THE DESIGN AND TESTING OF A NOISE-REDUCTION ALGORITHM-BASED ON SPECTRAL SUBTRACTION

Three noise reduction algorithms based on amplitude subtraction were d esigned and used to process speech mixed with babble noise in two sign al-to-noise ratios. The estimation of the noise-magnitude spectrum was performed with a novel synchro method, which exploits specific charac teristics of the speech signal. The unprocessed and processed signals were evaluated psychoacoustically by means of a four-alternative-force d choice test with monosyllabic words (minimal pairs) in carrier phras es. The testing was carried out on groups of normally hearing and hear ing-impaired subjects and the long-term power spectra of the processed signals were shaped to be essentially identical with those from the c orresponding unprocessed signals. For the hearing-impaired subjects al l signals were spectrally shaped according to the POGO-fitting rule. N one of the algorithms improved speech intelligibility for any group of listeners and no change in the overall pattern of confusion was obser ved.