885095949

much bcttcr for high F_L values. Thus it is concluded that detection of inharmonicity (experiments l and 2) was limited by factors other than those rcsponsible for finc pitch discriminations. These results also havc implications with respect to thc origin of human listcners’ melodie oc-tave templates.

9:45

8MU4. Sharpness measurements for musical instrument timbres. Pamela Goad (Systematic Musicology Próg., School of Musie, DN-IO, Univ. of Washington, Seattic, WA 98195)

The timbres of string and woodwind families were investigated by measuring sharpness and loudness over a 2-octave range and between individual performers. Sharpness is defined as a freąuency-dependent weighting of the loudness-critical band ratę patiem. It is promising for timbre perception research bccause it addresses the most salient timbral dimension. Digital recordings were collected with performers 4 m from the microphone in a musie rehearsal hall. Perfomiers played diatonic scales at two extreme musical dynamie lcvcls, pianissimo and fortissimo. Across all families, sharpness inereases with playing frequencies and with dynamie level, and lies between 0.5-2.0 aeums. Loudness variance is larger than sharpness variance. O/crtap in sharpness is greatest at the pianissimo levels with the exception of the double reeds. The string family smoothly changes in sharpness across thc entire playing rangę at both musical dynamics levels. The clarinet family has less overlap at the fortissimo levcl and a slight inerease in sharpness for lower pitches. Double reeds exhibit the greatest variance in sharpness. Performer-specific variations are evident, but the generał trends in sharpness rc-main the same.

ltfcOO

8MU5. Normalization process in the human auditory system. Jennifer L. Cho, Michael D. Hall, and Richard E. Pastore (Dept. of Psychol., State Univ. of New York, Binghamton, NY 13901)

Irrelevant variability of source characteristics imposes problcms in maintaining perccptual aecuracy of an intended message. The percep-tual system appears to adjust for diffcrences between sources in a time-consuming manner callcd normalization which has been investigated with speech stimuli. The present experiment sought to establish auditory normalization with musie stimuli using an AX chord discrimina-tion task. Stimuli consistcd of four chords produced by five diflFcrent instruments. Degree of instrument variability was manipulated within subjects by presenting stimuli in three normalization blocks: single instrument, multiple instrument, and multiple-mixed instrument. In these conditions instrument selection remained the same throughout a błock of trials, differed across trials, or differed within and across trials. Sig-nificant expccted diłferenccs in response latcncics and task accuracy were observed for trials that were composed of different instruments, compared to trials having the same instrument. Results clcarly indicate normalization in the perception of musie. [Work supported by NSF.]

10:15

8MU6. Measuring slmilarity of musical timbres. Paul Iverson and Carol L. Krumhansl (Dept. of Psychol, Uris Hall, Cornell University, Ithaca, NY 14853)

Similarity scaling, adjective scaling, and acoustic measurements were used to assess the similarity of 16 digitally recorded tones produced by orchestra! instruments playing the same pitch. In similarity scaling experiments, subjects rated the similarity of the tones on a con-tinuous scalę. Similarity scaling experiments were run on the complete tones, the first 85 ms of the tones, and on the tones with the first 85 ms removed. In an adjective scaling experiment, subjects rated each complete tonę on 21 different adjective scales. Acoustic measurements in-cluded correlations between the amplitudę spectra of the tones, and

1988 J. Acoust. Soc. Am., Voł. 89. No. 4, Pt. 2, April 1991

correlations between amplitudę spectra at regular intervals within each tonę. The results of these experiments were analyz^d using correlations, multidimensional scaling. factor analyses, and linear regressions. The ratings from the three similarity scaling experiments were highly corre-lated with each other, and they were each moderately correlated with the results of the adjcctive scaling experiment. Acoustic attributes were identified that correlated with the results from the other two methods, thus isolating factors common to all three methods. [Work supported by Sigma Xi, The Scientific Research Society.)

10:30

8MU7. Hystcresis in the perception of directional pitch change. Betty Tuller and Janice Giangrande (Ctr. for Complex Systems, Florida Atlantic Univ., Boca Raton, FL 33431)

It has long been known that pairs of complex tones that are clearly different in pitch may be ambiguous as to the direction of the difference [e.g., R. N. Shepard, J. Acoust. Soc. Am. 36, 2346-2353 (1964)). The present experiment explores whether the ambiguity in pitch change perception results from the dynamie naturę of the perccptual States. Twelve stimulus tones were generated, each 120 ms in duration, and each composed of 10 sequential, harmonically related sinusoida! components. The 12 tones differed by a chromatic step change of the 10 components. Tonę pairs consisted of a standard complex tonę (D#) followed by one of thc other 11 tones. Three conditions were uscd: (1) randomized presentation of the tonę pairs; (2) tonę pairs in which the comparison tonę inereased by one chromatic step in successive tonę pairs (1 vs 2; 1 vs 3; etc.); and (3) tonę pairs in which the comparison tonę decreased by one chromatic step in suecessive tonę pairs (1 vs 12; l vs 11; etc.) Consistent with previous reports, when the tonę comparisons were randomized the direction of pcrceived pitch change was ambiguous as the frequency change neared the half-octave. Signature properties of non-linear dynamical Systems (hysteresis effects and bistability) were ob-served in conditions 2 and 3, for all subjects. Specifically, the region of ambiguity shifted markedly as a function of the ordering of tonal comparisons. Implications for geometrie modcls of pitch perception will be discussed. [Work supported by NIDCD and NIMII.)

10:45

8MU8. Perception of smali rhythmic variations. Peter Marvit (Dept. of Psychol, Univ. of California at Berkeley, Berkeley, CA 94720)

A study to determine the mechanisms and parameters of the perception of smali timing differences in short rhythmic sequences was performed using 13 subjects (which yielded 10 valid subjects) in a completely within-subjccts forced choice factorial design. Manipulated variables included three “base” internote intervals (100, 200, 400 ms— corresponding to different tempos) with rcspective total durations of thc rhythmic figures (1200, 2400, 4800 ms), three catcgories of two pat-tems each of varying rhythmic complexity from Povel and Essens [Musie Percept. 2, 411-440 (1982)], and six variations in internote notę onset timings (0-50 ms). Subjects were played pairs of rhythms, one with the timing variation and one without and then vice versa, and were asked to decide whether the rhythms were the same or different. Scores were based on correct discrimination; rcaction time was recorded, but varied too much to bc useful. All possiblc combinations of factors wcrc presented to each subject and order of presentation varied for each subject. Results based on differences between interval variations, plus the strong showing of one of the rhythmic pattems, strongly supported several aspects of a model for an intemal clock plus a threshold of 30 ms for detection of timing variation. Except for that one pattem, differences between the pattems did not emerge. This casts doubt on the generality of Povel and Essens’ complexity determination of rhythmic figures for pure perception (rather than production). Trends of differences in “base” interval (or tempo) and thus the total duration of the rhythmic sequences, although not statistically significant, suggested possible con-firmation of the limits of precategorical acoustic storage or a model of rhythmic regularity. The unintentional factor of presentation order

121st Meeting: Acoustical Society of America 1988