Deprecated: mysql_connect(): The mysql extension is deprecated and will be removed in the future: use mysqli or PDO instead in /home/ampsocie/public_html/include/database.php on line 29

Warning: session_start(): Cannot send session cookie - headers already sent by (output started at /home/ampsocie/public_html/include/database.php:29) in /home/ampsocie/public_html/include/session.php on line 55

Warning: session_start(): Cannot send session cache limiter - headers already sent (output started at /home/ampsocie/public_html/include/database.php:29) in /home/ampsocie/public_html/include/session.php on line 55

Warning: session_regenerate_id(): Cannot regenerate session id - headers already sent in /home/ampsocie/public_html/include/session.php on line 56
AMPS: Past Seminar Series

Past Seminar Series

Select the that year you would like to see

2000 Seminar Series

A Pitch in Time: An Artificial Neural Network of Melodic Expectancy
Speaker : Kate Stevens MARCS, University of Western Sydney
Date :24th Mar


A contemporary theory of melody cognition - the implication-realization (IR) theory (Narmour, 1990, 1991, 1992) - draws on Gestalt principles of organization such as proximity, similarity, and closure as being the "given code" for perception of melodic patterns. Interestingly, a range of experimental evidence is emerging that suggests that human listeners are sensitive to the principles of IR theory (Cuddy & Lunney, 1995; Schellenberg, 1996; Thompson, Cuddy & Plaus, 1997). One assumption of Narmour's model is that many expectancies in melody cognition can be explained, to a large degree, by structures that arise from local, tone-to-tone transitions. Artificial neural networks operate according to similar bottom-up processes and if such networks are to be regarded as valid and informative models of human perception and cognition then they need to reflect principles espoused in contemporary theory and their results need to be compared directly with human performance. This study examines aspects of Narmour's theory and explores the development of expectancies and representation of pitch within a neural network framework.

It was hypothesized that an artificial neural network model of melody cognition would exhibit signs of IR principles during the course of learning and while performing a melody prediction task. Specifically, after exposure to examples of Western tonal melodies, the network constructs a set of connection strengths and hidden unit activations that permit a) prediction of the next note in familiar melodies; and b) prediction of the note following an implicative interval that conforms with Narmour's principles. Two feed-forward networks were constructed and exposed to identical sets of Western tonal melodies as training and test patterns. The first back-prop network was designed to encourage the construction of an interval code; an interval code uses the pitch distance between each two successive tones and ignores the actual pitch values. Melody and interval prediction accuracy of this network was compared with output from a single-layer linear network. Performance of the linear network was surprisingly good although relatively poor recognition of transposed melodies indicated that there was minimal representation of key.

The two models provide an existence proof that principles central to Narmour's model of bottom-up melodic expectancy can be learned by exposure to a set of Western tonal melodies. Design modifications are planned to enable inclusion of duration, amplitude features, and a more coarse coding of pitch. The existence of "exceptions" to IR principles in music motivates a model of expectancy based on adaptive mixtures of local experts (Jacobs, Jordan, Nowlan & Hinton, 1991). Proposed models that examine the development of pitch, interval, scale step and key relations from exposure to the auditory environment are also discussed.

The Point of the Dot. A perceptual approach to understanding performance practice issues
Speaker : Emery Schubert & Dorottya Fabian Somorjay University of NSW
Date :7th Apr


Historically informed performance is based, to a large degree, on information from writings of contemporaries, and assertions made by musicologists and performers. These approaches to informedness, we hypothesise, ignore one of the crucial aspects of the musical experience - the perception of the listener. In a controlled study, 70 volunteers made responses to an excerpt from Bach's Goldberg Variations (number 7, gigue) in which note durations, tempo and inter-onset-intervals were manipulated. Participants rated each excerpt according to 'dottedness', preference and perceived stylishness. Preliminary results suggest that there are important individual differences in responses, but that, in general, there is more to the ratio of note durations than just the dotting ratio. For example, the gap (or 'kerning') between notes affects the perception of dotting. This is a factor which has not received enough emphasis from past performance practitioners, yet the finding has serious implications for the character and stylishness of a piece. This presentation discusses information about these issues in the context of the recently collected data.

Input/output: The strange cases of two organists and a king
Speaker : Robert Walker University of NSW
Date :5th May


I am no Skinnerian, but one cannot ignore the issue of input/output in human behaviour. An alternative to the influence of the environment is determinism. Of course it cannot be all one and not the other, but I am only concerned with the environment since there are no reliable data on deterministic influences, nor is our present state of science capable of providing any. Following Kate Stevens' interesting and lucid exposition of her work with artificial neural networks, I thought it might be interesting to try and articulate how the ultimate neural network - the actual brain - might operate in humans but purely from the known details of the input from the environment to the output of observable musical behaviour. The actual networks involved must remain, in this case, as a black box, because there is no way that human brains can be studied in order to account for the kind of things I am interested in musical behaviour as defined in musical terms.

Anecdotally, several psychologists in the USA and the UK have said to me that psychology can only exist as the psychology of something specific, like fishing or piano playing. Others have suggested that psychology now has only one of two ways forward: chemistry and physics, or biography and ethnography. I am using the latter in this paper.

I present 3 case studies whereby I explain in as much detail as time allows the respective musical environment, the type of musical experiences and other influences which provided musical 'input', including attitudes about music, in the early, formative years, of an 18th century English organist, composer and novelist who knew Handel, John Alcock, a famous choral conductor and choir trainer of the second half of the 20th century who is still alive, Sir David Willcocks, born in 1919, and the 'king' himself, Elvis. I compare and contrast the environment of each and influences on each, and the musical behaviour of each in order to try and show some connections between 'input' and 'output' in musical terms. In the process I will try and shed light on what might be impossible about the task of artificially replicating the development of the human brain in processing music so as to induce musical behaviour.

The arguments will be entirely musical but with ethnographic and psychological flavouring. I show the music which each of the three would have experienced as children, and compare these musical experiences with what they each produced in adulthood. Details on the two organists have been drawn from recent investigations I have carried out from primary sources in the UK. In the case of the 'King' I have relied on some existing secondary sources in published materials about his life. Video clips and/or audio tapes will demonstrate the music of each as both 'input' and 'output'.

A Cross-Cultural Investigation of Musical Tastes of Korean, American, and Australian University Students Focusing on Informal and Formal Musical Experiences
Speaker : Myung-sook Auh University of Technology-Sydney & Robert Walker U
Date :19th May


The purpose of the study was to determine whether Korean students' musical taste has become westernized by comparing their musical taste with American and Australian students' musical tastes. Total 260 students-92 Korean, 65 American, and 103 Australian-participated in answering the Musical Taste Questionnaire (MTQ) and the Musical Experiences Questionnaire (MEQ) on their university campus. The results showed that: 1) Korean students' ratings for Korean Popular music were significantly higher than their ratings for Western Popular music; 2) Korean students' ratings for Korean traditional music was significantly higher than American and Australian students' ratings for Asian traditional music; 3) There were no significant differences in ratings for Western Classical music between Korean and American students; and 4) Korean students' ratings for Western Classical music were significantly higher than their ratings for Korean traditional music. The results indicate that Korean musical taste has not become westernized completely. It retains an element of uniqueness.

Computerized Adaptive Testing of Music Listening Skill
Speaker : Walter Vispoel University of Iowa
Date :13th Oct


Recent advances in computer technology and measurement theory have led to the development and growing use of computerized adaptive tests (CATs) in large-scale assessment programs. CATs are appealing to many test users because such tests can reduce test length and testing time (often by 50%) but still maintain high measurement precision by tailoring the administration of items to each examinee's estimated skill level. During this presentation, I will discuss recent research findings, which demonstrate that CATs may significantly reduce problems of inefficiency and low measurement precision that have long hindered the assessment of music listening skills.

Promoting Will, Skill and Thrill in the Classroom: An Integrated View of Learner Motivation in Music Contexts
Speaker : James Austin University of Colorado at Boulder
Date :20th Oct


Social cognition provides the most useful theoretical paradigm for understanding and enhancing learner motivation. Factors that limit the application of motivation research to teaching and learning include conceptual imprecision, theoretical proliferation, polarization of research agendas, and limited contextual grounding. Sample studies of motivation conducted in music contexts will be presented and future directions for motivation research will be explored with audience members.

Understanding Choice through Categorizations of Self and Context: A new theory for the research area of musical instrument choice
Speaker : Samantha Pickering University of Sydney
Date :27th Oct


In the investigation of musical instrument choices of children, research frequently suggests that children hold gender-typed beliefs about instruments and that these beliefs inform instrument choice. However, these gender-typed beliefs are typically assumed from the patterns of choice behaviour identified in groups of children; the views that these children may have held about instrument choice are ignored. Consequently, it is proposed that further research should focus on the views of participants by making use of the notion of categorization. A database analysis of demographics, choice and scores of Higher School Certificate (HSC) music students was conducted revealing the relationship between instrument choice and a number of factors, including gender, elective focus and score in the performance examination. Future studies involving questionnaire and interview data from selected schools will highlight whether such factors are consistent with the views that music students hold while in the context of instrument choice for the HSC.

The Perception of Musical Sounds with Cochlear Implants
Speaker : Thomas Stainsby MARCS, University of Western Sydney
Date :10th Nov


Following widespread success with speech communication, the recipients of cochlear implants often express a desire for the improved perception of music. Inspired by the ultimate aim of delivering better musical appreciation to these listeners, the project described here investigated the perception of musical timbre. Three psychophysical experiments were used to study the perceived frequency spectra from the steady-state portions of 10 musical instrument sounds. The shapes of the internal spectra were measured using a forward-masking paradigm, and the same stimuli were subsequently employed in discrimination and identification tasks. The results reveal much information about the frequency selectivity and spectral detail available to normally hearing listeners and cochlear implant users.