Core audio watermarking software technologies
Audio Watermarking Tools is a family of software digital audio watermarking core DSP engines. The family consists of several different technologies, each having its unique features and capabilities, thus making them applicable in different applications and use-cases.
AWT1 |
Audio Watermarking Tools 1 (AWT1) — a watermarking, "digital signing" solution that implements a so-called "strict watermarking" approach in the sense that the non-watermarked source audio data is required to find and decode the watermark in the watermarked recording. The watermark extraction is performed by “comparing” the source data with the watermarked stream. A simple and effective watermarking solution that demonstrates almost insane watermark robustness. |
AWT2 |
Audio Watermarking Tools 2 (AWT2) — a flagship forensic watermarking and "digital signing" solution of the AWT family. AWT2 implements a so called "blind watermarking" approach in the sense that the watermark can be extracted directly from the watermarked audio recording, the source audio is not required for watermark detection and extraction. AWT2 offers industrial-grade performance, features, robustness and data rate. Designed for anti-piracy, monitoring, tracking and forensic watermarking purposes over digital and analog audio channels. |
AWT3 |
Audio Watermarking Tools 3 (AWT3) — an audio trigger and watermarking solution specifically designed for inaudible and reliable far-field, over-the-air digital signaling in applications such as kick-shopping, interactive museum tours and the like. Details and a demo, including an Android app allowing to detect watermarks over-the-air, are available upon request. |
AWT4 |
Audio Watermarking Tools 4 (AWT4) — an audio triggering, signaling and watermarking solution providing semi-inaudible time-precise "digital stamping" with a high data rate. Enables placing/extracting capacious digital codes in audio streams at precise points in time to trigger time- and data-specific events with very high temporal accuracy (~30ms). |
AWT5 |
Audio Watermarking Tools 5 (AWT5)
— an audio triggering and watermarking solution specifically designed for over-the-air (loudspeaker-to-microphone) triggering in applications requiring transmission of lossy-compressed audio and over-the-air triggering in reverberant, real-world environments such as in living rooms / halls via FM radio, TV broadcasts, etc. on the user's mobile device. AWT5 applications include "second screen" experiences, kick-shopping, forensic watermarking, identification and more, including triggering with high temporal accuracy. Details and a demo, including Android app allowing to detect watermarks over-the-air, are available upon request. |
AWT0 |
Audio Watermarking Tools 0 ("AWT-Zero") — inaudible signaling via lossless audio PCM channels. Allows watermarking of extremely short wave files and triggering over completely lossless audio channels. |
Refer to a table below containing a detailed comparison of the 5 watermarking solutions.
Core audio fingerprinting and acoustic matching software technology
AFT |
Audio Fingerprinting Tools (AFT) — robust audio (acoustic) fingerprinting solution. Content matching, identification and monitoring by means of acoustic hashing. |
Web-services
TrustedAudio.com — secure, watermarked audio delivery platform. An easy and reliable way to share, deliver, and distribute watermarked audio files
TrustedAudio offers manual operation via web interface as well as automation via API.
Visit TrustedAudio.com...
AudioTag.info — a free music recognition robot with multi-million songs database:
AudioTag.info is a free music recognition robot offering its service as a web-application and via API. AudioTag can be used commercially.
Visit AudioTag.info...
Detailed comparison of the core watermarking AWT solutions
| Applications and uses | AWT1 | AWT2 | AWT3 | AWT4 | AWT5 | AWT0 |
| Robust watermarking designed specifically for forensic purposes and anti-piracy | Yes | Yes | No | No | Yes | Yes |
| General purpose watermarking and digital signing enabling authentication, anti-piracy, back-tracing and forensic uses | Yes | Yes | Partially | Partially | Yes | Yes |
| Data-hiding, data-transmission mean for general purposes | Yes | Yes | Yes | Yes | Yes | Yes (via lossless PCM channel only) |
| Reliable over-the-air sound transmission and triggering (e.g. for second-screen uses) | No | No | Yes | No | Yes | No |
| Live broadcast monitoring (easy/fast automated watermarks search) | Partially | Yes | Yes | Yes | Yes | Yes (via lossless PCM channel only) |
| Events triggering by signaling via audio stream | No | Yes | Yes | Yes | Yes | Yes (via lossless PCM channel only) |
| Algorithm characterization | AWT1 | AWT2 | AWT3 | AWT4 | AWT5 | AWT0 |
| Decoder needs source (not watermarked) audio to find/extract watermark | Yes | No | No | No | No | No |
| Overal watermarks robustness | Very high, insane | Very high | High | High | Very high | Minimal |
| Watermarks are inaudible | Yes | Yes | Yes | Almost | Yes | Yes |
| Watermarks survive wave editing (cutting, cropping, padding) | Yes | Yes | Yes | Yes | Yes | Yes |
| Watermarks survive time-stretching (playback speed variation) | No | Yes | Partially | Yes | Yes | No |
| Watermarks survive mixing | Yes | Partially | Yes | Partially | Partially | No |
| Watermarks survive lossy transcoding (MP3, Youtube, etc.) | Yes | Yes | Yes, with proper settings | Yes | Yes | No |
| Watermarks survive bandwidth limiting | Yes, fully | Yes, fully | Typically not | Yes, with proper settings | Yes | N/A |
| Watermarks survive FM radio transmission | Yes | Yes | Typically not | Yes, with proper settings | Yes | N/A |
| Watermarks survive transmission over telephone lines | Yes | Yes | No | Yes, with proper settings | Yes | N/A |
| Watermarks reliably survive air-transducing (sound transmission over the air, loudspeaker -> microphone) | Yes | Partially (on short distances) | Yes (even on long distances) | Partially (with proper settings) | Yes | No |
| Watermarks can be used for over-the-air triggering in reverberant environments | No | No | Yes (even on long distances) | Partially (with proper settings) | Yes | No |
| Very low-power watermark detection | No | No | Yes | No | Yes | No |
| Watermarks are fully secret and cannot be detected by expert signal inspection | Yes | Yes | Partially | Partially | Partially | Yes |
| Minimal watermarkable audio duration | ~ 5 sec | ~ 5 sec | ~ 3 sec | ~ 3 sec | ~ 1 sec | ~ 20 ms |
| Watermark payload length | Up to 20 bytes* | Up to 120 bytes* | Up to 30 bytes* | Up to 48 bytes* | Up to 20 bytes* | Up to 128 bytes* |
| Multi-layer watermarking capabilities (multiple independent watermarks in one stream) | No | Yes, dual-layer | Partially | Yes, multiple layers | Yes | No |
| Time-accurate watermark placement and detection | No | ~8 sec precision | ~150 ms precision | ~30 ms precision | Up to ~150 ms precision | Up to ~5 ms precision |
| Default/typical watermark carrier frequency region | Low+mid range, fixed | Low+mid range, fixed | Ultra-sonic, configurable | Fully configurable | Low+mid range, ultrasonic, flexible | N/A (not frequency-based) |
| Types of AWT-watermarks that can co-exist in the same stream | AWT3, AWT4 | AWT3, AWT4 | AWT1, AWT2, AWT4 | AWT1, AWT2, AWT3 | AWT1, AWT2, AWT3, AWT4 | N/A |
| Encoding speed | Very high | Very high | High | Very high | Very high | Very high |
| Watermarking domain | Uncompressed | Uncompressed | Uncompressed | Uncompressed | Uncompressed | Uncompressed |
| Software and Deliverables | AWT1 | AWT2 | AWT3 | AWT4 | AWT5 | AWT0 |
| Free evaluation package available | Yes | Yes | Yes | Yes | Yes | Yes |
| Cross-platform software (Windows, Linux, Mac OS, mobiles) | Yes | Yes | Yes | Yes | Yes | Yes |
| Command line (CLI) file-to-file tools for easy automation and scripting on servers | Yes | Yes | Yes | Yes | Yes | Yes |
| SDK available | Yes | Yes | Yes | Yes | Yes | Yes |
| SDK with real-time audio streams encoding & decoding capabilities | Yes | Yes | Yes | Yes | Yes | Yes |
| Detailed documentation | Yes | Yes | Yes | Yes | Yes | Yes |
| Android app available | No | Yes | Yes | No | Yes | No |
(*) subject to limitations in different packages
