Title :
A perceptual metric for masking
Author_Institution :
Dept. of Electr. Eng., California Univ., Los Angeles, CA, USA
Abstract :
Analytical tools were developed to predict masking of formant frequencies of steady-state vowels in the presence of a noise masker. Two kinds of masking were studied: within-band masking (where the signal frequency was within the bandwidth of a noise masker) and above-band masking (where the signal frequency was above the upper cutoff frequency of the masker). It was found that results of auditory-masking theory can be used to predict accurately within-band masking. Above-band masking, however, could not be predicted well from existing procedures, and an algorithm was developed to predict this kind of masking.<>
Keywords :
filtering and prediction theory; hearing; speech intelligibility; algorithm; auditory-masking theory; formant frequencies; noise masker; perceptual metric; steady-state vowels;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1993. ICASSP-93., 1993 IEEE International Conference on
Conference_Location :
Minneapolis, MN, USA
Print_ISBN :
0-7803-7402-9
DOI :
10.1109/ICASSP.1993.319411
