Audio Watermarking via Phase Constellation Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio watermarking methods face challenges in robustly embedding and extracting data from audio signals, particularly under conditions of noise and compression, such as those encountered in broadcast and storage applications, where the embedded data can be overwhelmed by noise and distortion, leading to errors in decoding.
Innovation Solution
The proposed audio watermarking system employs a modulator/encoder that modifies the phase values of frequency coefficients in an audio signal using a phase constellation, allowing for error correction and iterative processing to embed and extract information bits, while minimizing audible distortion by varying the phase constellation size based on frequency sensitivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If phase values of frequency coefficients are modified to embed watermark data, then data embedding capacity is improved, but audible distortion increases
Solution Approach 1:
The patent applies different phase constellation sizes to different frequency bands based on human auditory sensitivity. Specifically, smaller phase constellation sizes (e.g., 2-4 phases) are used in frequency ranges where phase distortion is more audible (typically lower frequencies), while larger phase constellation sizes (e.g., 8-16 phases) are used in frequency ranges where phase distortion is less noticeable (typically higher frequencies). This local adaptation of modulation depth resolves the contradiction by embedding sufficient data while maintaining audio quality in critical frequency regions.
2Productivity
If larger phase constellation sizes are used to increase data capacity, then watermark embedding rate is improved, but phase distortion increases
Solution Approach 1:
The patent dynamically adjusts the phase constellation size based on frequency-dependent auditory sensitivity characteristics. The system transitions from static watermarking approaches to dynamic adaptation, where the modulation depth (phase constellation size) varies continuously across the frequency spectrum according to human perception thresholds. This dynamic approach enables higher embedding rates in less sensitive frequency regions while maintaining low distortion in sensitive regions.
3Reliability
If phase modification is applied to all frequency coefficients, then data embedding robustness is improved, but audio quality deteriorates
Solution Approach 1:
The patent selectively applies phase modification only to frequency coefficients within specific frequency bands that are less sensitive to phase distortion. By identifying and targeting these specific frequency regions rather than applying uniform modification across all frequencies, the system achieves robust watermark embedding while preserving audio quality in sensitive frequency regions. This selective local modification resolves the contradiction between robustness and quality.
Data Source
AI summary
An audio watermarking system conveys information using an audio channel by modulating an audio signal to produce a modulated signal by embedding additional information into the audio signal. Modulating the audio signal includes segmenting the audio signal into overlapping time segments using a non-rectangular analysis window function produce a windowed audio signal, processing the windowed audio signal for a time segment to produce frequency coefficients representing the windowed time segment and having phase values and magnitude values, selecting one or more of the frequency coefficients, modifying phase values of the selected frequency coefficients using the additional information to map the phase values onto a known phase constellation, and processing the frequency coefficients including the modified phase values to produce the modulated signal.