Dynamic Speech Watermarking for Conference Call Source Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In voice-based conference calls, it is challenging to identify the source of unauthorized copying or dissemination of sensitive information due to the difficulty in determining the origin of unauthorized recording and dissemination among multiple participants.
Innovation Solution
A method and system that dynamically watermark a conference call signal with a unique identification signal, imperceptible to humans, which identifies the destination end user device by monitoring characteristics of the speech signal and embedding the watermark in real-time, allowing for the identification of the source of unauthorized copying.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a watermark signal is embedded in the speech signal to identify the source of unauthorized copying, then the ability to identify the source of unauthorized copying is improved, but the complexity of the system increases
Solution Approach 1:
The system embeds a watermark signal containing destination end user device identification into the speech signal before transmission. This preliminary action ensures that if unauthorized copying occurs, the source can be identified by analyzing the watermark in the copied signal, thus improving identification accuracy without requiring complex post-detection systems.
Solution Approach 2:
The watermark signal acts as an intermediary element that carries identification information through the speech signal transmission chain. By embedding this intermediate identifier, the system enables source tracking without requiring direct complex communication between all system components, thereby managing system complexity while maintaining identification capability.
2Object-affected harmful factors
If the watermark signal is adjusted dynamically based on speech signal characteristics, then the imperceptibility of the watermark is improved, but the processing time increases
Solution Approach 1:
The watermark signal characteristics are adjusted dynamically based on the monitored speech signal characteristics. This dynamic adaptation ensures the watermark remains imperceptible by matching it to the temporal and spectral properties of the speech signal, thereby reducing watermark perceptibility while accepting increased processing time as a necessary trade-off for maintaining audio quality.
Data Source
AI summary
An apparatus and method for determining a source of an unauthorized copy of speech signals in a conference call session. A conference bridge receives a speech signal during a conference call session with a plurality of end user devices attending. The conference bridge provides one or more dynamic, end user specific watermarks corresponding to the one or more end user devices. Each watermark is adjusted based on characteristics of the speech signal to make the watermark imperceptible to a human being. The speech signal is then embedded with the adjusted watermark in real time to generate a modified speech signal.