What is a Spectrogram?
A spectrogram is a visual representation of audio frequencies over time. The horizontal axis shows time, the vertical axis shows frequency, and the brightness/color indicates amplitude (loudness).
Puzzle creators can embed images or text into audio files by converting visual data into specific frequency patterns. When you view the audio as a spectrogram, the hidden content becomes visible.
Famous example: Aphex Twin's "Windowlicker" contains a hidden face visible only in the spectrogram — this technique has since been adopted by puzzle creators worldwide.
Tools for Spectrogram Analysis
Free Desktop Software
Audacity (Windows, Mac, Linux)
Free, open-source audio editor with built-in spectrogram view. Most popular choice.
View → Track → Spectrogram
Sonic Visualiser (Windows, Mac, Linux)
Dedicated audio analysis tool with excellent spectrogram features and customization.
Spek (Windows, Mac, Linux)
Lightweight, fast spectrogram viewer. Just drag and drop audio files.
Online Tools
Spectrum Analyzer (academo.org)
Browser-based real-time spectrogram — works with microphone or uploaded files.
Online Spectrogram Viewer
Various free online tools — search "online spectrogram viewer" for current options.
Mobile Apps
- Spectroid (Android) — Real-time spectrogram
- SpectrumView (iOS) — Audio spectrum analyzer
Step-by-Step: Using Audacity
- 1
Open the audio file
File → Open, or drag and drop the audio file into Audacity
- 2
Switch to Spectrogram view
Click the track dropdown (▼) → Spectrogram
- 3
Adjust spectrogram settings
Track dropdown → Spectrogram Settings
- Try different window sizes (higher = better frequency resolution)
- Adjust frequency range if needed
- Try different color schemes (Grayscale often works best)
- 4
Zoom and scan
Use Ctrl+scroll to zoom. The hidden image might only appear in certain sections.
- 5
Look for patterns
Text, numbers, coordinates, or images should become visible as visual patterns in the frequency display.
Pro Tip: Hidden content is often in the higher frequencies (above 15kHz) which are inaudible to most people. Make sure your frequency range includes up to 20kHz+.
What to Look For
Text and Numbers
The most common hidden content is text — coordinates, hints, or code words rendered as visual patterns in the spectrogram.
Examples: "N 52.12345" | "LOOK UNDER BRIDGE" | "GC12345"
Images
Puzzle setters may embed actual images — maps, symbols, photos, or diagrams.
QR Codes
A QR code might be visible in the spectrogram. Take a screenshot and scan it with a QR reader app.
Patterns
- Morse code dots and dashes as short/long frequency bursts
- Binary patterns (high frequency = 1, low = 0)
- Bar codes or other scannable codes
Other Audio Puzzle Techniques
Morse Code
Audio beeps encoding Morse code are common. Listen carefully for patterns of short and long tones separated by pauses. Use our Morse Code tool to decode.
DTMF Tones
Phone keypad tones (DTMF) can encode numbers. Each digit produces a unique dual-tone. Tools like DTMF Decoder can extract the numbers from audio.
Reversed Audio
Sometimes the audio is simply played backwards. In Audacity: Effect → Reverse.
Speed Changes
Audio might be sped up or slowed down. Try adjusting playback speed if spoken words sound garbled.
Stereo Channel Separation
Different content might be in the left vs right channel. Split stereo tracks and analyze each separately.
Creating Spectrogram Art (For Cache Owners)
If you want to create your own audio spectrogram puzzles:
Free Tools
Coagula
Windows tool that converts images to audio spectrograms. Free and easy to use.
Photosounder
Professional tool for spectrogram image creation. Has a free trial.
Python + Libraries
For programmers: scipy and librosa can create custom spectrogram audio.
Troubleshooting Tips
- Can't see anything? Try different spectrogram settings — window size, frequency range, and color scheme all affect visibility.
- Image looks distorted? Adjust the aspect ratio by zooming horizontally or vertically until proportions look correct.
- Only see noise? The hidden content might be in only part of the audio — scan through the entire file.
- Low quality image? Some spectrograms are intentionally low-res. Try taking a screenshot and adjusting contrast.
- Content in high frequencies? Ensure your frequency range goes up to 20kHz+ and your audio file is high quality (not low-bitrate MP3).