Esophageal Speakers Research Articles

Speaking proficiency variations according to method of alaryngeal voicing

Twelve judges, with no previous exposure to laryngectomees, rated the speaking proficiencies of 33 laryngectomees divided into the following groups: esophageal speakers (n = 12); electrolarynx speakers (n = 11); and tracheoesophageal puncture speakers (n = 10). In addition, the speech of ten normal subjects was rated. Specific speaking parameters examined included voice quality, pitch, loudness, intelligibility, rate of speaking, visual presentation during speech, extraneous speaking noise, and overall communicative effectiveness. Multiple discriminant analyses performed on the ratings made by each judge revealed significant differences in ratings for various speaking parameters in the four subject groups. Results generally support the stance that tracheoesophageal speech is perceived as superior to communication with either an electrolarynx or with traditional esophageal speech, even though it is not viewed as comparable to normal voice.

Perceptual dimensions of alaryngeal vowels

The present study uses multidimensional scaling to examine the perceptual dimensions underlying the perception of alaryngeal vowels and compares these dimensions with those commonly found in MDS studies utilizing normally voiced vowels. A set of 11 vowels ([iɪεæaΛoᴜᴜuɔɚ] produced by a competent esophageal speaker represented the stimulus vowels. These vowels were compared (pairwise) on a nine‐point similarity/dissimilarity scale by 20 listeners and the resulting perceptual distance data were analyzed using INDSCAL. Three perceptual dimensions were obtained (significantly correlated with F2, F1‐F0, and F3, respectively), which accounted for 67% of the variance. These dimensions are similar to those obtained in studies using normally voiced vowels and suggest that alaryngeal voice quality does not disrupt the use of these common perceptual dimensions (which depend only upon vocal tract configurations). The nature of these perceptual dimensions will be discussed in terms of both acoustic (i.e., formants) and auditory (i.e., bark‐transformed formants) measures.

Pharyngoesophageal myotomy for vocal rehabilitation of laryngectomees.

Pharyngoesophageal myotomy, performed in 16 laryngectomees to improve vocal rehabilitation, proved to have some negative side effects. Although no failures of speech rehabilitation using a button were consequently found, a considerable number of button assisted esophageal speakers had a breathy voice. Also, the number of poor injection-esophageal speakers was much higher than in the group of nonmyotomized laryngectomees. A "normal" pseudoglottis could only rarely be identified by fluoroscopy in the myotomized group. A significantly lower intratracheal pressure appeared to be required for tracheo-esophageal phonation after myotomy. To improve the slightly disappointing voice rehabilitation results of the myotomized laryngectomees, a modified myotomy is proposed.

A British experience of surgical voice restoration techniques as a secondary procedure following total laryngectomy.

An earlier paper (Cheesman et al., 1985) described assessment procedures for 'failed' oesophageal speakers who were referred to Charing Cross Hospital for possible Surgical Voice Restoration in 1983-4. This paper presents the results of surgical voice restoration in the first fifty laryngectomized patients referred and analyses the reasons for failure in those where surgical voice restoration was unsuccessful. 94 per cent of patients who underwent secondary voice restoration were successful at two weeks post-surgery, but this success rate dropped to 73 per cent at three months. Reasons for this are explained and the pre-requisites for a successful surgical voice restoration are indicated.

A comparison of the speech acceptibility of good and excellent esophageal and tracheoesophageal speakers

Thirteen esophageal speakers (8 male, 5 female) and 12 tracheoesophageal (t-e) speakers (7 male, 5 female) were placed into eight groups based on gender, voice type, and speaking proficiency (excellent or good). Audio recordings of each speaker reading the Rainbow Passage (Fairbanks, 1960) were made. The resultant samples were judged for acceptability of speech by 25 naive listeners trained in use of a five-point equal-appearing-interval scale. An analysis of variance of the listener data indicated that speaker proficiency but not voice type had a significant effect on judgments of acceptability, indicating that the listeners percieved no difference in the acceptability of esophageal and t-e speech.

Rhythm in Thai Esophageal Speech

Duration measurements associated with rhythmic aspects of speech produced by one Thai esophageal speaker were compared to previously published identical measures for normal Thai speakers. Measures of central tendency and variability were taken of the duration of syllables, pauses, phrases, and interstress intervals. Despite a slower-than-normal overall speaking rate and a shorter-than-normal phrase length, the esophageal speaker was able to maintain normal relative temporal relations among syllables within phrases. Although his speech was dysrhythmic at the overall discourse level, it was rhythmic at the phrase level. Findings are interpreted to highlight the importance of differentiating features of alaryngeal speech that are common across languages from those that are specific to particular languages.

Pitch extraction of esophageal speech

A number of studies have been devoted to the description of the prosodic characteristics of esophageal speakers. However, pitch extraction of esophageal speech is difficult because of its low pitch, of the often quasi-periodic character of the vibration of the pharyngo-esophageal segment, and because stoma noise may be added to the signal. Commercially available hardware F0 extractors often produce artefacts in the analysis of esophageal speech. Consequently it may be asked whether the harmonic source model of laryngeal speech is inappropriate for esophageal speech, or whether it simply constitutes an extreme of this model in terms of the proportion of noise it contains. In a number of recent studies, cepstral analysis has been applied to pitch extraction of esophageal speech. This procedure handles low fundamentals well, but is computationally slow and complex. In this study, several other pitch extraction algorithms which are computationally simpler than the cepstrum will be compared to cepstral analysis in regard to their adequacy for F0 extraction with esophageal speech. Some of these, especially the harmonic product spectrum and the harmonic sum spectrum, have been shown to be superior to the cepstrum in the analysis of noise speech signals. [Work supported by the Norris Cancer Hospital, USC.]

Ratings of intelligibility of esophageal and tracheoesophageal speech

Audio-recorded oral reading and continuous discourse speech samples were obtained from esophageal and tracheoesophageal speakers. The samples were rated twice by each of two groups of listeners, one experienced and the other inexperienced with alaryngeal speech. The main finding was that reliable mean ratings of intelligibility of both types of speech can be obtained from experienced and from inexperienced judges. The experienced and inexperienced listeners tended to rank the speakers similarly with regard to the intelligibility of their speech samples. The reading and discourse samples also tended to be ranked similarly with regard to intelligibility. The tracheoesophageal speakers' samples as a group were rated as more intelligible than were those of the esophageal speakers.

Pitch Perturbation Measures of Voice Production of Laryngectomees After the Amatsu Tracheoesophageal Shunt Operation

Twenty tracheoesophageal speakers, 5 esophageal speakers, and 10 laryngeal speakers were subjected to pitch perturbation measurement in the sustained vowel production. Mean jitter, jitter ratio, average perturbation, and relative average perturbation were calculated from the successive vocal period. Jitter ratio values were 9.7 for laryngeal speakers, 29.6 for tracheoesophageal speakers, and 58.3 for esophageal speakers. Relative average perturbation values were 6.1 X 10(-3) for laryngeal speakers, 18.2 X 10(-3) for tracheoesophageal speakers, and 38.1 X 10(-3) for esophageal speakers. Results indicate that tracheoesophageal speech using exhaled pulmonary air is more stable than conventional esophageal speech.

The role of the upper oesophageal sphincter in voice rehabilitation after laryngectomy and Staffieri's procedure.

The Staffieri method was used for voice rehabilitation after total laryngectomy in our patients. Leakage of saliva has been a problem in some. It has been suggested that aspiration could be due to specific pressure relations in the pharynx, upper oesophageal sphincter and cervical oesophagus during deglutition (Mann et al., 1980). Regarding voice production, the upper oesophageal sphincter is probably the main sound source, functioning in essentially the same way as in oesophageal speech. Intraluminal pressure recordings at rest and during swallowing were used in 16 Staffieri speakers and 9 oesophageal speakers. Upper oesophageal sphincter dysfunction was seen more often in Staffieri speakers with aspiration problems than in Staffieri speakers without aspiration problems, but the relation is not clearly understood. A definite relation could be established between upper oesophageal sphincter function in terms of relaxation and coordination and aerodynamic pressure measured in the tracheostoma, necessary to sustain phonation. After dilatation of the oesophagus, produced by the inflow of air, a reflex rise in the upper oesophageal sphincter pressure is thought to occur. The pressure necessary to sustain phonation probably depends on the extent to which this reflex mechanism occurs and on the ability of the patient to induce a relaxation of the upper oesophageal sphincter before air expulsion. Selective myotomy will possibly lower the necessary pressure and enhance vocal rehabilitation. Whether this will solve a part of the aspiration problems in Staffieri speakers is completely unpredictable, as the relation of the motor function of the upper oesophageal sphincter and the aspiration problems is not well understood.

Recent advances in the assessment of 'failed' oesophageal speakers.

ABSTRACTBy means of videofluoroscopy laryngectomy patients can be clearly categorised into five mutually exclusive groups. It has been shown that difficulties in oesophageal voice acquisition are directly related to adequacy of the pharyngo‐oesophageal (P‐E) segment, which is a direct result of the type of surgery used at the time of initial laryngectomy. This paper gives indications for intervention in each of the five groups described and emphasises the need for earlier objective evaluations in order to remedy these difficulties.

Production of syntactic stress in alaryngeal speech.

An acoustical investigation of syntactic stress in alaryngeal speech was undertaken. Measurements were made of fundamental frequency, relative intensity, vowel duration, and intersyllable duration from noun compounds and noun phrases spoken by four normal speakers, four esophageal speakers, four Blom-Singer tracheoesophageal speakers, four users of the Western Electric 5A electrolarynx, and one user of the Servox electrolarynx. Results indicated that esophageal and tracheoesophageal speakers were able to manipulate the relative magnitudes of fundamental frequency, intensity, and vowel duration in a manner comparable to those of normal speakers. With the exception of one user of the Western Electric 5A electrolarynx, who manipulated both fundamental frequency and duration, electrolarynx users generally manipulated only duration properties. All alaryngeal speakers manipulated intersyllable duration, whereas normal speakers did not. The degree of acoustical contrast for both peak fundamental frequency and vowel duration was significantly greater on syllables in final position than in initial position for normal and alaryngeal speakers alike. These findings suggest that stress contrasts in alaryngeal speech are based on a complex of acoustic cues which are influenced by linguistic structure.

Location of the vibratory segment in tracheoesophageal speakers.

The Singer-Blom tracheoesophageal puncture procedure for surgical-prosthetic voice restoration has proved to be a viable option for alaryngeal speech rehabilitation. Following tracheoesophageal puncture, occlusion of the tracheostoma shunts pulmonary air through the Blom-Singer prosthesis into the cervical esophagus. The pulmonary air passing through the cervical esophagus and into the hypopharynx causes a portion of the upper alimentary tract to vibrate in a manner similar to that of the pharyngoesophageal segment during the production of esophageal speech. To study the location and shape of the vibratory segment in tracheoesophageal speakers, videofluoroscopy and simultaneous voice recording were performed with 16 patients. To analyze the vibratory segment(s), photographs were made of the videotaped image while it was stopped during the patients' production of the /a/ sound. The most frequent location of the vibratory segment was in the lower third of the neck, which corresponds to cervical vertebrae C5 through C7. Five of the subjects had two separate vibratory segments and two other subjects had long vibratory segments. The vibratory segment(s) in tracheoesophageal speakers was found to be similar to the vibratory segment(s) in esophageal speakers.

Fundamental frequency and intensity measurements in laryngeal and alaryngeal speakers

Fundamental frequency measurements and voice sound pressure levels were examined in 10 tracheoesophageal, 10 esophageal, and 10 laryngeal speakers during three phonation tasks. The results indicated that tracheoesophageal speakers had a fundamental frequency approximately 25 Hz higher than esophageal speakers. Intensity levels for laryngeal and tracheoesophageal speakers were similar. Discussion includes support for this new voice restoration technique.

Evaluation of speech of laryngectomees after the amatsu tracheoesophageal shunt operation

Fourteen patients were evaluated by a number of tests to analyze the voice produced following the senior author's surgical technique for speech rehabilitation. The parameters used were sustained duration of the prolonged vowel, counting per breath, syllables per breath, words per breath, and words per minute for quantitative determination. The qualitative parameters used were pronunciation accuracy, level of intelligibility and adequacy. The results obtained were better than those of esophageal speakers and approximate laryngeal voice. A comparison of relative advantage over other rehabilitative procedures is presented.

Utterance and pause lengths and frequencies in oral readings by superior esophageal speakers and by normal speakers

Utterance and pause lengths and frequencies in oral readings by superior esophageal speakers and by normal speakers

Acoustic characteristics: Tracheoesophageal speech

This study examined differences among five tracheoesophageal, four esophageal, and five laryngeal speakers on the acoustic parameters of intensity and duration. Each subject repeated 14 prerecorded model sentences and prolonged phonation of three vowels. Analogue prints of the voice samples were obtained, and measures were subjected to parametric statistical analysis. Differences among groups indicate that the tracheoesophageal speakers are more like normal speakers than are esophageal speakers, indicating that surgical procedures such as the tracheoesophageal puncture are a significant rehabilitative alternative for the laryngectomee.

Perception of Intonational Contrasts in Alaryngeal Speech

The achievement of intonational contrasts by 4 normal and 16 laryngectomized speakers was assessed in this project. The laryngectomized subjects represented four clinical subgroups of alaryngeal speech: esophageal, tracheoesophageal, Western Electric #5 electrolarynx, and Servox electrolarynx. High-quality tape recordings of each subject's productions of two pairs of sentences (Bev loves Bob) spoken in statement and question from were presented to 40 listeners for evaluation using a two-interval-alternative-forced-choice procedure. Intonational contrasts were achieved in a highly effective manner by the normal, esophageal, and tracheoesophageal speakers. In contrast, users of electronic artificial larynges were generally unable to achieve these intonational distinctions except for one user of the Western Electric #5 electrolarynx.

Acoustical Characteristics of Intended Syllabic Stress in Excellent Esophageal Speakers

The ability of excellent esophageal speakers to manipulate acoustical characteristics associated with intended syllabic stress was studied. Five excellent esophageal speakers and five sex- and age-matched normals produced 10 sentence pairs, each containing a bisyllabic stimulus item differing only in primary stress placement. The mean fundamental frequency, sound pressure level, and duration of the stressed and unstressed vowel nuclei were analyzed. Although some differences in absolute levels were apparent, only sound pressure level differences reached statistical significance. For both groups, intended primary stress was associated with a comparable pattern of increased fundamental frequency, sound pressure level, and duration. The present findings suggest that excellent esophageal speakers are capable of producing some correlates of primary syllabic stress in a fashion remarkably similar to but somewhat less consistent than normals. The implications of these data for long-range clinical planning are discussed.

Perception of contrastive stress in alaryngeal speech

This project was a perceptual investigation of contrastive stress in alaryngeal speech. The ability to signal contrastive stress in American English was assessed by obtaining high quality tape recordings of sentence production from four normal and sixteen laryngectomized speakers using four different types of alaryngeal speech. The recordings of these sentences, paired on the basis of differences in the location of contrastive stress, were presented to forty listeners for perceptual evaluation of stress location using a two-interval-forced-choice procedure. As expected, the four normal speakers achieved high (95% or above) levels of stress contrast. Contrastive stress patterns were also realized in a highly effective manner (93% or above) by three esophageal speakers, three tracheoesophageal (Blom-Singer) speakers, and three users of the Western Electric # 5 artificial larynx. Contrastive stress patterns were realized in a lower (86–88%), but reasonably effective manner by two users of the Servox artificial larynx. The findings were interpreted to highlight the contributions that study of clinical samples may make to questions about linguistic theory, speech production and speech perception.