Long-term average spectrum of Spanish

Abstract

The long-term average spectrum of Spanish, as spoken in Mexico, was obtained for thirty subjects—fifteen men and fifteen women. To obtain the average spectrum, use was made of a digital system comprised of an FFT analyzer interfaced to a microcomputer. All data acquisition, storage and processing was thus digital. The main features of the average spectrum are: peaks in the region of 240 Hz and 500 Hz, a −20 dB/octave slope from 0·5 kHz to 1 kHz; a flat region between 1 kHz and 2 kHz, a further decline of −10 dB/octave to ≈3 kHz and, finally, a large flat region (within 5 dB) from 3 kHz up to 10 kHz. These results agree reasonably well with some other studies for English. The technique employed here is thus adequate for this kind of study, being relatively fast and easy to implement.