Audio spectrogram viewer

What it does

Visualise a short audio recording as frequency energy over time and export a labelled image.

How to use it

1. Choose an audio file up to 10 MB, 30 seconds and eight channels, in a browser-supported format.
2. The first channel is analysed initially. After decoding, select another available channel or an arithmetic mono average, choose FFT size and generate again.
3. Download PNG or frequency-data JSON with the selected channel identified. Averaging opposite-polarity channels can cancel signals; compare separate channels when needed.
4. Processing stays in this browser. No sign-up or upload.

Example

A steady 1,000 Hz tone appears near 1 kHz throughout the recording.

Methodology

The first input channel is analysed by default, with an explicit choice of another decoded channel or arithmetic mono average. A Hann-window FFT reports one-sided amplitudes on linear frequency axes with a −100 dBFS floor. The selected channel and source channel count accompany PNG and JSON exports.

Limitations

10 MB, 30 seconds, 1–8 channels, FFT sizes 256–2,048 and up to 600 frames. Browser-supported audio only; decoding may resample. Decode deadline 15 seconds; FFT worker deadline 30 seconds, both cancellable. Optional mono averaging can cancel opposite-polarity signals. Not calibrated sound pressure or a diagnostic tool.

Automation

This tool runs in your browser. No upload API or MCP endpoint is available.