METHOD OF ENSURING THE CONFIDENTIALITY OF SPEECH NEGOTIATIONS TAKING INTO ACCOUNT PSYCHOACOUSTIC MASKING
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
The article considers the problem of ensuring the confidentiality of speech signals in communication systems under conditions of limited resources, strict requirements for delay and incompatibility with the existing infrastructure, where traditional cryptographic methods often prove unsuitable. A new approach to noise masking is proposed, aimed not at minimizing theoretical mutual information, but at maximizing the degree of psychoacoustic masking — that is, at effectively hiding the useful signal by using the properties of human auditory perception. An optimization model for the synthesis of an adaptive noise additive is formalized, which simultaneously satisfies technical constraints (frequency band 100–4000 Hz, power limitation), functional requirements (reproduction quality for a legal user, MSE < 0.01) and provides the possibility of synchronization. A bell-shaped spectrum is used as a parametric noise model, which allows reducing a high-dimensional problem to the optimization of several key parameters. Using the example of a numerical experiment with a speech signal of 1 s duration, it is shown that the proposed method significantly increases the degree of masking — up to 87% of the useful signal energy becomes unacceptable for perception by an unauthorized listener — while maintaining high quality signal recovery. The results confirm the promising approach for building practically implemented protection systems that combine security, efficiency and compatibility with existing communication standards.
Fields of Study & Subjects
Wireless Communication Security Techniques
Speech and Audio Processing
Cybersecurity and Information Systems
Publication details
- Journal / source
- Cybersecurity Education Science Technique
- Volume
- 4
- Issue
- 32
- Pages
- 565–565
- Publication date
- 26 March 2026
- Language
- EN