How LeakGuard Works

A comprehensive guide to embedding and detecting audio watermarks using spread spectrum technology.

Embedding a Watermark

  1. Select Your Audio File

    Start by choosing the WAV file you want to protect.

    • LeakGuard supports standard WAV formats including 16-bit, 24-bit, and 32-bit float audio.
    • Files are processed entirely in your browser, so there is no file size limit imposed by upload restrictions.
    • For best results, watermark your master or final mix before any distribution.
    • The tool preserves the exact audio quality of your original file, maintaining bit depth and sample rate.

  2. Enter a Unique Identifier

    Create a distinctive identifier for this copy of your track.

    • Identifiers can be up to 32 characters long, using letters and numbers.
    • Choose identifiers that help you track recipients, such as names, dates, or codes.
    • Examples: "JOHN-REVIEWER", "ATLANTIC-A&R", "PROMO-0115", "MIXFEEDBACK".
    • The identifier is embedded invisibly into the audio signal and can be recovered later.

  3. Process and Download

    The watermark is embedded and you receive your protected file.

    • Processing happens instantly using Web Workers for optimal performance.
    • The output file maintains identical audio quality to your original.
    • The watermarked file is saved with a modified filename indicating the identifier.
    • Create multiple copies with different identifiers for different recipients.
Start Watermarking Now

Detecting a Watermark

  1. Upload the Suspected File

    Load the audio file you want to analyze for watermarks.

    • The tool analyzes WAV, MP3, and AAC files for embedded watermarks.
    • Even heavily compressed versions of your track can be analyzed.
    • The detection process works on partial clips as well as full tracks.

  2. Automatic Analysis

    LeakGuard scans the audio for embedded identifiers.

    • The detection algorithm searches for spread spectrum watermark signals.
    • Analysis completes in seconds, even for long audio files.
    • The process checks across the entire frequency spectrum for hidden data.

  3. View Results

    Get a clear answer about the watermark presence and origin.

    • If a watermark is found, the embedded identifier is displayed.
    • Results show which copy of your track was used as the source.
    • No watermark found means the file was not created with LeakGuard or was heavily processed.

Spread Spectrum Technology

LeakGuard uses spread spectrum watermarking, a technique originally developed for secure military communications. The watermark signal is spread across a wide range of frequencies, making it resistant to removal and difficult to detect without the proper decoding algorithm.

Unlike simple LSB (Least Significant Bit) steganography that embeds data in the quietest parts of a signal, spread spectrum distributes the watermark energy across the entire audio spectrum. This redundancy means the watermark survives even when portions of the frequency range are removed or modified.

The spread spectrum approach provides two key advantages. First, the watermark is imperceptible because it is spread so thin across frequencies that it falls below the threshold of human hearing. Second, it is robust because damaging any single frequency band does not destroy the entire watermark.

Surviving Compression

When audio is converted to MP3 or AAC, lossy compression removes frequencies deemed inaudible to human ears. This process destroys simple watermarks but spread spectrum watermarks are designed to survive it.

Our watermark signal is embedded primarily in perceptually significant frequency ranges, the same frequencies that compression algorithms work hardest to preserve. This means the watermark travels with the audio through format conversions.

Testing shows reliable detection after WhatsApp sharing (which uses Opus compression), Telegram transmission, and standard MP3 conversion at 128 kbps and above. The watermark can even survive multiple compression cycles.

Understanding Limitations

No watermarking technology is indestructible, and we believe in transparency about what LeakGuard can and cannot do. Understanding these limitations helps you make informed decisions about protecting your music.

Extreme audio processing will remove the watermark. This includes significant pitch shifting (more than a few semitones), substantial time stretching, and applying heavy distortion or filtering effects. However, these modifications also significantly alter the audio itself.

Re-recording from speakers or using analog capture will destroy the watermark, as will uploading to platforms like YouTube or SoundCloud that apply aggressive audio processing. The watermark is designed to survive casual sharing methods, not determined attacks.

The watermark works best on musical content with dynamic range and frequency variation. Very sparse or highly repetitive material may offer less signal for the watermark to hide in, potentially affecting detection reliability.

Complete Privacy

Every operation in LeakGuard happens locally in your browser. When you upload a file for embedding or detection, it is processed using JavaScript and Web Workers on your own computer. The audio data never leaves your device.

We do not store any information about your files, identifiers, or usage patterns. There is no cloud processing and no analytics tracking what you do with the tool. Even if our servers were compromised, there would be no user data to steal.

This architecture was a deliberate design choice. Musicians sharing pre-release material have legitimate privacy concerns, and we wanted to create a tool that eliminates the need to trust any third party with sensitive content.

Ready to Protect Your Music?

Start watermarking your tracks today. Works entirely in your browser.

Get Started Now