songsee

To generate an audio spectrogram, run the songsee binary on your audio file to produce a single-panel PNG visualization. It validates input paths before processing and supports reading from files or stdin.

Beyond a standard spectrogram, you can visualize mel, chroma, and MFCC audio features. The tool generates multi-panel visualizations to render these different audio characteristics simultaneously from your input file.

You need ffmpeg only for non-native audio formats. The tool requires a Go-installed songsee binary for core processing, while ffmpeg handles format conversion for unsupported inputs before spectrogram generation.

Spectrogram visualization supports podcast quality control by rendering audio content insights. You can analyze frequency patterns and features across specific time ranges to identify issues in your recordings.

The best way to compare MFCC and mel features is generating multi-panel visualizations from the same audio file. This produces side-by-side PNG outputs showing different feature extractions across your specified time ranges.

Spectrogram generation fails when input paths are invalid or non-native formats lack ffmpeg for conversion. The tool validates paths before processing and requires proper format support to render spectrograms successfully.