpu

# Scene-based rendering with genre preset
python main.py song.wav -o output.mp4 --scenes --genre electronic --pipe

# With beat effects
python main.py song.wav -o output.mp4 --beat-zoom --beat-flash --style terrain_3d_gpu

# GIF for Discord
python main.py song.wav -o preview.gif --gif --gif-preset discord --max-duration 10

# With lyrics (Whisper transcription)
python main.py song.wav -o output.mp4 --lyrics --style sphere_3d_gpu

---

Performance

Renderer	Style Type	Speed	Memory
FFmpeg Pipe	GPU	~30 fps	Low (streaming)
FFmpeg Pipe	Matplotlib	~0.3 fps	Low (streaming)
Standard	GPU	~25 fps	Medium
Standard	Matplotlib	~0.2 fps	High (RAM)

GPU styles are ~100x faster than Matplotlib styles.

---

Interactive Demos

Explore the capabilities interactively:

• Style Showcase - Browse all 19 visual styles with filtering
• Audio Analysis - STFT pipeline and feature extraction deep dive
• Scene Detector - Visualize automatic scene detection
• Beat Effects - Configure zoom, flash, and shake effects
• Genre Presets - Explore 5 genre configurations
• CLI Builder - Build commands interactively

---

Documentation

• Architecture - System design and modules
• Features - Complete feature reference
• Audio Analysis - STFT and feature extraction
• Usage Guide - CLI reference and examples

---

License

All Rights Reserved. This showcase is for demonstration purposes only. The source code is not publicly available.

---

About

Music Video Creator was developed to provide musicians with professional-quality visualizations for their audio content. It’s optimized for YouTube publishing with automatic VP9 codec selection and supports various social media platforms through GIF presets.

Tech Stack: Python, NumPy, SciPy, Matplotlib, ModernGL, FFmpeg, faster-whisper