Three tools in one: generate a visual spectrogram from any audio file, convert an image into sound, or decode a spectrogram image back to audio. All in your browser — no upload, no signup.
Drop any audio file to instantly visualise its frequency content over time. Download the result as a high-resolution PNG image — perfect for sharing, analysis, or embedding in your ARG.
A spectrogram plots time (x-axis) against frequency (y-axis), with brightness representing amplitude. It reveals the frequency content of sound invisible to the ear.
AudioCipher's HZAlphabet produces a characteristic stepped-frequency pattern — each letter is a distinct horizontal band. The shape spells the message on the spectrogram.
Artists like Aphex Twin have hidden images inside music tracks. Use the Image → Sound tab to encode your own picture into audio, then view it here.
Convert any image into audio. Pixel brightness maps to frequency amplitude — brighter pixels produce louder tones at that frequency. The resulting audio reveals the image when viewed in a spectrogram. Inspired by artists like Aphex Twin who hid images inside music.
The image is scaled to a fixed height (frequency bins). Each column of pixels becomes one moment in time. Pixel brightness sets the amplitude of each frequency at that moment.
High-contrast images with clear shapes produce the most recognisable spectrograms. Simple logos, text, and line drawings work best. Busy photos become noisy audio.
Hide an image in your music, podcast, or Discord audio. Listeners who run the audio through a spectrogram analyser will see your hidden picture — a classic ARG technique.
Upload a spectrogram image and recover the audio it represents. Use this to decode images created by the Image → Sound tool, or to attempt reconstruction of audio from any spectrogram screenshot.
Works best on spectrograms created by this tool's Image → Sound panel. Screenshots of external spectrograms can be reconstructed — match the frequency range settings to the original.
The image rows map linearly from min-freq (bottom) to max-freq (top). Set these to match the spectrogram you're decoding for accurate reconstruction.
Phase information is lost in a spectrogram image — reconstructed audio approximates the original. Heavily compressed or low-resolution images produce noisier results.