“The Science of Vocal Pedagogy” | Open Indiana

As a speech sound, each of the fifteen phonemes [i], [e], [I], [ē], [ɛ], [æ], [a], [ɑ], [ɔ], [ð], [o], [u], [U], [ʌ], and [ɜ] maybe considered basic because each is a phoneme. Dialectal tendencies in speech permit a leniency of departure from each phonemic entity, but communication is always consummated. (See Fig. 105A.)

As sung sounds, phonemes required in the proper pronunciation of a word are most frequently altered from their phonemic centers and substituted by the singer as he selects a certain timbre, color, or articulatory position. This faulty process tends to place voice quality, timbre, and color ahead of phonemic integrity. In such a process of substitution the meaning of the word is often impaired. (See the first objective on p. o.)

Accurate phonemic selection is achieved by the singer in his production of isolated vowel sounds only when his selection of a phoneme is disciplined by his aural awareness of that phoneme. The phoneme cannot remain a concept; it must become a sound. All singers need much more specific training to develop this aural awareness.

Accurate phonemic identification is attained by the listener when the sung phoneme is nearest to that acoustic position in which the phoneme is properly pronounced. Cues for its identification are glides, caused by consonantal movement to and from the phoneme; however, a proper singing technique reduces glides to a minimum. (See steady state formant structure characteristics, p. 228.)

The singing of any vowel sound is a conceptual process involving the selection of some phonemic entity within a tonal environment suggested by the proper pronunciation of a word. A testing procedure must seek answers to the following question: Which of the numerous phonemic choices available to the singer should be considered the proper “home base” position that will serve as a point of reference from which all sung phonemes may be instantly determined by the singer to preserve the meaning of the word.

The recognition tests I (male) and II (female) were designed by the author to determine which of fifteen sung phonemes were the most frequently identified correctly under controlled conditions.

To construct Tests I and II the author, a bass, and a professional soprano, sang each of the vowel sounds [i], [e], [æ], [ɑ], [ô], and [u] and made them conform to a musical standard of vocal utterance that was acceptable to a selected group of faculty members at Indiana University. To control uniformity of each vowel, well-defined physiological positions of the tongue, labial orifice, velum, mandible, and larynx were used. (See p. 231.) The positions of these physiological parts were determined by means of cinefluoroscopy and X-ray photographs. The intensity of each vowel was subjectively judged by the faculty members to be mezzo forte but not forte. Although the integrity of the vowels [i], [ɑ], [ô], and [u] are affected by the intensity level at which they are produced, it was reasoned that to assure a uniformity of identification all phonemes should be sung at the same level. Each vowel was analyzed acoustically and positioned upon a formant chart. The frequency of formants one and two determined each position.

To arrange a random sampling of the six phonemes, each phoneme was written upon six cards and shuffled ten times. The random sampling for males differed from that for females. Each card was then numbered and a master tape was made by a dubbing process, repeating each vowel in its numerical random order upon the tape. Each sung vowel on the test was identified by speaking the number assigned to each vowel one second before the sung sound was heard.

This process provided the necessary contrast to the sung sound and enabled those judging the sounds to keep accurate track of each vowel and record their judgments in the proper numbered squares upon the intelligibility sheets.

These tests were presented to 224 students enrolled in classes in phonetics at the University of Michigan or in classes in vocal pedagogy at Indiana University. All students participating in the tests had been drilled in phonetic transcription.

The high percentage of recognition of these six vowels when sung in random order are displayed in Tests I and II.

Tests III (male) and IV (female) were designed in the same manner as Tests I and II by the same singers to determine the extent of phonemic recognition when nine additional phonemes were sung with the six predetermined basic vowels. These nine phonemes were [ē], [I], [ɛ], [a], [ɔ], [o], [U], [ʌ], [ɜ]. The tests were presented to the same classes at the University of Michigan and Indiana University.

The results of the tests, displayed in Charts II and III, have determined the acoustic position of each of the fifteen phonemes used in the phonetic system within this work. Vowels with the highest frequency of recognition proved to be the predetermined basic vowels [i], [e], [æ], [ɑ], [ô], [u]. The vowels of lower frequency of recognition [ē], [I], [ɛ], [a], [ɔ], [o], [U], [ʌ], [ɜ], were considered to be the quality alternate of a particular basic vowel.

Test I. Frequency of Recognition of Basic Vowels, Male Voices

The figures on the diagonal line indicate the frequency of correct recognition of the basic vowels [i], [e], [ɜ], [ɑ], [ô], and [u] in 1,344 responses by 224 persons. The numbers on either side of the central line indicate the frequency of error and the vowels mistaken for the basic vowels.

Test II. Frequency of Recognition of Basic Vowels, Female Voices

The larger figures on the central line indicate the frequency of correct recognition of the basic vowels [i], [e],[æ], [ɑ], [ô], and [u] in 1,344 responses by 224 persons. The numbers on either side of the central line indicate the frequency of error and the vowels mistaken for the basic vowels.

Test III. Frequency of Recognition of All Phonemes, Male Voices

The figure on the diagonal line indicates the frequency of correct recognition of that sung phoneme in 1,344 responses by 224 persons. The numbers on either side of the central line indicate the frequency of error and the vowel mistaken as the sung vowel.

Test IV. Frequency of Recognition of All Phonemes, Female Voices

Since all sung vowels are altered by variations in intensity and frequency the name “quality alternate” means that it is a color substitute for that vowel, which must be pronounced in a specific manner to preserve the meaning within the word. The quality alternate vowel is not controlled by frequency or intensity as is the basic vowel. (The application of these vowels in the singing process may be found on p. 226.)

Two vowels, [i] and [a] are considered to be basic vowels although they were not most frequently recognized as the intended phoneme. The vowel [i] which can only be sung properly at intensity levels of pianissimo or piano was sung mezzo forte and had migrated to its quality alternate position of [I] or [e] and was so identified. (See Tests III and IV.) The vowel [ɑ] at mezzo forte intensity had migrated into the neutral vowel position of [ʌ] and was properly recognized.

TABLE III. Percent of Recognition of Basic Vowels

The percent of recognition of each vowel sung in Tests I and II is indicated above. The vowels most often mistaken as the sung vowel are indicated under greatest conflict of identification for each vowel.

TABLE IV. Percent of Recognition of Basic Vowels

The percent of recognition of each vowel sung in Tests III and IV is indicated above. The vowels most often mistaken as the sung vowel are indicated under greatest conflict of identification for each vowel.

The Science of Vocal Pedagogy