songsee

Generate audio spectrograms and multi-panel feature visualizations from audio files.

Updated May 4, 2026

Search Tags:#cli #visualization #audio #feature extraction #spectrogram #mfcc #music analysis

Ready to run

npx skills add https://github.com/JamesFincher/gengar --skill songsee-jamesfincher

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: songsee
Source: https://github.com/JamesFincher/gengar/tree/main/skills/media/songsee
Command: npx skills add https://github.com/JamesFincher/gengar --skill songsee-jamesfincher

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of translating raw audio into visual spectrograms and audio feature plots so you can quickly understand sound characteristics without manually running complex audio analysis tools.

Core Features & Use Cases

Generate spectrograms and music feature visualizations: Create standard and musically-informed representations such as mel, chroma, MFCC, loudness, tempogram, and spectral flux.
Support multiple diagnostic views in one output: Render several visualization types together as a single multi-panel image for side-by-side comparison (e.g., melody vs. onset behavior).
Work on full tracks or specific time slices: Produce focused visuals for a particular region using start time and duration, such as analyzing a chorus or a brief event.

Quick Start

Run songsee on an audio file like track.mp3 to generate a spectrogram image you can inspect and compare against other feature outputs.

Frequently Asked Questions about songsee

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I generate an audio spectrogram from an MP3 file?▼

You can generate an audio spectrogram from an MP3 file by running a CLI tool that processes input audio and renders visual representations. Optional ffmpeg is required for decoding formats beyond WAV and MP3.

What audio features can I visualize for music analysis?▼

Audio features available for music analysis visualization include mel, chroma, MFCC, loudness, tempogram, and spectral flux. These derived feature types can be rendered together as a single multi-panel image for side-by-side comparison.

Do I need ffmpeg to extract audio features and create spectrograms?▼

You need ffmpeg to extract audio features and create spectrograms only when decoding formats beyond standard WAV and MP3 files. The core processing handles WAV and MP3 natively without requiring additional dependencies.

Can I visualize a specific time slice of a track instead of the full audio?▼

You can visualize a specific time slice of a track instead of the full audio by specifying a start time and duration parameter. This produces focused feature visualizations for analyzing particular regions like a chorus or brief audio event.

What is the best way to compare melody and onset behavior in audio tracks?▼

The best way to compare melody and onset behavior in audio tracks is by rendering multiple visualization types together as a single multi-panel image. This approach produces side-by-side diagnostic views for direct inspection of different audio feature representations.

Does songsee require any external dependencies to produce audio feature images?▼

songsee requires the songsee CLI installed via Go to produce audio feature images. It has no mandatory external dependencies, though optional ffmpeg is needed for decoding audio formats beyond WAV and MP3.