885095941

THURSDAY AFTERNOON, 2 MAY 1991

INTERNATIONAL C, 1:30 TO 3:45 P.M.

Session 7PP

Psychological and Physiological Acoustics: Hearing Aids and Audiological Measurement

Lynne Marshall, Chair

Nawal Submarine Medical Research Laboratory, Box 900, Groton, Connecticut 06349

Contributed Papers

1:30

7PP1. Constraints on nonlinear amplification for recruitment of loudness compensation. Apichat Tungthangthum and Janet C. Rutledge (Elect. Eng./Comput. Sci. Dept., Northwestern Univ., Evanston, IL 60208)

A nonlinear amplification techniquc has bcen deve!oped to compen-sate for recruitment of loudness in sensorineural hearing losses [J. C. Rutledge, “Timc-Varying, Freąuency-Dependent Compensation for Recruitment of Loudness,” Ph.D. thesis, Georgia Inst. of Technol., 1989]. This techniquc uses a sinusoida! speech model to incorporate a model of the psychoacoustic masking of sinusoids in normal hearing and in hearing-impaired persons. It operates on both a time-varying and frequency-dependent basis. The drawback is that the model processing may distort the vowcl formant structure and thus cause confusions between vowels. Therefore constraints on the relativc amplitudę levels of the sinusoidal components at different frequency regions should be incorporated into the model. These constraints have bcen studied and will be discusscd.

1:45

7PP2. Loudness matching for compressed speech signals. Matthew H. Bakke, Arlene C. Neuman, and Harry Levitt (Ctr. for Res. in Speech and Hcar. Sci., CUNY Graduate Ctr., 33 W. 42nd St., New York, NY 10036)

The purpose of this expenment was to determine a rule for matching the loudness of uncomprcssed and compressed speech samples. Eighl normal hearing and 16 sensorineural hearing-impaired subjects (dividcd into four groups of differing sevcrity and configuration) listened to two continuous speech samples (małe and female talkers) under conditions of linear and compression amplification. Compression conditions con-sisted of 12 selected combinations of varying compression ratio, thresh-old, and release time (compression ratio = 2:1, 4:1, 8:1; knee point = 5, 10, 15, and 20 dB below the highest speech peak; release time = 20, 200 ms). Subjects matched the loudness of the compressed signals to a reference uncompressed signa) using an adaptive procedurę. Overall levels and cumulative distributions of the processed speech signals were obtained. Estimates of equa! loudness based upon rms levcl and 90th percentile of the cumulative distribution were both found to closely match the gain selected by subjects. [Work supported by NIH Grant No. 2P01 DC00178.]

2:00

7PP3. Reliability, sensitlyity, and validity of magnitude estimation, pa i red comparisons, and category scaling. Suzanne C. Purdy (Audiol.

1974 J. Acoust. Soc. Am., Vol. 89, No. 4, Pt 2, April 1991

Sec., Dept. of Physiol., Univ. of Auckland, Privatc Bag, Auckland, New Zealand) and Chaslav V. Pavlovic (Univ. of Iowa, Iowa City, IA 52242)

This study invcstigated the reliability, sensitivity, and validity of speech intelligibility judgments for hearing aid evaluation. Subjects judged the intelligibility of sentences using either magnitude estimation, category scaling, or paired comparisons. Speech recognition scores for CID sentences and NU-6 words were also obtained. Test conditions were selected using articulation index (Al) theory. Al theory predicts that speech intelligibility is monotonically related to the articulation index which depends on the available speech dynamie rangę and the imporianee [/(/)] of each frequency band for speech intelligibility. The speech signal was bandpass filtered so that the Al inereased monotonically regardless of /(/). Filter conditions were selected that produce an inerease in Al for speech materials spanning the rangę of probable importance functions, namely nonsensc syllablcs, casy speech, and av-erage speech [C. V. Pavlovic, J. Acoust. Soc. Am. 82, 413-422 (1987)]. Speech scores and intelligibility judgments were obtained for each filter condition. Test-retest reliability was poorest for paired comparisons and CID sentence scores. There were no differences in sensitivity among the scaling procedures. Intelligibility judgments and NU-6 scores were morę sensitivc than CID sentence scores to differences among conditions. The results indicated that intelligibility judgments are valid mea-sures of speech recognition.

2:15

i

7PP4. Supra-aural and insert earphone occlusion effeets. Diana C. Wright and Joseph Angelclli (Dept. of Commun. Disord., Pcnn State Univ., University Park, PA 16801)

Monaural and binaural occlusion effeets were determined for 30 subjects using mastoid placement of a B-71 bonę vibrator using supra-aural earphones (TDH-49/P/N 51) and insert carphones (Eartone-3A/Earlink-3A), which had a shallow or fuli insertion depth. Statisti-cally significant differences were found for earphone type, frequency, and insert earphone insertion depth. Overall, the mean occlusion effect for all conditions dccrcascd as frcqucncy incrcascd froni 250-2000 Hz and was greater for the binaural than the monaural occlusion and for the supra-aural than the insert earphone occlusion. Also, the mean occlusion effect was greater for the shallow compared with the fuli insertion depth for the insert earphone indicating the occlusion effect inereased as the volumc of air between the occluder and the ear drum inereased. The standard deviations were not consistently higher or lower across frequency for any one occlusion condition. Clinical impli-cations regarding the magnitude of the occlusion effect with supra-aural and insert earphone systems will be discussed.

121st Meeting: Acoustical Sodety of America 1974