Sonify — turn an image into sound
Build a sound whose spectrogram is your picture. Drop an image, type a word, or draw one, hit play, and watch a live FFT pull your image back out of the noise. Download the WAV and open it in any spectrogram viewer to find it hiding there.
🔒 100% private — your image never leaves your deviceYour image — what goes in
Live spectrogram — what comes outpress play
How it works
A spectrogram is a picture of a sound: time runs left to right, pitch runs bottom to top, and brightness is how loud each pitch is at each moment. Sonify runs that idea backwards. It takes your image and treats every row of pixels as one pitch, spread across the frequency range you choose. As it sweeps across the image column by column, the brightness of each pixel becomes the loudness of its tone right then. Stack a couple hundred of these tones and play them together, and the combined sound, viewed as a spectrogram, redraws your original picture.
How does this run in a browser?
The whole thing is additive synthesis built on the Web Audio API. Sonify creates one sine oscillator per frequency line and automates each one’s volume over time from a row of your image, then renders the mix offline with an OfflineAudioContext so it computes faster than real time and never stutters. The result is normalized and encoded to a 16-bit WAV in JavaScript. When you press play, the sound is routed through an AnalyserNode, which runs a real FFT many times a second, and the frequencies it reports are painted onto the right-hand canvas — so you literally watch your image reappear out of the audio. Because every step happens on your machine, your image is never uploaded, and it all keeps working with your internet off.
Good for
Hiding a face or a message in a track the way Aphex Twin famously did, making a secret a friend can only read with a spectrogram app, sound-design textures with real spectral shape, and showing students what the Fourier view of sound actually looks like. Download the WAV and open it in Audacity (Spectrogram view), Spek, or Sonic Visualiser to reveal the picture anywhere.
Is my image uploaded anywhere?
No. Every step, from reading pixels to rendering audio, runs locally in your browser with the Web Audio API. Your image never touches a server, which is why it is instant, private, and works offline.
Why does the sound resemble noise?
Hundreds of pure tones playing at once average out to a dense, hiss-like texture for your ears. The structure is still there — it only becomes visible when you look at the frequencies over time in a spectrogram.
How do I reveal the image again?
Press play here to watch the built-in live spectrogram redraw it, or download the WAV and open it in any spectrogram viewer such as Audacity, Spek, or Sonic Visualiser.
Which settings read best?
Short words and bold, high-contrast shapes. Linear scale keeps the picture’s proportions faithful; log scale spreads out the low frequencies. A wider frequency range and more detail lines sharpen the image at the cost of a longer render.
Is it really free?
Yes, completely free, no sign-up, no watermark, and no limits.
I build a tool like this every day.
Senior full-stack engineer, available for senior or contract work, fully remote. See the rest of the lab or get in touch.