Experienced observers were asked to identify, in a four-level 2AFC situation, the longer of two unfilled time intervals, each of which was marked by a pair of 20-msec acoustic pulses. When all the markers were identical, high-level (86-dB SPL) bursts of coherently gated sinusoids or bursts of band-limited Gaussian noise, a change in the spectrum of the markers generally did not affect performance. On the other hand, for I-kHz tone-burst markers, intensity decreases below 25 dB SL were accompanied by sizable deterioration of the discrimination performance, especially at short (25-msec) base intervals. Similarly large changes in performance were observed also when the two tonal markers of each interval were made very dissimilar from each other, either in frequency (frequency difference larger than 1 octave) or in intensity (level of the first marker at least 45 dB below the level of the second marker). Time-difference thresholds in these two latter cases were found to be nonmonotonically related to the base interval, the minima occurring between 40-and 80-msec onset separations.For the past decade, readers of the psychophysics and speech literature have witnessed a growing interest in the perception of temporal aspects of speech (for a review, see Studdert-Kennedy, 1975). Many of the studies generated by this interest deal with the perception of brief time intervals defined by speech sounds. Several phenomena that have been revealed in this area, ones that relate to the perception of voice onset time (VOT), vowel transitions, syllabic duration, etc., are intriguing from the psychoacoustical point of view. Yet, for two reasons, these phenomena, as well as most of the others related to speech perception, do not lend themselves easily to detailed psychophysical analysis. First, the prerequisite for such an analysis would be the definition of what the exact timing cues are in a given speech sound ensemble. Because of the enormous

To test the adequacy of French regulations for sound isolation in buildings, subjects were asked to rate their annoyance with samples of music filtered by electronic "insulation curves" representing different party walls. The insulation curves differed in their shape but all provided an A-weighted le•,el difference of 51 dB with a pink noise source, measured over a 1/3-octave bandwidth of 40 Hz-10 kHz. However, the different insulation curves did not provide the same degree of sound isolation with various music samples due to source spectral differences. A statistically significant correlation was observed between annoyance and the A-weighted level difference ratings of the insulation curves when bandlimited pink noise (125 Hz-4 kHz)was used as a source. This correlation was not present when broadband (40 Hz-10 kHz) pink noise was used for the performance rating. Subjects showed a preference for insulation curves with steeper slopes (9 and 12 dB/oct), thus preferring a greater relative attenuation at higher frequencies. Additionally, the presence of coincidence dips was found to have an effect on subject preference that appeared to depend upon both the frequency range at which they occurred and the slope. The bandwidth of the music signals and the intelligibility of speech in the intruding sounds were also found to influence the annoyance ratings. These results indicate that the level difference method for rating sound insulation could better predict occupant response if the above results were accounted for in the procedure.

Twenty-eight audiologically normal adult subjects participated in a study designed to assess how well six noise-rating indices would predict the annoyance caused by 3-min recorded samples of traffic noise obtained from both nominally constant-speed and stop-and-go traffic. The study was performed in a laboratory simulating a home environment. Annoyance judgments were obtained through the use of a magnitude estimation technique involving a 10-point scale. Subjects were also asked if they could accept each of the 24 traffic sounds if heard on a regular basis in their homes. Data obtained indicate that the simpler noise-rating indices, such as the average sound level and the level exceeded 10% of the time, predict annoyance as well as, if not better than, complicated schemes incorporating a measure of either variability or rate-of-change of levels with time. Thus it appears that the measurement and computational burdens associated with these complicated schemes are unwarranted.

Experienced listeners identified the longer of two unfilled time intervals, each bounded by acoustic markers. Time-difference thresholds (d′ = 1.0) were measured under a variety of conditions in a four-level, 2AFC procedure. When all four tonal markers were identical and at a high level (86 dB SPL), tonal frequency did not affect performance. For 1-kHz markers, performance deteriorated as their intensity was creased below 25 dB SL. When the markers of each interval differed in intensity or in frequency, performance also decreased and the time-difference threshold was not monotonically related to base duration. This nonmonotonicity was also observed when the time intervals were hounded by a 10-msec burst of bandpass (800–2000 Hz) noise at 71 dB SPL and by a 500-msec buzz formed from 100 pulses/sec, filtered with “formants” at 630 and 1000 Hz. The best relative threshold for base durations of 10, 25, and 60 msec, was found at 25 msec. When these CV-like, noise-buzz patterns were passed through a series of one-third octave filters, some filter conditions yielded performance as good as with the entire spectrum, while some did not. The principal factor was the intensity difference between the noise and buzz energy in the particular one-third octave band. We conclude that time-marking information is not integrated across frequency bands, and that the temporal discrimination is controlled largely by temporal masking in that spectral region where masking is smallest.

In 1948, Henry [J. Exp. Psychol. 38, 734–743 (1948)] found that duration of pure tones is most discriminable in the 0.5–2-kHz frequency range. The present study represents an attempt to examine whether frequency effects can be observed also in the discrimination of time intervals marked by two bursts of pure tones, identical in frequency. The frequency of the markers covered a range from 0.5 to 4 and their level was 86 dB SPL. Each tone burst marker had a 20-msec total duration and its envelope was smoothened so that the major part of the sound energy be kept within the 50-Hz range surrounding the tonal frequency. Data of three trained listeners, obtained for interburst intervals (t) ranging from 20 to 140 msec, indicate that the logarithm of the interburst interval difference discriminable at threshold (Δtd′=1.0) is a nearly linear function of log t. The slope of this function is close to 0.5 for the 0.5- and the 4-kHz markers, whereas it is approximately 0.8 for the 1- and 2-kHz markers. Pitch difference cues in the pairs of markers may only partly account for the observed frequency effect. [Supported by NINDS Grant No. 03586.]