Biomedical subjects
Publications and source records attributed to C Kamm.
This paper discusses some of the aspects of task requirements, user expectations, and technological capabilities that influence the design of a voice interface and then identifies several components of user interfaces that are particularly critical in successful voice applications. Examples from several applications are provided to demonstrate how these components are used to produce effective voice interfaces.
Computer methods, based on the theory and application of signal processing, combined with numerical methods for simulating mathematical processes, facilitate greater objectivity in most aspects of speech intelligibility testing, including specification of the stimuli, control of the tests, evaluation of the responses, corrective feedback, and automatic interpretation. This paper discusses several basic issues in digital signal processing, and also describes the application of computer-aided procedures for recording and delivery of speech materials for audiologic research. Examples of the use of computer procedures for manipulation of digitized stimuli demonstrate the increased efficiency and versatility of these procedures compared to more conventional tape recording methods. In addition, the use of digitized recordings allows more reliable specification of speech levels than conventional calibration methods involving observations of signal peaks on a VU-meter.
A simple up-down adaptive procedure was used to estimate the 50% point on the psychometric function for loudness discomfort level (LDL) and the two functions describing the range of most comfortable loudness (MCL) for listeners with sensorineural hearing impairment. For pure tone and speech stimuli, median LDL and MCL levels were observed at relatively constant SPLs for subjects with hearing loss less than or equal to dB HL and at progressively higher SPLs with further increase in hearing loss. Correlation analysis verified a statistically significant relationship between LDL and magnitude of hearing loss. The nonlinear relationship between LDL and hearing loss together with the large intersubject variability in the data suggest that prediction of LDL from hearing threshold would often be highly inaccurate. These results also demonstrate that averaging LDL data across a group of subjects with a wide range of hearing loss may lead to inaccurate conclusions regarding the effects of sensorineural hearing loss on LDL.
The influence of middle-ear muscle (MEM) contraction on auditory threshold has been measured for pure tones of 0.25, 0.5, and 1.5 kHz. The reflex-activating signal was a 3-kHz pure tone. Signal paradigms were chosen to reduce or eliminate the effects of binaural loudness summation, contralateral direct masking, and contralateral remote and backward masking effects, and to maximize the influence of MEM contraction. Results indicate that under no condition was behavioral threshold affected by the MEM contraction induced using a pure-tone stimulus of 3 kHz, 105 dB SPL.
Performance-intensity (PI) functions for phonetically balanced (PB) word lists were obtained for a group of normal listeners (27 ears), and for two groups of patients with cochlear (89 ears) and retrocochlear disorders (eight ears). Listeners with normal hearing or cochlear disorders exhibited mild to moderate reductions in discrimination score as the speech level was raised above the PB maximum. In contrast, patients with retrocochlear disorders showed a pronounced rollover phenomenon, characterized by a rapid decline in performance as the speech level was raised above the maximum discrimination score.
In a previous article we reported on the standardization of normal hearing for bone conduction. All of our findings were reported in force units. Because many international users of this information express such findings in units of acceleration, we have, at the Editor's request, converted our force units into acceleration units to increase the scope of our results and to enhance their comparison with other investigations. This note presents data which may be useful to organization concerned with the development of standards for normal hearing by bone conduction.
Adaptive procedures were used to determine psychometric functions for loudness discomfort level (LDL) and most comfortable loudness (MCL) for pure tones and speech using normal and hearing-impaired listeners. For the LDL, both groups demonstrated steeply rising functions with the 50% point at approximately 100 dB SPL. The MCL data resulted in two functions, one (Function A) differentiating MCL from less intense stimulus levels and the second (Function B) differentiating between MCL and more intense levels. Function A may be considered a lower bound and Function B an upper bound for MCL. For the normal listeners, the difference between the functions at 50% response ranged from 9.9 to 19.9 dB depending upon the experimental condition. For the hearing-impaired subjects, this range was restricted to approximately 4.5 dB, primarily as a result of a shift in Function A toward higher sound pressure levels.