A theory toward discontinuous perception of speech

Abstract

An experiment by Guberina (1963) was replicated for perceptual and acoustical analyses. The vowel /i/ spoken by a female voice and recorded on a tape was presented via a loudspeaker to normal hearing listeners through the following conditions: (1) no filter; (2) the band with cutoff frequencies of 200–400 Hz at most confortable level (MCL); (3) the band with cutoff frequencies of 2400–4800 Hz at MCL; and (4) two bands just below threshold of intelligibility simultaneously. The results were summarized as follows: (1) the vowel /i/ was perceived in no filter condition; (2) the vowel /u/ was perceived through the 200–400 Hz. Only the first formant was observed; (3) the vowel /i/ was perceived through the 2400–4800 Hz. Only the second and third formants were observed; and (4) when the two discrete bands were combined, the vowel /i/ was perceived. Only the first formant was observable. It was speculated that the suprathreshold low frequency speech energy enhanced the perception of the sub‐threshold high frequency speech energy and as a result the vowel /i/ was perceived. A theory toward discontinuous perception of speech will be discussed especially as it is applied to the rehabilitation of hearing‐impaired listeners.