Intelligibility of Prolonged Vowels in Classical Singing*
Gregg, Jean Westerman, Scherer, Ronald C., Journal of Singing
MANY PEOPLE ATTENDING CONCERTS SUNG IN ENGLISH readily state that they are unable to understand the words being sung. Even the Metropolitan Opera Company provides titles for operas sung in English, which have included in recent years: Billy Budd by Britten, The Ghosts of Versailles by Corigliano, The Voyage by Glass, The View From the Bridge by Bolcom, and The Great Gatsby by Harbison. These observations suggest that a closer look at the intelligibility of sung English is in order.
The question arises that perhaps a failure for high intelligibility for sung English in opera may be due to the use of only a few English vowels while vocalizing during singing instruction and warm up. Intelligibility for the rest of the English vowels might then be relatively low because of their lack of use in a singer's vocalizing exercises. Accordingly, this study focused on the intelligibility of sung American English vowels relative to the vowels used in vocalises by singing teachers.
During speech, vowel targets are reached only briefly and are influenced by coarticulation. During singing, however, the vowel must be prolonged according to the rhythmic duration of the word, prescribed by the composer. Hence, the ability to hold a relatively constant vocal tract configuration must become highly developed in the singer. Additionally, which vowel configuration is chosen, and which vowel is perceived by the listener, will determine the conveyance of the meaning of the word, as the sequence "meat, mate, met, mat, moat, moot" suggests. If the vowel sound is inaccurate, so might be the received meaning.
The study of vowel intelligibility during singing has a rather rich history. Typically, as the fundamental frequency rises during singing by women, the first formant will begin to rise ahead of the fundamental frequency to preserve beauty and loudness but not preserve vowel intelligibility.1 The typical finding is that higher sung frequencies tend to have lower vowel intelligibility.2
Of these respondents, forty-four voice teacher/singers plus one husband/singer volunteered to participate in a research project on vowel accuracy in singing. Information collected from this subgroup included the age of the subject, number of years of voice study, number of years of teaching voice (including type/level of studio), and the vowels used in their own daily practice. In view of the far larger percentage of use of the five cardinal vowels, it was hypothesized that the intelligibility of sung vowels would decrease if the sung vowel was not used in daily practice by the singer.
The research was designed with the following objectives: 1) to determine whether there is greater intelligibility of sung vowels used in daily practice compared to those not used in daily practice, and 2) to determine the relation between the intelligibility of eleven sung vowels and the singer's intended vowel as judged by a different group of voice teachers and a group of speechlanguage pathologists.
Perceptual Rating Form
Eleven words-"beat, bait, Bob, boat, boot," representing the most frequently used vowels in practice, and "bit, bet, bat, bought, but, book," representing the other six vowels used less frequently, were arranged in six random orders. Judging sheets were devised (Figure 2), listing each of the eleven vowels across the top of the sheet. IPA (International Phonetic Alphabet) symbols were given, accompanied by a word (different from those above) containing the vowel. Spaces were provided for noting the perceived vowel choice of the sung sounds by each subject. There was a space on the far right for the level of confidence of the judgment, ranging from zero (low confidence) to five (high confidence), so that the judge could indicate how sure he or she was of the vowel identified. The judges all had separate judging sheets for each of the subjects.
A DAT recorder (Tascam Model DA-P1) was patched to a Mackie (1402 VLZ Pro 14) Channel Mic Line Mixer. …