UDC 621.395
OPTIMIZATION OF ENERGY THRESHOLDS IN WAVELET TRANSFORM FOR SPEECH SIGNAL COMPRESSION BASED ON PARTICLE SWARM OPTIMIZATION
V. T. Dmitriev, Dr. in technical sciences, department of radio control and communication, Head of the de partment, RSREU, Ryazan, Russia;
orcid.org/0000-0001-5521-6886, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Vu Hoang Son, post-graduate student, RSREU, Ryazan, Russia;
orcid.org/0009-0004-7428-5296, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
An adaptive method for optimizing energy thresholds in discrete wavelet transform (DWT) based on Particle Swarm Optimization (PSO) algorithm for speech signal compression has been proposed and inves tigated. The method enables automatic selection of optimal energy retention ratios at each level of wavelet decomposition. Optimization is performed to maximize the compression ratio while simultaneously satisfying predefined constraints on the quality of reconstructed speech at a receiver. The quality of reconstructed speech is evaluated using two objective metrics: segmental signal-to-noise ratio (SegSNR) and perceptual speech quality metric ViSQOL. Experimental studies conducted on standard speech signals recorded in ac cordance with GOST R 50840-95 demonstrate that the method proposed achieves a compression ratio of 87 % while maintaining high reconstruction quality: SegSNR = 9,5 dB and ViSQOL = 3,9 points. According to integral efficiency criterion, the proposed approach outperforms classical methods by 7…14 %.
Key words: : discrete wavelet transform, particle swarm optimization, energy-based thresholding, param eter optimization, SegSNR, ViSQOL, speech signals.