REVIEW OF THE MODERN METHODS OF MASKING OF THE ACOUSTIC SPEECH SIGNALS FOR THE PRIVACY PURPOSES

Abstract

In today’s world, the importance of protecting speech privacy spans from safeguarding personal conversations to securing national interests. Among the effective strategies for mitigating speech eavesdropping is acoustic speech masking, which introduces background noise to reduce the intelligibility of private conversations. This article reviews and compares various modern methods of the acoustic speech signals masking for privacy purposes, with a primary focus on their effectiveness, subjective comfort, and complexity of implementation. The article begins the review from stationary colored noise maskers, white and pink noise, emphasizing the latter’s favorable spectral characteristics and broader adoption in office environments due to its deeper, less intrusive sound. Then, the discussion turns to the speech-shaped noise (SSN) technique, which refines this approach by matching the spectrum of typical human speech, yielding equivalent privacy at lower sound pressure levels while being perceived as more natural. Further, the babble noise masking approach is reviewed, a mixture of multiple simultaneous human voices, is examined for its capacity to combine both energetic and informational masking, providing enhanced effectiveness over previously reviewed masking noises. Besides that, the article reviews adaptive speech-dependent masking and ultrasonic jamming approaches. The paper analyzes masking performance, annoyance levels, and deployment feasibility for each of the discussed approaches, highlighting the trade-offs between masking efficiency, comfort and simplicity. The paper concludes that while pink and SSN maskers are widely used in commercial environments, advanced methods like babble and adaptive maskers offer improved performance but are less common due to complexity.