Audio Watermark Encoding Using Self-Correlation for Reverberation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio watermarking technologies face challenges in detecting watermarks in audio data transmitted between devices due to reverberation effects, leading to inefficiencies in implementing unique functionalities like wakeword suppression and local signal transmission.
Innovation Solution
The system embeds audio watermarks using a bi-layer encoding structure that incorporates Eigenvectors and a sign sequence, allowing for effective detection despite reverberation through a self-correlation algorithm, enabling efficient wakeword suppression and local signal transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional audio watermarking is used, then watermark embedding is simple, but watermark detection accuracy deteriorates due to reverberation effects
Solution Approach 1:
The audio watermark is divided into multiple segments that are distributed across different time frames of the audio signal. Each segment can be independently detected and correlated, allowing the system to overcome reverberation effects that may affect individual segments. The segmentation enables robust detection by aggregating correlation results across multiple segments.
Solution Approach 2:
The system performs preliminary encoding of the watermark into the audio signal before transmission, using a known encoding scheme that incorporates redundancy and error correction capabilities. This preliminary action ensures that even when reverberation degrades the signal, the original watermark information can be recovered through correlation with the known encoded pattern.
2Adaptability or versatility
If audio watermarks are embedded to enable unique functionalities, then functionality is enhanced, but system complexity increases
Solution Approach 1:
The audio watermarking system is designed to support multiple functionalities through a universal encoding and decoding framework. The same encoding mechanism can embed watermarks for various purposes such as wakeword suppression, local signal transmission, and device identification. The decoding side can interpret different watermark types using a unified correlation-based approach, reducing the need for separate specialized systems.
Solution Approach 2:
The patent introduces an intermediary encoding layer that translates high-level functionality requirements into standardized watermark patterns. This intermediary layer handles the complexity of different functionality types by converting them into a common watermark format that can be processed by the correlation decoder, thereby shielding the overall system from excessive complexity while enabling diverse functionalities.
3Reliability
If watermark detection is performed in noisy environments, then local signal transmission is enabled, but detection reliability decreases
Solution Approach 1:
The system employs feedback mechanisms where the decoder uses the known encoding structure and correlation results to iteratively refine watermark detection. By comparing detected segments against the expected encoded pattern and using feedback from correlation strength measurements, the system can adjust detection thresholds and parameters to maintain reliability in noisy and reverberant environments.
Solution Approach 2:
The encoding scheme incorporates redundant watermark segments and excessive encoding information beyond the minimum required. This partial or excessive action ensures that even when noise and reverberation corrupt parts of the watermark signal, sufficient intact information remains for reliable detection through correlation with the known encoded pattern.
Data Source
AI summary
A system may embed audio watermarks in audio data using a sign sequence. The system may detect audio watermarks in audio data despite the effects of reverberation. For example, the system may embed multiple repetitions of an audio watermark before generating output audio using loudspeaker(s). To detect the audio watermark in audio data generated by a microphone, the system may perform a self-correlation that indicates where the audio watermark is repeated. In some examples, the system may encode the audio watermark using multiple repetitions of a multi-segment Eigenvector. Additionally or alternatively, the system may encode the audio watermark using a binary sequence of positive and negative values, which may be used as a shared key for encoding/decoding the audio watermark. The audio watermark can be embedded in output audio data to enable wakeword suppression (e.g., avoid cross-talk between devices) and/or local signal transmission between devices in proximity to each other.