songsee

To generate a spectrogram, run the songsee CLI with your audio input and specify an output image file. This renders frequency content over time, producing a visual plot to analyze energy distribution across the track.

You can visualize MFCC, chroma, spectral flux, loudness, and tempogram features. These plots help analyze pitch, timbre, and onset characteristics to support music information retrieval and track comparison.

Yes, you can produce time-slice outputs to crop specific segments. This allows you to inspect and document particular moments in a track, such as analyzing a isolated guitar riff or debugging a specific onset.

You can debug synthesis pipelines by generating multi-panel outputs combining spectral flux and chroma views. Comparing these visualizations helps detect onset timing issues and explain why generated sounds feel out of sync.

Yes, generating visualizations requires the songsee CLI to render selected visualization types like mel spectrograms or MFCC plots. The tool then outputs the results as configurable image files for documentation.

The best way to compare features is generating combined multi-panel grids. This outputs multiple visualization types side-by-side, allowing you to compare mel-scaled energy views with chroma pitch content simultaneously.