Auditory Pattern Recognition Research Articles

The pattern-transformation model of pitch.

A new approach to pitch perception is outlined in this paper. It is based on what might be called auditory pattern recognition. The general approach is formalized in a mathematical model, the so-called “pattern-transformation model.” In this model an acoustic stimulus is first transformed by the sense organ into a pattern of peripheral neural activity. This peripheral pattern is assumed roughly to represent the power-spectrum of the stimulus. Thus, the temporal fine structure of the stimulus is virtually ignored; the model is phase insensitive. The peripheral pattern is then assumed to be Fourier transformed into another pattern of activity. This second pattern roughly represents the autocorrelation function of the stimulus. Pitch is derived from the positions of maximal activity in this pattern. From preliminary tests it appears that the model can successfully predict the pitch of many types of complex stimuli. In addition, the model provides estimates of pitch “strength” or “clarity.” These estimates also agree, at least qualitatively, with available data.

Auditory Processing by Dolphins of Signals Related to Signature Whistles

It is now well established (Caldwell and Caldwell, 1970) that one of the major forms—if not the major form—of dolphin communication is the signature whistle. In an effort to obtain information concerning the types and levels of auditory discriminations dolphins must be able to make if they are to process these calls, ah extended series of related studies is being conducted. In the first investigation, the research focus was on auditory pattern recognition within three domains: those of frequency, time, and acoustic configuration. Specifically, a “master” dolphin whistle was patterned and synthesized—as were three other whistles varying only with respect to their frequency, temporal, or spatial characteristics, respectively. Additionally, sets of synthesized whistles were obtained for several equally spaced intermediate points along each of the three experimental continua; all signals were generated through the courtesy of the Speech Transmission Laboratory, Royal Institute of Technology, Stockholm, Sweden. For the three substudies, an experimental animal (an Atlantic Bottlenose dolphin) was administered 50 trials (25 stimuli each) of the “master” whistle and one of the experimental stimuli. Following completion of this procedure, a second set of 50 trials would compare the master whistle with another set of experimental signals; these sets were always ordered from the most different comparisons to the least different within a substudy. Once all data were obtained from one substudy (frequency, say), the whole procedure was replicated for the next. With one exception, the experimental animal was able to discriminate between all stimuli—even those that were almost identical. A discussion of these results and a review of the direction of subsequent research will be provided.

The Pattern-Transformation Model of Pitch

This report describes a new approach to pitch perception, based on a kind of auditory pattern recognition. The general approach is formalized in a mathematical model, the “pattern-transformation model.” In the first stage of this model, an acoustic stimulus is assumed to be “transformed” by the sense organ into what is called a neural activity pattern. This pattern is thought to represent roughly the power spectrum of the waveform. Thus, the model is phase insensitive; the temporal fine structure of the stimulus waveform is ignored. The first-stage pattern (power spectrum) is then assumed to be Fourier transformed (neurally) into another pattern of neural activity. This second pattern roughly represents the autocorrelation function of the stimulus. Pitch is derived from the positions of maximal activity in this pattern. From preliminary tests, it appears that the model can accurately predict the pitches of many types of acoustic stimuli. In addition, the model provides numerical estimates of pitch “strength” and “ambiguity.” While experimental confirmation of these predictions is lacking, the strength estimates agree at least qualitatively with available data. [Research supported by a National Science Foundation Postdoctoral Fellowship awarded to the author and the Institute for Perception—TNO, Netherlands.]

Stimulus information vs processing time in auditory pattern recognition

Previous studies have shown that pitch perception improves with increases in tone duration. An increase in tone duration increases the information available in the stimulus presentation. Other recent experiments have shown that sufficient perceptual processing time is also necessary for accurate pitch perception. The present study determines the relative contribution of stimulus information and processing time in an absolute pitch indentification task. The results indicate that processing time is more critical than stimulus information. The results are described accurately by a model that describes the perceptual process in terms of the information in the stimulus and the time the information is available for perceptual processing.

Design and Development of an Adaptive, Auditory, and Distractive Stressor

An auditory, distractive stressor was developed that automatically adjusts its rate of presenting random digits to human subjects. The rate of digit presentation serves as an inverse index of the amount of attention the subject can spare from a primary task, such as driving a simulated car. The auditory pattern recognition of the device was evaluated under several speaking conditions and for a variety of human speakers. The distractive stressor has a satisfactory digit recognition accuracy, and consequently adapts its digit presentation rate quickly according to how well a subject repeats its given numbers.