Music Recognition

メディア

スペクトログラムと分析ツールでオーディオを可視化

できること

  • 9 visualization types — Spectrogram, mel, chroma, HPSS (harmonic-percussion separation), self-similarity, loudness, tempogram, MFCC, spectral flux
  • Multi-panel layouts — Combine multiple visualizations in a single grid image
  • Time slicing — Analyze specific sections with start time and duration controls
  • Color palettes — Classic, magma, inferno, viridis, or grayscale
  • Advanced controls — FFT window/hop size, frequency range filtering, output dimensions
  • Flexible input — Process audio files directly or pipe via stdin
  • Output formats — PNG or JPG with configurable resolution

    • 試してみる質問

  • "Generate a spectrogram of this song with the viridis palette"
  • "Show harmonic vs percussion separation for the first 30 seconds"
  • "Create a 4-panel grid: spectrogram, chroma, loudness, and tempogram"
  • "Analyze the self-similarity matrix to find verse/chorus structure"
  • "Show spectral flux for the drum track in magma colors"
  • "Export a 1920x1080 mel spectrogram"

    • プロのコツ

  • Chroma visualizations reveal chord progressions and harmonic structure
  • HPSS separates harmonic and percussive elements — great for mix analysis
  • Self-similarity matrices automatically highlight repeated sections (verse, chorus)
  • Multi-panel analysis combines spectral, rhythmic, and harmonic views
  • Frequency filtering lets you zoom into bass, mids, or highs specifically
  • Native WAV/MP3 decode; other formats use ffmpeg if available