Gregory P. Widin

Noncategorical perception of stop consonants differing in VOT.

The discriminability of bilabial stop consonants differing in VOT (the Abramson-Lisker bilabial series) was measured in a same-different task, an oddity task, and a dual response, discrimination--identification task. Subjects showed excellent within-category discrimination in all three tasks after a moderate amount of training in a same-different task with a fixed standard and with feedback. In addition, discrimination performance continuously improved with increasing stimulus difference for both intra- and intercategory comparisons. Also, subjects were able to alter their identification responses so that well-defined category boundaries fell at arbitrary values determined by the experiments. These results are not compatible with a strict interpretation of the categorical perception of stop consonants.

Intensive and temporal effects in pure‐tone forward masking

Intensive and temporal characteristics of forward masking were assessed using a pure‐tone masker and a 1‐kHz probe. For a given interval between masker and probe, the slope of the masking function tends to decrease with increasing masker frequency, but is always less than unity. When the masker frequency equals that of the probe, increasing the masker‐probe interval decreases the slope of the masking function. Implications of these data for forward‐masked psychophysical tuning curves are discussed. A model, based on a linear combination of exponential decays and an initial logarithmic transform of stimulus intensity, predicts the general form of the masking function for varying interval between masker and probe, and suggests time constants which decrease with increasing masker level.

Masker interaction in pure‐tone forward masking

Forward masking of a 1-kHZ, 20-ms probe was examined using two temporally distinct 1-kHZ, 20-ms maskers, separated by 6.5 ms. The interval between the second masker and the probe was the same as the interval between the first and second masker. Masking functions were obtained for each masker separately, and for varying levels of one masker with a fixed level of the other. The presence of the first masker at a fixed level raised the probe threshold over that obtained with the second masker alone: this elevation was approximately constant, independent of the level of the second masker. The results indicate that masking effects of the two maskers are not independent, i.e., the maskers interact. Several models of masker interaction are discussed.

Learning and generalization of intraphonemic VOT discrimination

A recent study [A. E. Carney, G. P. Widin, and N. F. Viemeister, J. Acoust. Soc. Am. 62, 961–970 (1977)] has shown that with appropriate training listeners can consistently discriminate intraphonemic differences in a synthetic bilabial VOT stop consonant series. We examined whether listeners trained to discriminate intraphonemic differences on one VOT series can also make intraphonemic discriminations on a different VOT series. Identification and discrimination data were initially obtained from a group of six listeners for both a bilabial and a velar VOT stop consonant series. Half the listeners then received discrimination training on the bilabial series and half received training on the velar series. Discrimination and identification data were subsequently obtained from all listeners for both series. Improved final discrimination on the nontraining series indicates that listeners had learned to discriminate VOT per se, while a failure to improve discrimination on the nontraining series indicates that sp...

Time‐intensity envelope cues for consonant recognition

A primary source of perceptual information for the cochlear implant user in the time‐intensity envelope of the speech waveform. The purpose of this study was to estimate the amount and type of information for consonant recognition potentially available in the time‐intensity envelope of speech. The experimental stimuli were generated from the time‐intensity envelopes of 19 /aCa/ utterances (C = /p, t, k, b, d, g, f, θ, s, ∫, v, ð, z, ȝ, m, n, r, l, j/), spoken by a male talker. The envelopes were obtained by full‐wave rectifying and low‐pass filtering the speech waveforms at 2000, 200, and 20 Hz. These envelopes were used to multiply noise with a 3‐kHz bandwidth, producing three sets of envelope stimuli with identical, flat spectra that differed in the amount of time‐varying in their amplitude envelopes. The unprocessed speech waveforms and the three sets of envelope stimuli were presented to 12 normal‐hearing subjects in blocked, closed‐set consonant recognition tests. Individual and group confusion matri...

Short-term spectral effects in pure-tone forward masking

Measurements of the masking function with different masker envelopes suggest that short‐term spectral broadening at offset affects the frequency extent of forward masking, at least for masker frequencies above the probe frequency.

Intensity effects in forward masking: Implications for psychophysical tuning curves

The effects of masker intensity and masker frequency on the threshold of a 1‐kHz, 20‐ms probe were assessed in forward masking, with a masker‐probe delay of 5 ms. In general, probe threshold is a power function of masker intensity, with an exponent considerably less than unity; the exponent and constant of proportionality depend upon masker frequency. For example, with a 1‐kHz masker, a 10‐dB increase in masker level increases probe threshold by only about 5 dB, over a 60‐dB range of masker intensities. The nonlinear relation between probe threshold and masker intensity is consistent with the increased “selectivity” sometimes observed in forward‐masked tuning curves as compared with those obtained in simultaneous masking. These data then emphasize the importance of nonlinear intensity effects, in addition to possible changes in true frequency selectivity, in determining the form of psychophysical tuning curves obtained in forward masking and their relation to those obtained in simultaneous masking. [Suppo...

Perceptual similarity of complex signals presented by electrical stimulation of the auditory nerve

Electrical stimulation of the auditory nerve bypasses the transduction mechanisms of the cochlea. Thus the properties of electrical stimuli that make them perceptually similar or dissimilar may differ from those of acoustical stimuli. We have examined the influence of bandwidth, center frequency, peak factor, spectral shape, and tonality on perceptual similarity. Nineteen stimuli were presented electrically to four cochlear implant subjects and acoustically to two normal‐hearing subjects. All subjects first adjusted the stimuli for equal loudness and then rated their similarity in a triadic comparison task. The similarity matrices from the triadic mask were submitted to multidimensional scaling analyses using the KYST‐2A and SINDSCAL programs. The obtained stimulus configurations and dimensions for implanted and normal‐hearing subjects will be discussed and interpreted in terms of the psychophysics of hearing with electrical and acoustical stimulation.

Loudness of complex signals presented by electrical stimulation of the auditory nerve

An understanding of the factors affecting the loudness of complex electrical stimulation is necessary if improved speech processing schemes are to be developed for cochlear implants. This experiment used a variety of complex signal types to examine the relationship between perception of loudness and physical characteristics of electrical stimulation. Loudness was studied for a set of 19 stimuli differing in bandwidth, center frequency, peak factor, spectral shape, and tonality. Four cochlear implant subjects and two normal‐hearing subjects made equal‐loudness judgments for the stimuli using the method of adjustment. A 200‐Hz tone at a comfortable listening level served as the reference. Equated stimuli delivered to the electrode were analyzed for amplitude and frequency characteristics. The relationship of these statistics to perceived loudness will be discussed.

Psychometric functions in pure‐tone masking

Psychometric functions were obtained for detection of a 20‐ms, 1‐kHz sinusoid, in the presence of forward or simultaneous pure‐tone maskers. The psychometric functions are considerably steeper in simultaneous than forward masking. For example, the range of the psychometric function for on‐frequency masking between the 60% and 90% correct points is 3–4 dB for simultaneous masking and about 10 dB for forward masking. The slope of the functions appears to vary little as masker frequency is changed, for constant probe frequency, in either simultaneous or forward masking. Thus, comparisons of selectivity in simultaneous and forward masking will not depend on performance level. The steep slope in the simultaneous‐masking condition is consistent with earlier observations. [L. A. Jeffress, J. Acoust. Soc. Am. 58, 399–403 (1975); D. M. Green and J. Nachmias. J. Acoust. Soc. Am. Suppl. 1 65, S59(A) (1979)], and may reflect relatively high uncertainty regarding signal onset and duration for short signals. [Research ...