David H. Raab

Visual detection of signals in the presence of continuous and pulsed backqrounds

The detectability of intensity increments in the presence of continuous and pulsed backgrounds was investigated using a two-alternative, temporal, forced-choice procedure. Differences were found in intensity-duration reciprocity relations, the form of the Weber function, and the shape of the psychometric function between continuous- and pulsed-detection conditions. In a second study, signals were added either to a steady background (simple detection) or to background plus pedestal (pedestal detection). Two unusual phenomena were noted, namely, “negative masking” and the “pedestal effect.” The interpretation of the results in terms of a simple, Poisson-detection model is discussed.

Magnitude Estimation of the Brightness of Brief Foveal Stimuli

Eighteen observers judged the apparent brightness of light flashes that varied in both duration and luminance. The median numerical estimations (made relative to a standard flash) confirmed three principles: the reciprocity between luminance and duration (Blochs law), the enhancement of brightness at about 50 milliseconds (Broca-Sulzer effect), and the power-law relation between brightness and energy (Stevenss law).

Pure‐tone intensity discrimination: some experiments relating to the “near‐miss” to Weber's law

Intensity DLs for pulsed tones (f = 1000 Hz) were determined at two sensation levels. As has previously been noted, discrimination was better at the higher level. The possibility that this effect reflects the subjects use of information from aural harmonics was tested using stimuli specifically designed to disrupt such information. These stimuli did affect discrimination, but only when they were at levels in excess of those which would be required by the usual estimates of aural harmonic magnitudes. Further experiments with interfering stimuli used three types of noise: wide‐band, high‐pass, and band‐stop. The noises had equal energies in the critical band around 1000 Hz and had spectrum levels in their respective pass‐bands of −60, −30, and −35 dB relative to the 1000‐Hz test tone. The noises had no effect on performance at the lower sensation level. At the higher level, performance was impaired, the greatest impairment being produced by the band‐stop noise, the least by the wide‐band noise. This was in...

Effects of Waveform Correlation and Signal Duration on Detection of Noise Bursts in Continuous Noise

Studies of the masking of noise bursts by continuous noise revealed the pedestal effect previously reported by others for 100‐cps tones. The correlation ρ between mask and probe waveforms was found to affect detection: the pedestal effect—evidenced with ρ = 1 or with ρ = −1—was absent with ρ = 0. In a second experiment, the pedestal effect was found to diminish as probe bursts were shortened in duration from 100 to 5 msec. The results of these and related experiments provide support for the energy‐detection model of Pfafflin and Mathews. Tone and noise probes are considered to be detected by virtue of the energy changes they produce. Some features of the energy‐detection model are discussed.

Noise‐intensity discrimination: Effects of bandwidth conditions and mode of masker presentation

Noise‐intensity discrimination was studied as a function of both signal and masker bandwidth. Five bandwidths of noise—ranging from 100 to 10 000 Hz—were employed. Maskers were presented at each of three spectrum levels (5, 25, and 45 dB re 0.0002 μbar). Discrimination thresholds were relatively unaffected by changing bandwidth over a two‐decade range when the signal and masker were filtered together, with either continuous or gated presentation of the masker. When the masker bandwidth was greater than that of the signal, the reciprocity between signal power density and signal bandwidth was found to be 5 dB per log unit of bandwidth with continuous maskers and 5–10 dB (depending upon level) with gated maskers. The results were compared with predictions of energy‐detector models of noise‐intensity discrimination. Modifications of simple energy‐detection schemes were discussed.Subject Classification: [43]65.58, [43]65.75, [43]65.50; [43]50.70.

Audotory intensity discrimination with bursts of reproducible noise

Auditory−intensity discrimination was investigated using gated bursts of reproducible noise. Weber fractions were determined at a number of energy levels with stimuli that differed in intensity, duration, and bandwidth. Weber’s law was approximated in all cases. In addition, discriminability was found to be somewhat dependent on burst duration and relatively independent of noise bandwidth; threshold signal−to−noise ratios improved when duration was increased from 10 to 100 msec, but not when bandwidth was increased from 500 to 5000 Hz. The data were compared with the results of a second experiment in which the stimuli were random—rather than pseudorandom—noises. Finally, the results of both experiments were compared with predictions about intensity discrimination derived from two theoretical models: (1) a Poisson−counting mechanism and (2) an energy detector incorporating additive internal noise.Subject Classification: 65.35, 65.58.

Forward and Backward Masking between Acoustic Clicks

The masking of one click by another was studied as a function of the time interval between the pulses. Two‐alternative forced‐choice procedures were employed to measure thresholds before and after the masking click. Both forward and backward masking were found; the forward effect was more pronounced and longer lasting. Backward masking studied in this way extends beyond what can be explained by peripheral intensity‐latency conversions.

Effect of temporal overlap on brightness matching of adjacent flashes.

Flashes having durations of 10, 20, 50, and 100 msec were matched in brightness to an adjacent 200-msec standard. When the standard and test stimuli terminated together, the results confirmed earlier demonstrations of the Broca–Sulzer phenomenon. Different functions relating the growth of brightness to duration were generated, however, when the flashes matched to each other had coincident onsets or coincident mid-durations. The matches made with stimuli terminating together were shown to exhibit transitivity: Targets which are as bright as a standard are approximately as bright as each other. Finally, an effect of relative luminance on apparent temporal position is described.

Click‐Intensity Discrimination with and without a Background Masking Noise

A study of intensity discrimination employing click stimuli revealed three unusual effects. First, clicks are difficult to differentiate with respect to intensity. Second, the Weber function differs from those usually obtained with stimuli of longer duration. Third, click‐intensity discrimination is, improved by the addition of a continuous background noise. Since click stimuli have virtually no energy variations, these results cannot be explained by stimulus‐oriented theories of detection and discrimination.

Fluctuations of N1 Amplitude in Relation to Click‐Intensity Discrimination

Round‐window potentials evoked by clicks were recorded from anesthetized guinea pigs. By comparing N1 amplitudes on line, it was possible to compute physiological difference limens for click intensity. The method requires no assumptions about the underlying distributions of neural effect. Weber functions were generated with and without a continuous masking noise. The effects of click intensity and of background noise paralleled those previously reported for human listeners. A neural model was described in which variability of N1 amplitude turned out to be the principal parameter.