Audio Watermarking Tools 1 (AWT1)
change log

Please refer to for more details about AWT1 package.



   New generation of the AWT1 technology

  • Completely re-designed encoding (watermarking) engine
  • SDK for real-time (live) audio streams watermarking is now available: low latency, all typical sampling rates are supported (from 8 up to 192 KHz), fast, low RAM; The SDK is cross-platform and is available in a form of linkable object library
  • Significantly improved watermarking transparency, new “-quality” parameter for transparency adjustment
  • Rewritten from scratch, new file encoder (watermarker): much faster files processing, low RAM consumption, support for compound payloads, adjustable watermarking transparency, adjustable output limiter
  • Improved watermark search/extraction by the decoder



   Major technology update

  • (decoder) drastically faster sync stage and immediate 'no watermark' detection in case of acoustically different files
  • (decoder) new '-fast_sync' switch, leading to twice faster synchronization (safe to use!)
  • (decoder) changed math for even faster overall decoding operation
  • (decoder) code optimization and speed-up in watermark extraction
  • (decoder) Improved non-standard waves handling
  • ARM Linux binaries are now available


  • fixed decoder in case of empty audio file


  • fixed occasional decoder crash during file conversion


  • fixed occasional decoder crash


  • several bug-fixes in the encoder and decoder, stability improvements.


  • introduced "-no_limiter" command line switch in the encoder, allowing to force normalizing the output audio instead of limiting it
  • fixed critical bug in the GUI front-end: "-auto_answer=a" switch was not specified in the decoding command line, as a result, proper decoding of cropped signals was not possible

  • improved handling of non-standard Pro Tools 10 waves
  • minor improvements

This is major release.
  • Significantly improved handling of different wave formats (WAVE_FORMAT_EXTENSIBLE), multichannel (>2) files are now fully supported
  • new -tmp_file command line parameter in decoder (allowing to define temporary file name used by FFmpeg for conversion)
  • fixed FFmpeg call
  • various optimizations and clarifications
  • error codes revesied (refer to AWT1 Guide)

  • Significant speed improvements in encoder and decoder
  • clarified WAV reading errors

  • Further code optimizations, significant speed and memory improvements

  • Further code optimizations, speed and memory improvements
  • The decoder is now faster and consumes at least twice (!) less memory

  • Code optimizations resulted in decoding speed increase up to 30%
  • New -fast_sync switch for even faster synchronization in the decoder

  • In case of possible clipping of the encoder output, the signal is now limited and isn't normalized (as previously) to prevent RMS decrease
  • Improved decoder reliability with two significant bug fixes
  • Fixed bug with -8bits switch

  • Minor bug fixes
  • Mac OS X version of binaries is available upon request

  • Convenient GUI front-end has been added for Linux and Windows platforms (with drag'n'drop support)
  • FFmpeg is now distributed as a part of AWT package (under LGPL license)

  • faster and less memory consuming decoding
  • more user friendly messages

  • twice faster encoding
  • twice smaller executables
  • twice faster synchronization in decoder
  • re-arranged exit codes (see AWT Guide)

  • memory optimizations
  • better error reporting
  • more informative messages

  • Major AWT upgrade
  • significant improvement in watermark robustness (refer to the Examples section)
  • watermark aural transparency is now even better than previously; undistinguishable watermarking to average listener on audio equipment of any quality.
  • higher watermarking rate (15 bps with 4-bytes watermarking payload and default parameters)
  • possibility of false detection significantly reduced
  • bug fix: two payloads allowed in demo encoder did not work mistakingly
  • extended AWT Guide

  • AWT decoder now supports various audio file formats via FFmpeg (Windows / Linux). If specified file(s) is not wave 44.1 KHz, the decoder will automatically attempt to convert it into wave by running FFmpeg. FFmpeg binary is expected to be located in the current directory. FFmpeg is not distributed with AWT package, you should download FFmpeg binaries by yourself. For your convenience, you can download FFmpeg binaries for Windows and Linux here, or search google.

  • AWT binaries are now available also for Linux x86 and x64 (and actually all *nix systems compatible with ELF binaries).

  • Fixed bug in the encoder: output WAV file format is now correct (previously it could prevent some players from playing the output waves correctly)
  • The encoder now notifies the user when output audio stream is normalized

  • RAM consumption of the decoder has been significantly decreased (by a factor of 4 at least)
  • decoding (extraction) speed has been significantly increased (now 2-3 times faster)
  • memory leakage fixed

  • fixed issue in decoder with long payloads (>15 bytes)

  • allowed demo-version watermarking payloads list extended for better evaluation convenience

  • first public release