Methods of forming acoustic jamming for counteraction means against unauthorized documentation of speech information

Abstract

The article is devoted to solving the actual problem of protecting confidential speech information by analyzing and improving methods of forming acoustic interference. The aim of the work is to systematize methods of acoustic jamming formation and experimentally substantiate the effectiveness of spectrally balanced and adaptive methods for counteracting unauthorized recording. The article provides a comprehensive analysis of existing approaches, among which white noise, tonal, speech-like, and formant jammers are highlighted. Special attention is paid to the scientific and methodological aspects of jammer synchronization with the useful signal.

Based on software modeling in the Python environment using signal processing libraries, a comparative evaluation of the effectiveness of four types of jammers was conducted. Effectiveness was assessed using a complex of objective metrics: PESQ (Perceptual Evaluation of Speech Quality) and STOI (Short-Time Objective Intelligibility), which provided verified quantitative results. It has been experimentally established that the highest efficiency is provided by the proposed method of formant jamming, which ensures targeted masking of the most informative spectral components of speech. It is calculated that the formant jammer provides the lowest quality scores (PESQ) and intelligibility scores (STOI) compared to other types of jammers. The research on the dependence of the STOI intelligibility index on the distance between the jammer source and the recording device confirmed the critical influence of spatial conditions on information security. The results prove the feasibility of using intelligent jamming formation methods in modern active acoustic protection systems to increase their effectiveness.