Bernd Weinberg

Enhancement of female esophageal and tracheoesophageal speech

Qi [J. Acoust. Soc. Am. 88, 1228-1235 (1990)] has demonstrated that (1) linear predictive (LP) methods can be used to separate vocal tract transfer functions from source functions of vowels produced by alaryngeal talkers and that (2) vowels synthesized with reconstructed transfer functions and totally synthetic voicing excitation sources have improved source-related properties over those present in the original vowels. Here, an extension of this work which is directed to the general goal of developing systems (devices) to enhance the quality of alaryngeal speech is reported. The specific goal of the present project was to determine whether speech, i.e., words spoken by female esophageal and tracheoesophageal talkers, could be enhanced by means of LP-based analysis and synthesis methods. Words spoken by four female alaryngeal talkers were analyzed and synthesized. A perceptual evaluation was completed to permit the quality of the synthetic and the original words to be compared. Listeners generally preferred to listen to the synthesized words, indicating that alaryngeal speech enhancement was accomplished.

Speaker Sex Recognition of 5‐ and 6‐Year‐Old Children's Voices

Listeners were generally able to identify the sex of 5‐ and 6‐year‐old children through recordings of their speech. Differences in average fundamental frequency characteristics among the children did not account for the recognition performance of listeners.

Long‐time spectral and intensity characteristics of esophageal speech

Long-time spectral characteristics and distributional properties of speech intensity produced by esophageal speakers were measured and compared with those of normal speakers. Substantial differences were observed in the distributional properties of speech intensity produced by normal and esophageal talkers. On the average, esophageal speech was about 10 dB less intense than normal speech. In addition, the average long-time spectrum for esophageal speech was characterized by a flattened spectral envelope. Other spectral differences between esophageal and normal speech were found and interpreted in terms of the known variations in voice and speech production of these two divergent forms of oral communication.

Sexual characteristics of preadolescent childrens' voices.

This investigation was undertaken to enlarge current understanding of the acoustic properties which influence the perception of maleness and femaleness in the voices of prepubertal children. Perceptual judgments of sexual identity were obtained in response to tape recordings of whispered and normally phonated vowels, normally spoken sentences, and sentences spoken in a monotonous fashion. Seventy-three children provided recordings. The four utterance types were chosen to experimentally manipulate selected physical properties of speech thought to exert an influence on listener judgments of sexual identity. The results of this work suggest that cues stemming from differences in vocal tract dimensions and/or articulatory behaviors provided the primary cues about the sexual identity of these preadolescent children. Although laryngeal source cues could have provided relevant information about the sex of a few children, this variable was felt to play a relatively minor role in the sex recognition process. New information was uncovered about the role certain suprasegmental factors play in the identification of child sex.

Minimizing the effect of period determination on the computation of amplitude perturbation in voice

Current methods of computing amplitude perturbation present in human voices depend upon being able to accurately determine fundamental period. In this paper, two methods of estimating the amplitude perturbation present in human voices, which do not depend on accurate determination of the boundaries between fundamental periods, are described. In both of these methods, amplitude perturbation is computed as the variance of an ensemble of periods calculated after these periods have been aligned in time. In one method, time alignment is accomplished using zero-phase transformation. In the second method, an unconstrained dynamic programming procedure is used. The accuracy of estimating amplitude perturbation by these two methods is evaluated using synthetic and natural voice signals and is also compared with an estimation using zero-padding based time alignment. The unconstrained dynamic programming method is shown to provide accurate estimation of voice amplitude perturbation over a variety of signal conditions.

On the Relationship between Vowel Height and Fundamental Frequency: Evidence from Esophageal Speech

Data from esophageal speech produced by laryngectomized patients was used to provide a reexamination of major contemporary hypotheses advanced to account for intrinsic variation in vowel fundamental frequency (F0). 4 speakers provided recordings of 15 tokens of three CVC syllables [hid, hud, had]. Analysis of these utterances revealed that intrinsic variation in vowel F0 is clearly evident in esophageal speech. These findings were interpreted (1) to support neither the source-tract coupling nor the tongue-pull hypotheses and (2) to highlight the need to test alternative explanations to account for the intrinsic variation in vowel F0.

Influence of Postvocalic Consonants on Vowel Duration in Esophageal Speech.

The durations of vowels preceding voiced-voiceless stops in utterances produced by three esophageal speakers and three normal speakers were measured. Vowels were significantly longer before voiced stops than before voiceless stops for both esophageal and normal speakers. For either postvocalic consonant environment, the average absolute durations of the vowels of the esophageal speakers were longer than those of the normal speakers. Speaker-group comparisons revealed no significant difference in vowel duration preceding voiceless stops; however, the average absolute duration of vowels preceding voiced stops was significantly longer for the esophageal group. Speaker-group comparisons revealed no significant differences in the relative differences in vowel duration preceding voiced-voiceless stops. The data were interpreted to support the view that the vowel length variation observed is a language-specific behavior governed by a phonological rule of the English language, rather than simply a language-universal behavior governed by inherent physiological characteristics of the speech-production mechanism. Moreover, it was suggested that this phonological rule is a lengthening rule, which lengthens vowels before voiced consonants, rather than a shortening rule, which shortens vowels before voiceless consonants.

Fricative duration in esophageal speech

Comparative observations were made of fricative (/s/ and /z/) durations of normal and esophageal speakers. Differences between the average durational characteristics of fricatives spoken by normal and esophageal speakers varied systematically as a function of their voicing features. Findings were interpreted to lend additional support to the following views: (1) laryngeal amputation occasions widespread changes in speech production, (2) esophageal speakers exhibit articulatory behaviors, and (3) esophageal speakers retain fricative durational voicing contrast features.

Production of Syntactic Stress in Alaryngeal Speech.

This project was a perceptual investigation of syntactic stress in alaryngeal speech. The ability to signal syntactic stress in American English was assessed by obtaining tape recordings of minimally differentiated noun compound-noun phrase pairs (e.g., GREENhouse vs. green HOUSE) from four normal and 16 laryngectomized speakers using four different types of alaryngeal speech. The recordings of these compounds and phrases were presented to 30 listeners for identification using a two-interval-forced-choice procedure. As expected, the four normal speakers were able to achieve high (96% or above) levels of stress contrast. The compound/phrase stress contrasts were also achieved at a high level of proficiency (93% or above) by all four conventional esophageal speakers, all four tracheoesophageal shunt speakers, all four users of the Western Electric #5 electronic artificial larynx, and two of four users of the Servox electronic artificial larynx. Comparisons were made with the results of similar experiments on intonation, contrastive stress, and lexical stress. The findings highlight the contributions that study of clinical samples may make to questions about speech production and perception.

Estimation of Glottal Volume Velocity Waveform Properties: A Review and Study of Some Methodological Assumptions

Publisher Summary This chapter discusses methods used to estimate glottal waveform properties. The chapter presents the results of a project undertaken to assess critical assumptions underlying the use of reflectionless tube methodology. The methods used to estimate glottal waveform properties can be divided into two categories. The first involves those that measure the vibratory pattern of the vocal folds. The basic rationale for pursuing this approach is that the shape of the glottal pulse is largely influenced by the configuration of the glottis and by the size and duration of medial contact of the vocal folds during vibration. The second category involves methods that attempt to obtain a more direct estimate of the glottal volume velocity waveform by means of acoustic analysis