FIELD: medical engineering. SUBSTANCE: actual value of main current period and spectrum envelope are determined in sonicated sections of speech signal after they pass through auxiliary comb-like filter with spacing of peaks of amplitude-frequency characteristics equal to obtained periods of main current. Obtained values of spectrum envelopes and main current periods are used to form control signals which set up corresponding spacing of peaks and their relative values in amplitude-frequency characteristic of main filtration. Signal is subjected to main filtration after its time delay for noise-proof separation of comb-like filtering parameters. Upon filtering, signal is added to delayed signal after its band filtration. Summary signal obtained is supplied to patient's hearing system after amplification and electroacoustic conversion. Advance of above-indicated processing of signal is automatic adaptation of both frequency and amplitude characteristics of hearing apparatus to spectra of speech signal elements. This makes it possible to enhanced noise-proof feature of identification by ear of such key elements as stressed vowels, and some tyres of consonants which improves word intelligibility of speech percepted by patients with neurosensory hypoacusis against noise background. EFFECT: improved speech intelligibility. 2 dwg
