Audio Spectrum Visualizer
Generate a spectrogram with real FFT analysis — useful both as artwork and as a diagnostic for spotting lossy encoding.
Convert with this tool
or choose a file from your device — it is processed in this tab and never uploaded
Accepted: MP3, WAV, FLAC, M4A/AAC, OGG, Opus, video files
Files stay on your device. There is no upload, no queue and no watermark, so a conversion works even with the network switched off after the page has loaded.
Supported formats & processing
| Accepted input | MP3, WAV, FLAC, M4A/AAC, OGG, Opus, video files |
|---|---|
| Produced output | PNG spectrogram image at up to 1600 × 460 px |
| Processing | Web Audio API + canvas + native JavaScript encoders |
| Upload required | No — the file is read locally and never transmitted |
How to use this tool
- Add the audio fileIt is decoded to PCM at 44.1 kHz for the analysis.
- Choose the FFT size4096 for fine frequency detail, 1024 for speed on long files.
- Set the time columns and bandsMore columns means finer time resolution; more bands means finer frequency resolution.
- GenerateThe FFT runs across the whole file, which is the slowest step.
- Read the imageDark teal is silence, yellow and magenta are the loudest energy.
About this tool
A spectrogram plots frequency on the vertical axis against time on the horizontal, with colour showing energy. Where a waveform shows you how loud something is, a spectrogram shows you what frequencies are present — which makes it a diagnostic tool as well as a visual one.
The clearest practical use is spotting lossy encoding. An MP3 or AAC file has a hard frequency ceiling well below the Nyquist limit because the encoder discarded the high band; on a spectrogram that appears as an abrupt flat cut-off across the top. This tool runs a real FFT over the decoded samples and maps the result through a heat palette.
Features
- Real FFT analysis with a Hann window
- Adjustable FFT size from 1024 to 4096
- Logarithmic frequency band mapping for musical detail
- Heat colour mapping from silence to peak energy
- Reports the Nyquist frequency so you know the real ceiling
- Reveals the low-pass signature of lossy encoding
Frequently asked questions
How do I tell whether a file is genuinely high quality?
Look for the cut-off. Lossy encoders discard everything above a set frequency, so a 128 kbps MP3 shows a hard ceiling around 16 kHz and a 320 kbps MP3 around 20 kHz. A genuine lossless file shows content all the way to the Nyquist line at the top of the image.
Why does my spectrogram look blocky?
Time and frequency resolution trade off against each other. More columns give finer time detail but each column has fewer samples to analyse; more bands give finer frequency detail at the cost of time resolution. Adjust both and regenerate.
What is the Nyquist frequency?
Half the sample rate — 22.05 kHz for a 44.1 kHz file. Nothing above it can be represented, so the top of the image is the theoretical ceiling, not the content ceiling.
Can I use the spectrogram as artwork?
Yes, and it looks striking behind a track title. Generate at 1600 px wide with a dark background and composite it into your design.