INVESTIGATION OF THE EFFECTIVENESS OF WHITE, PINK, AND ADAPTIVE NOISE IN ACTIVE SPEECH MASKING

Madida Jusse Charlene, Ivanov Vladimir Alekseevich, Krepak Ivan Pavlovich

The article discusses the problem of active protection of speech information using acoustic masking methods. The main focus is on studying the effectiveness of various types of noise—white, pink, and adaptive—in suppressing speech intelligibility. The results of modelling performed using digital audio signals and software processing in the Python environment are presented. The effectiveness was evaluated using objective quality metrics — signal-to-noise ratio (SNR), speech intelligibility index (STOI), and speech perception index (PESQ). Experiments have shown that adaptive noise provides the best compromise between masking level and acoustic comfort. The results obtained can be used in the development of speech information protection systems in rooms for various purposes.

Keywords:active information protection, speech masking, acoustic noise, STOI, PESQ, adaptive noise control