UCR Research on
Auditory Event Perception



Research Program and Selected Projects

The goal of our auditory event perception research is to understandhow humans recognize natural sound sources, and how they use acousticinformation to guide behavior. Much of my work in this area hasbeen designed to test between the ecological approach and moretraditional theories of perception. Endemic to the ecologicalapproach is the priority to test subjects under natural, stimulus-richconditions. While this approach has been applied to a number ofissues in visual and haptic perception, very little work has beenconducted in acoustics from the ecological perspective. In fact,most research on auditory perception has focused on low-level,sensory functioning (e.g., sensation of loudness, pitch, timbre,and location of simple tones).

For a number of years, we have been investigating the perceptionof looming sound sources. An early project (Rosenblum,Carello, & Pastore, 1987) tested the relative effectivenessof three acoustic dimensions for specifying the time of passageof an approaching sound source. We found that the dimension whichis available in the most environmental contexts (intensity change)was the most salient for looming judgments. More recent researchtested the accuracy with which listeners can judge the time ofarrival of an approaching sound source when only hearing a smallpart of its approach. This work has practical implications fordeveloping training programs and prosthetic devices for the visually-impaired.One project tested the informational constraints on anticipatoryjudgments (Rosenblum, Wuestefeld, andSaldaña, 1993). Among other findings, these experimentsrevealed that listeners were as accurate at judging time of arrivalwhen hearing only the first 2/3 of a trajectory as when hearingthe full trajectory. This finding is important in demonstratingthe availability of anticipatory or prospective auditory information:a concept central to the ecological approach.

A second project used the auditory looming paradigm to examinethe effects of judgment feedback on perceptual learning (Wuestefeldand Rosenblum, 1996). Most traditional theories of perceptuallearning predict a reversion of performance when feedback is withdrawn.However, the support for this prediction has generally come fromexperiments which test unnatural stimuli with unnatural subjecttasks. In using our more natural auditory looming paradigm, weobserved sustained learning effects even after feedback was withdrawn:a result directly predicted by the ecological approach. We havealso been active in conceptual work on auditory looming whichis reflected in a review chapter (Rosenblum,1993).
We have also conducted research on auditory distance perception.One project (Rosenblum, Wuestefeld, &Anderson, 1996) tested listener accuracy in judging whetherthey could reach a sound source. We observed that blind-foldedlisteners were quite good at 'reachability' judgments. In fact,our listeners were much more accurate than would be expected givenpast research which used unnatural sounds, acoustic environments,and subject tasks. This project also revealed interesting findingsabout the body-scaled informational (acoustic) support for thesejudgments. It seems that the critical information for perceivedreachability might lie within the acoustic structure defined interaurally:a sensible hypothesis given that interaural distance is lawfully-relatedto maximum arm reach.

Our work in progress involves testing whether an oscillating auditoryflow field can induce perceived self-motion as well as posturalchanges in listeners. This project has involved constructing alarge, moveable 'acoustic room' which allows for the shiftingof an entire ambient acoustic space around a stationary listener.Thus far, we have found evidence for perceived self-movement therebyreplicating analogous results in optical flow research. Otherrecent work tests human echolocation. We are currently exploringwhether blindfolded subjects are better at judging the distanceof a sound-reflecting surface when moving or standing still.


Relevant References

Rosenblum, L.D., Carello, C., Pastore, R.E. (1987). Relativeeffectiveness of three stimulus variables for locating a movingsound source. Perception, 16, 175-186.

Rosenblum, L.D., Wuestefeld, A.P., and Saldaña, H.M (1993).Auditory Looming Perception: Influences on Anticipatory Judgments.Perception. 22, 1467-1482.

Rosenblum, L.D. (1993). Acoustical information for controlledcollisions. In A. Schick (Ed.), Contributions to PsychologicalAcoustics. Oldenburg, Germany: Bibliotheks- und Informationssystemder Carl von Ossietzky Universität Oldenburg.

Rosenblum, L.D., Wuestefeld, A.P., and Anderson, K.L. (1996).Auditory reachability: An affordance approach to the perceptionof sound source distance. Ecological Psychology. 8(3), 1-24.

Wuestefeld, A.P. & Rosenblum, L.D.(1996). Perceptual learning in an auditory time to arrival task.University of California, Riverside Cogntive Science TechnicalReport #UCR.96.011.


Abstracts

Rosenblum, L.D., Carello, C.,Pastore, R.E. (1987). Relative effectiveness of three stimulusvariables for locating a moving sound source. Perception, 16,175-186.
A study is reported in which it is shown that observers can useat least three types of acoustics variables that indicate reliablywhen a moving sound source is passing: interaural temporal differences,the Doppler effect, and amplitude change. Each of these variableswas presented in isolation and each was successful in indicatingwhen a (simulated) moving sound source passed an observer. Thesethree variables were put into competition (with each indicatingthat closest passage occurred at a different time) in an effortto determine their relative importance. It was found that amplitudechange dominated inteaural temporal differences which, in turn,dominated the Doppler effect stimulus variable. The results arediscussed in terms of two interpretations. First, it is possiblethat subjects based their judgments on the potential discriminabilityof each stimulus variable. However, because the stimuli used involvedeasily discrimable changes, subjects may instead have based theirjudgments on the independence of a stimlus variable from differentenvironmental situation conditions. The dominance ordering obtainedsupports the second interpretation.

Rosenblum, L.D., Wuestefeld, A.P.,and Saldaña, H.M (1993). Auditory Looming Perception: Influenceson Anticipatory Judgments. Perception. 22, 1467-1482.
Several studies in the auditory perception literature hint thatlisteners may be able to anticipate the time of arrival of anapproaching sound source (Rosenblum, L D, Carello, C, Pastore,R E, 1987; Schiff, W, Oldak, R, 1990). In two experiments listenersjudged the time of arrival of an approaching car based on variousportions of its auditory signal. Subjects pressed a computer keyto indicate when the car would have just passed them assumingthe car maintained a constant approach velocity. A number of variableswere tested including (a) the time between the offset of the signaland the virtual time of passage (b) duration of the signal (c)feedback concerning judgment accuracy. Results indicate that increasingthe time between signal offset and virtual time of passage decreasesjudgment accuracy while the actual duration of the signal hadno significant effect. Feedback significantly improved performanceoverall.

Rosenblum, L.D. (1993). Acousticalinformation for controlled collisions. In A. Schick (Ed.), Contributionsto Psychological Acoustics. Oldenburg, Germany: Bibliotheks- undInformationssystem der Carl von Ossietzky Universität Oldenburg.
From the ecological approach to perception, all behavior can beunderstood as the control of collisions with objects of the environment.It is incumbent on the ecological psychologist to understand theprospective information which allows an animal to anticipate upcomingcollisions. In this chapter, prospective acoustic informationfor controlled collisions is discussed borrowing from work onvisual time-to-arrival information. Potential acoustic dimensionswhich might specify time-to-arrival are surveyed. Two specificexamples of auditorily-controlled collisions are examined: 1)a moving observer approaching a stationary surface, and 2) a loomingobject approaching a stationary observer. The research conductedon these two scenarios is reviewed along with conjectures abouttheir informational support.

Rosenblum, L.D., Wuestefeld, A.P.,and Anderson, K.L. (1996). Auditory reachability: An affordanceapproach to the perception of sound source distance. EcologicalPsychology. 8(3), 1-24.
The literature on perception of sound source distance revealsa wide range of listener accuracy. Most experiments have listenersperform unintuitive tasks, using unnatural sounds presented inimpoverished acoustic environments. The present experiments implementan affordance paradigm for which listeners judge the 'reachability'of a natural, live sound source in a familiar acoustic environment.Results reveal that listeners are quite accurate in judging whetherthe source is reachable and are sensitive to the advantage affordedby two vs. one degree of freedom reaches. Further analyses revealthat when scaled to an intrinsic bodily dimension, judgment differencesbetween listeners disappear, implicating intrinsically scaledspecificational information. A follow-up experiment explores thepotential informational support for these judgments testing theusefulness of head movements and binaural hearing. Results revealthat whereas head movements had no bearing on either judgmentaccuracy or consistency, binaural information did enhance listenerconsistency. This could suggest that the allometric relation betweeninteraural distance and arm length might provide a basis for auditoryreachability judgments.

Wuestefeld, A.P. & Rosenblum, L.D.(1996). Perceptual learning in an auditory time to arrival task.University of California, Riverside Cogntive Science TechnicalReport #UCR.96.011
The present research examines accuracy in an auditory time toarrival task when performance feedback is first provided to listenersand is subsequently withdrawn. Previous research testing visualtime to arrival suggests that while performance feedback improvesaccuracy, this improvement is not maintained after feedback iswithdrawn. While these findings are supportive of a Cognitiveapproach to looming perception, they are not supportive of anEcological approach which predicts a retention of performanceupon feedback withdrawal. In an effort to determine which approachbest construes the perceptual learning occurring for auditorytime to arrival tasks, listeners were provided with feedback whichwas then withdrawn. Listeners made judgments about the time toarrival of an approaching car based on various portions of theevent. Subjects participated in three experimental sessions onconsecutive days. The experimental group received no feedbackduring the first session, feedback during the second session,and no feedback during the final session. When feedback was withdrawn,the higher performance level attained during training was retained.These results support the Ecological approach to perceptual learningin auditory time to arrival tasks.

Hits since 1/14/98