Deaf Speakers Research Articles

Factors Contributing to the Development of Intelligible Speech Among Prelingually Deaf Persons

This paper is a descriptive study of the incidence of semi-intelligible or better speech among 108 NTID students with congenital hearing loss greater than 91 dB. Statistical comparisons between semi-intelligible (in spontaneous speech) speakers and unintelligible deaf speakers were made in the areas of academic achievement, parental and educational background, and oral-aural skills.

Stress contrasts in the speech of the deaf: acoustic measurements

The study examined the acoustic manifestations of contrastive stress in perceived correct and incorrect utterances of a deaf speaker. The utterances were of the form /əpipap/ or /əpapip/ with stress on either /i/ or /a/. Judgments of perceived stress were obtained from inexperienced listeners of deaf speech and tokens were then sorted according to the perceived and intended productions. Acoustic measurements were made of parameters known to vary with stress in normal speakers including changes in vowel duration, fundamental frequency, amplitude, and formant frequency. No obvious consistent strategy was apparent in the deaf speaker's manipulation of the acoustic characteristics of correct and incorrect utterances. Relationships among these parameters in correct or incorrect productions will be discussed. [Work supported by NINCDS and BRSG.]

Acoustic characteristics of the /I‐i/ confusion in deaf children's speech

The purpose of this paper is to report preliminary data on the acoustic characteristics of error patterns in deaf children's speech. The /I‐i/ confusion was analyzed first because of its high weighting with respect to reduction in percent intelligibility on a multiple linear regression analysis. An acoustical analysis was performed on /i/ in CV syllables which were produced in a sentence context by normal hearing and deaf speakers. Based on the transcriptions of a group of phoneticians, the deaf speakers were then divided into two groups: (1) those who correctly produced the /i/ and (2) those who substituted /I/ for the intended vowel /i/. The data revealed systematic differences in formant values for the three groups of speakers. The two groups of deaf speakers showed the greatest deviation from the normal group in the frequency location of F2. Speakers who produced the /I/ for the /i/ showed the greatest deviation from normal. Measurement of vowel duration revealed no differences between the mean vowel duration for the two groups of deaf speakers, although there were some individual variations. These data indicate that spectral differences associated with the /I‐i/ confusion show a more consistent pattern of deviation from normal than do durational differences. [Research supported by PHS grant 09252 from NINCDS.]

New computer‐based speech‐training aid for the deaf

A computer‐based speech‐training aid, under development, employs the form of template analysis and comparison used in current commercial systems for the recognition of isolated words or short phrases. In this aid, as in such systems, a talker's own utterances are acoustically analyzed to form reference templates to which later utterances are compared. In one training mode, the aid will identify which of a stored list of words has been spoken, and demonstrate to the deaf speaker that utterances which are produced reliably are recognizable, whether or not they are intelligible. This same mode can be used to give a child control over gamelike devices, to promote a high rate of self‐paced oral output. In a second, and more novel, training mode, the child's previous most intelligible speech tokens will be used as reference templates in “successive‐approximation” procedures for improving the intelligibility of important words or phrases. The use of reference templates selected by perceptual criteria makes it unnecessary that this aid be selectively sensitive to the acoustic correlates that characterize deaf speech. Preliminary data demonstrate the general feasibility of this proposed aid. [Work supported by NIH.]

Tongue-lip coordination in speech production in a deaf speaker

Acoustic electromyographic measures were made of an adult deaf speaker producing 15 repetitions of simple nonsense utterances of the form [əpapipə] with stress on either the [i] and [a] vowel. Conventional hooked wire electrodes were inserted into the genioglossus muscle, while surface recordings were made from the orbicularis oris. The gross acoustic results show that the speaker executes stress variations in a conventional fashion, that is, stressed syllables have greater acoustic amplitude and longer duration, although, as is usual for deaf speakers, durations are prolonged overall. At the muscle activity level, the deaf speaker resembles a normal in some respects, but not in others. The general pattern of onsets and offsets for oris activity was appropriate, although muscle activity levels were abnormally high, as had been previously found [D. Huntington, K. Harris, and G. Sholes, J. Speech Hear. Res. 11, 147–158 (1968)]. The deaf speaker was like normals, also, in showing a somewhat greater activity level for [p] closure preceding stressed than preceding unstressed vowels. For the genioglossus muscle, activity was greater for the vowel [i] than for [a], and greater for stressed than for unstressed vowels, on the average; again, as we would expect from normal data. However, individual tokens showed greater variability for the deaf speaker, thus, genioglossus activity occurred during [a] for some tokens. This result can be interpreted to suggest that the vowel neutralization previously observed for deaf speakers [R. Monsen, J. Phonetics 4, 189–198 (1976)] may result, not from tongue inactivity, but from an inability of the speaker to consistently activate some muscles while suppressing others. Additional light will be shed on this supposition by detailed token-by-token analysis of the physiological and acoustical signals, now in progress. [Work supported by NIH Grants NS-13870, and NS-13617 to Haskins Laboratories.]

The effects of vowel height on fundamental-frequency control in the speech of profoundly deaf boys

The effects of vowel height on fundamental-frequency control were examined for eight profoundly deaf and five normal-hearing 9–13-year-old boys. Each speaker read a set of 40 phrases and 12 sentences containing from four to eight repetitions of the vowels /i, ɪ, ɛ, æ, a, ʌ, ɜ̀, ᴜ, u/ and the diphthong /ay/. Fundamental frequency was measured from computer-generated F0 contours near the center of each vowel token. On the average, cross-vowel variations in F0 were comparable for both deaf and hearing speakers, with fundamental frequency tending to increase with vowel height. The deaf speakers, however, appeared to fall into two groups: those for whom cross-vowel variations in F0 were relatively small, and those who produced the high vowels /i/, /ɪ/, and /u/ with excessively high F0. Differences between these two groups of deaf speakers have been examined relative to their articulatory proficiencies, and mechanisms responsible for the F0 variations within the latter group will be discussed. [Work supported by NIH.]

Acoustic qualities of phonation in young hearing-impaired children.

Monosyllables were recorded from a group of 24 hearing-impaired and 6 normal-hearing children between three to six years of age. Tokens of the syllables /ba/ and /bo/ were placed on listening tapes and a panel of 10 experienced listeners made ratings of relative voice quality. The phonation samples were subjected to both a gross (spectrographic) and a fine (computer-aided) acoustic analysis. The following acoustic characteristics were examined: (1) mean fundamental frequency, (2) duration, (3) mean period-to-period changes in intensity and in fundamental frequency, (4) spectral energy ratio above and below 1000 Hz, and (5) intonation contour. The type of intonation contour appeared to be the most important characteristic separating the better from the poorer deaf speakers. The hearing-impaired subjects produced four different types of deviant intonation contours. In addition, two abnormalities of phonation, diplophonia and breathiness, are described.

The effect of deafness on the generation of voice

The glottal source characteristics of 20 hearing‐impaired adolescents were investigated with a reflectionless tube [R. B. Monsen and A. M. Engebretson, J. Acoust. Soc. Am. 62, 981–993 (1977)]. Speakers phonated a neutral vowel in normal and soft voice and in polysyllabic words with primary stress on one syllable. These data were analyzed and compared with data from normal male and female adults and with glottal waves generated by the Flanagan‐Ishizaka two‐mass model of vocal‐fold vibration. It was found that (1) the glottal pulse produced by the deaf subjects, when considered as an isolated entity, is comparable to that of normal‐hearing speakers, although certain phonation abnormalities, such as diplophonia and creaky voice, occur quite commonly among the deaf subjects; and (2) the change of the glottal source over time, as controlled by the interaction of vocal‐fold tension and subglottal air pressure, is a primary deviant factor for many of the deaf speakers, and in particular the use of vocal‐fold tension as a frequency changing mechanism seems impaired by the subjects' deafness. [Supported by NINCDS Grant 03856.]

Effectiveness of linguapalatal place cues for a deaf speaker

The effectiveness of linguapalatal contact place cues were examined in a study of a profoundly deaf adult speaker. A 96-electrode pseudopalate was used to sense linguapalatal contact during utterance of isolated sounds, words, and short sentences. Contact was displayed by an LED array bounded by the maxillary teeth. Initial shaping procedures were directed toward establishing the/t, s, ∫/ sounds. The repertoire was then expanded to include contrating posterior plosives and the afficate /t∫/. Some attention was given to establishing high front vowels. Tongue contact patterns attained on the goal sounds from 20 treatment sessions approximated those of “normal” speakers. Pre- and postintervention intelligibility tests on nontreatment words indicated correct placement of the consonant sounds in over 80% of their occurrances. Physiological and perceptual characteristics of the changes will be discussed. [Supported by NIH Grant NS-11852-02 and HEW Grant RSA 16-P5680 7/4-11.]

An electromyographic investigation of articulation and phonation patterns in the speech of deaf adults

An electromyographic investigation was carried out on the speech of normal-hearing and deaf speakers; the research attempted to answer questions concerning articulation and phonation patterns in the speech of deaf adults. Recordings were made of each subject reading a constant carrier phrase into which a changing medial monosyllable was embedded. The deaf subjects utilized for this investigation were judged to be above average in intelligibility for a group of this type; they all exhibited a profound bilateral hearing loss of early onset. Surface suction electrodes were placed in two positions on the tongue. The EMG data were reproduced via the amplitude display of a two-channel Kay Sonagraph. The results indicated that the deaf speakers extended the durations of speech segments and exhibited difficulty co-ordinating articulation with phonation.

Speech quality of deaf children

A brief review is given of the problems involved in, and solutions suggested for, assessing the intelligibility of the speech of deaf children. A Study of 350 deaf children aged 15–16 is reported which uses a new test for the intelligibility of spoken words. Results, as a function of hearing loss, are compared with the intelligibility of spoken two-digit numbers and with teachers' ratings. In all cases the correlation with hearing loss is very high; the main effect occurring for hearing loss between 65–95 dB. Two-digit numbers are significantly more intelligible to untrained listeners than are words, and the role of reading ability in this connection is examined. With hearing loss taken out there is a highly significant correlation between intelligibility of spoken two-digit numbers and prose comprehension. But there is little evidence that the intelligibility advantage of spoken numbers over spoken words read by a deaf speaker, is due to poor reading ability.

A spectrographic investigation of consonant-vowel transitions in the speech of deaf adults

A spectrographic investigation was carried out on the speech of normal-hearing and deaf speakers; the research attempted to answer questions concerning formant transitions, coarticulation and neutralization of vowels in the speech of the deaf adults. Recordings were made of each subject reading a constant carrier phrase into which a changing medial monosyllable was embedded. The deaf subjects chosen for this investigation were judged to be above average in intelligibility (for deaf talkers that is); they exhibited a profound bilateral hearing loss of early onset. Wide band spectrograms were made of selected recorded stimulus items. The results indicated that the transitions of deaf speakers had a restricted range and slower rate of movement than did those of the normal talkers. Further, the speech of the deaf showed relatively few coarticulation effects, their vowel formants tended to be neutralized and they tended to begin any given articulatory sequence in a manner similar to any other regardless of context.

Auditory aid to deaf speakers in monitoring fundamental voice frequencies

Profound deafness is often accompanied by a loss in frequency‐discriminating ability. When such deafness dates from infancy, the deaf person may never learn to speak with the inflections of fundamental voice frequency that are part of the conventions of language and that contribute to speech intelligibility. Visual and tactile devices have been built to help the frequency‐insensitive deaf speaker‐monitor, and thereby learn to control, the pitch of his voice. This paper describes an auditory pitch‐monitor device, in which voice‐pitch information is presented by way of subsonic‐frequency, square‐wave modulations of the low‐frequency portion of speech. The modulation frequencies track the changing fundamental voice frequencies. The advantages sought in this device, compared to previous devices, are (a) the deaf person does not have to integrate signals from more than one sense into his perception of voice pitch and (b) the device is wearable, making possible a closer relationship between speech‐therapy sessions and everyday communication. Preliminary experimental results will be reported, and there will be a demonstration.

Durational aspects of vowel production in the speech of deaf children.

Two aspects of the duration control of vowel production were examined in deaf and normally hearing subjects: inherent durational differences between two closely related vowels (/i/ and / i /) and the modifying influence of the following consonant. Twelve deaf and five normal adolescents pronounced 56 CV(C) words containing a vowel, /i/ or / i /, followed by /t, s, n, d, z, ø/. Whereas in normal speech the tense vowel /i/ is only relatively longer than the lax vowel / i /, in the speech of the deaf subjects these two vowels occupy much more restricted durational ranges. The modifying influence of a following sound is different in the case of the deaf subjects. Differences between the deaf and normal subjects do not appear to be simply errors of speech production, but reflect instead a different type of linguistic structure. The vowel production characteristics of the deaf subjects account in part for the low intelligibility of consonantal sounds in the speech of the deaf, and imply that standard articulation tests may be an ineffective means of assessing speech production processes of deaf speakers.

An Acoustic and Electromyographic Investigation of Consonant-Vowel Transitions in the Speech of Deaf Adults

An acoustic and electromyographic investigation of the speech of normal-hearing and deaf speakers attempted to answer questions concerning formant transitions, coarticulation, and neutralization of vowels in the speech of deaf adults. In order to accomplish this, EMG and acoustic recordings were made simultaneously as each subject read a constant phrase into which a variable key word was embedded. The deaf speakers chosen for this study were judged to be above average in intelligibility for deaf talkers with a profound bilateral hearing loss of early onset. Acoustical analysis was performed by making wide-band spectrograms of selected recorded stimulus items. The amplitude display of the sonograph was used for relating electromyographic events to acoustic events in time. The results of this study showed that the transitions of deaf speakers had a restricted range of movement and a slower rate of movement than those of normal speakers. The speech of the deaf shows relatively small coarticulation effects, their vowels tend to be neutralized regardless of the context, and they tend to begin an articulatory sequence in a similar manner regardless of the context. The deaf speakers also exhibited difficulty coordinating articulation with phonation and their articulatory behavior showed greater variability than did the articulatory behavior of the normal speakers. The discussion of the results will indicate their implications for speech training of the deaf.

Tactile pitch displays for the deaf

Poor pitch control decreases the intelligibility of deaf speech. Small tactile pitch displays have the potential for supplying continuous corrective feedback for the improvement of the intonation patterns of deaf speakers. A simple pitch detector and tactile display proved effective in correcting a common intonation problem, indicating that work toward a wearable tactile aid is justified. A hypothesis is presented that explains the cause of the pitch problem and the success of the experimental system in terms of the need of the deaf speaker for voicing and frequency referents.

The operant manipulation of vocal pitch in normal speakers.

A device consisting of variable electronic filters and voice-actuated relays was used to raise or lower the vocal pitch of four normal-speaking subjects by the differential reinforcement of selected frequencies emitted by them during oral reading. Continuous, fixed interval, fixed ratio, and variable interval reinforcement schedules were applied to each subject. The results of the study indicated that fundamental vocal frequency is a manipulable operant response. All reinforcement schedules examined produced high rates of response at the selected frequency for all subjects. The variable interval schedules produced the most consistent, high durations of response for all subjects. The method of manipulation of fundamental frequency investigated in the study appears to have promise as a therapeutic technique, especially for deaf speakers and clients with functional pitch disorders.

A study of syllabic stress in some English words as produced by deaf and normally hearing speakers.

The aim of this investigation was to compare the syllable stress production of deaf and normally hearing speakers. Ten orally-trained deaf speakers and forty normally hearing speakers recorded six pairs of bisyllabic words which were either nouns or verb-adjectives depending upon the placement of primary stress. Syllabic differences in intensity, duration, and fundamental frequency were measured. In addition, a panel of expert listeners made judgments on which syllable had been given primary stress. Most (81%) of the words spoken by the hearing subjects were correctly judged , regarding the placement of primary stress, while relatively few (20%) of the words spoken by the deaf were judged correctly. Usually, greater intensity, longer duration, and higher fundamental frequency characterized the correctly identified primary syllable stress produced by both the hearing and the deaf subjects. However, in some instances all three acoustic cues were not necessary for the correct recognition of primary stress. Higher fundamental frequency was always present in the primary stressed syllable of correctly recognized nouns. Greater intensity and/or longer duration characterized the stressed syllables of correctly recognized verb-adjectives. Despite the presence of all three acoustic cues, some words spoken by the deaf subjects were not correctly identified. Inspection of spectrographic recordings suggested that these failures may have been due to lack of falling inflection on the initial syllable of nouns, inability to produce schwa vowels in verb-adjectives, and inappropriate and excessively long junctures between syllables.

Cinefluorographic investigation of velopharyngeal function in selected deaf speakers.

Cinefluorographic investigation of velopharyngeal function in selected deaf speakers.

Perceived nasality in the speech of the deaf.

Listener ratings of nasality were obtained from samples of deaf and normal speech presented in the backward play mode. In order to investigate the influence of rhythmic patterning on the perception of nasality, seven of the normal speakers also produced the samples in a word-by-word fashion. Statistically significant differences in mean ratings of nasality were found between the experimental and control groups and between the reduced-tempo and normal-tempo groups. In general, deaf speakers are perceived as more nasal than normal speakers. Furthermore, normal speakers are perceived as more nasal when speaking at a reduced tempo than when speaking at a normal tempo. Much of the perceived nasality in the speech of the deaf may be a natural consequence of reduced speaking tempo, and therapeutic techniques designed to increase speaking tempo may produce a concomitant reduction in perceived nasality.