摘要:

Sound‐pressure recordings made in real‐ear simulators are more realistic than those made in 2‐cc couplers. Unfortunately, most real‐ear simulators (1) require factory calibration and (2) were designed to simulate only the mean acoustic impedance of normal adult ears. Due to variability of ear‐canal geometry and eardrum impedance among subjects, the statistical possibility of any one subject exhibiting such “mean” characteristics—especially if that subject has a conductive pathology—is very remote. Reported herein is the development of a programmable artificial ear (PAE) which (1) requires no factory calibration and (2) can be programmed to account for individual variations in source impedance, ear‐canal geometry, and eardrum impedance. The PAE consists of a 2‐cc coupler, microphone, amplifier, and desk‐top scientific computer, equipment which is included in most laboratory/clinic inventories. Sound pressures recorded in the 2‐cc coupler are adjusted via the computer to what they would have been had the recordings been made in a particular subject's ear. Good agreement has been obtained between PAE data and real‐ear simulator measurements.

ISSN:0001-4966    

DOI:10.1121/1.2022147

出版商:Acoustical Society of America    

年代:1985

数据来源: AIP

摘要:

The development of practical methods of evaluating speaker recognizability in processed speech has been hampered by the lack of an adequate system, either physical‐acoustical or perceptual, for classifying voices. Previous efforts to develop taxonomies based on perceived voice traits (PVT) failed for a variety of reasons, including inappropriate scaling procedures and grossly inadequate voice samples. In the present attempt to remedy the situation, more than 500 potential voice descriptors were experimentally screened to yield a reduced set of rating scales with which listeners could reliably describe their perceptions of individual voices. Factor analysis of rating data for a sample of eighty voices revealed at least eight orthogonal PVTs, several of which had not been previously described. Methods of evaluating the effects of speech processing on speaker recognizability from its effects on the reliability and factorial structure of voice rating data are discussed and illustrated.

ISSN:0001-4966    

DOI:10.1121/1.2022150

出版商:Acoustical Society of America    

年代:1985

数据来源: AIP

摘要:

Thirty hearing‐impaired children were given two tests of intonation contour perception. The first involved a receptive task in which subjects had to identify the “odd man out” in three stimuli. The second involved an imitative task in which subjects had to imitate an intonation contour. The imitations in the second task were tape recorded for subsequent rating by five trained listeners on a ten‐point scale. The same stimuli were used for both tests, these being computer‐generated synthetic vowels with rising, flat, and falling contours. Among the older subjects (=>5 years), all the severely deaf (3 frequency average loss90 dB HL) passed both the receptive and the imitative task. Among the younger children (<4 years), the proportions passing the imitative task were the same as in the older group. Only one severely deaf child, however, was able to pass the receptive task. The fact that the receptive task was too difficult for the younger age group was confirmed using normally heating children. (Both hearing and deaf children, however, were able to perform a visual‐spatial analog of the task.) The findings suggest that the evaluation of intonation perception in young hearing‐impaired children should be based on imitative tasks, rather than receptive tasks. They further indicate that all severely deaf, and approximately 50% of profoundly deaf children should have sufficient auditory capacity for the perception of intonation contours.

ISSN:0001-4966    

DOI:10.1121/1.2022160

出版商:Acoustical Society of America    

年代:1985

数据来源: AIP

摘要:

Eight severely‐to‐profoundly hearing‐impaired children produced each of 14 phrase‐length utterances.F0 measures were obtained at several locations within each utterance (startingF0, finalF0, andF0 peaks). The relative timing of eachF0 measure (ms from onset of the utterance) was also determined. In addition, several difference measures were derived. Listeners experienced with the speech of the deaf served as subjects for the perceptual experiment and were asked to judge whether they heard a terminal fall or a flat final intonation contour. A multiple linear regression analysis was used to determine if any combination of the acoustic measures could predict listeners' responses. Results suggest that the only variable which made a significant contribution to the regression function was the temporal interval between the terminal peakF0 and the finalF0. The regression equation accounted for a statistically significant but not a large proportion of the total variance. The implications of these results will be discussed. [Work supported by PHS grant NS17764 to the Graduate Center, CUNY.]

ISSN:0001-4966    

DOI:10.1121/1.2022161

出版商:Acoustical Society of America    

年代:1985

数据来源: AIP