INVESTIGATION OF THE NATURE OF ACOUSTIC CUES FOR RECOGNIZING CZECH VOWELS: ARE THE TRANSITIONS SUFFICIENT CUES FOR RECOGNIZING CZECH VOWELS?
INVESTIGATION OF THE NATURE OF ACOUSTIC CUES FOR RECOGNIZING CZECH VOWELS: ARE THE TRANSITIONS SUFFICIENT CUES FOR RECOGNIZING CZECH VOWELS?

Summary/Abstract: It has been shown that English vowels in consonant-vowel-consonant (CVC) syllables can be perceived accurately even when most of the vowel has been silenced and only the consonants and the consonantal formant transitions are presented to listeners. We conducted a similar experiment for Czech in which 3 types of stimuli were used: (i) full CVC syllables, (ii) silent-center syllables, and (iii) silent-center syllables whose initial and final portions were produced by two speakers of different sex (mixed-sex). Identification rates of the full syllables were better than those of the silent-center and mixed-sex syllables which did not differ from each other. From that we infer that in Czech (like in English) silent-center identification is possible not because listeners would compute the missing formant contours (which would predict higher error rates in the mixed-sex condition) but because formant transitions themselves contain sufficient cues to vowel identity. This conclusion is discussed from a wider theoretical perspective.

• Details
• Contents