On the registration of time and the patterning of speech movements.

In order to study speech coordination we frequently average kinematic and other physiological signals. The averages are assumed to be more representative of the underlying patterns of production than individual records. In this note we outline different approaches to averaging and present a new nonlinear normalization technique that offers better information than ensemble averaging, linear normalization, or feature alignment methods. We suggest that this technique provides a clear estimation of pattern shape while preserving information on the variation over time.

Multidimensional scaling of synthetic musical timbre: perception of spectral and temporal characteristics.

The perceptual correlates of acoustic parameters involved in musical timbre were investigated by examining judgements of timbre dissimilarity. Nine synthetic sounds were created, derived from crossing three levels of spectral and temporal parameters (number of harmonics and rise time, respectively). Two separate conditions were tested, one using single tones, the other using short melodies. Fifteen musically untrained subjects were presented with pairs of stimuli and asked to judge dissimilarity on an 8-point scale. The spatial configuration resulting from multidimensional analysis of the data was best fit by a three-dimensional model, with the first two dimensions accounting for most of the variance. The perceptual space derived from the analysis indicates that these two orthogonal dimensions corresponded closely to the spectral and temporal differences inherent to the stimuli. Similar results were obtained with both melodies and single tones. A second experiment replicated the findings despite the introduction of random loudness variation, indicating that timbre judgements can be made independently of loudness. We conclude that even musically unselected subjects are sensitive to spectral and temporal information in musical tones, and are able to use them independently in making perceptual judgements of musical timbre.

Functional data analyses of lip motion.

The vocal tract's motion during speech is a complex patterning of the movement of many different articulators according to many different time functions. Understanding this myriad of gestures is important to a number of different disciplines including automatic speech recognition, speech and language pathologies, speech motor control, and experimental phonetics. Central issues are the accurate description of the shape of the vocal tract and determining how each articulator contributes to this shape. A problem facing all of these research areas is how to cope with the multivariate data from speech production experiments. In this paper techniques are described that provide useful tools for describing multivariate functional data such as the measurement of speech movements. The choice of data analysis procedures has been motivated by the need to partition the articulator movement in various ways: end effects separated from shape effects, partitioning of syllable effects, and the splitting of variation within an articulator site from variation from between sites. The techniques of functional data analysis seem admirably suited to the analyses of phenomena such as these. Familiar multivariate procedures such as analysis of variance and principal components analysis have their functional counterparts, and these reveal in a way more suited to the data the important sources of variation in lip motion. Finally, it is found that the analyses of acceleration were especially helpful in suggesting possible control mechanisms. The focus is on using these speech production data to understand the basic principles of coordination. However, it is believed that the tools will have a more general use.

Comparison of height acceleration curves in the Fels, Zurich, and Berkeley growth data.

A new type of smoothing spline designed for nonparametric estimation of growth acceleration curves was applied to longitudinal measures of height from the Fels, Zurich, and Berkeley growth studies. Landmark features of the individual acceleration curves were located and examined for site effects and site by sex interaction. The curves in each group were structurally averaged by transforming the age scale of each child to align the landmarks with overall means. The modal shapes of the structural average curves were found to be highly similar in the three studies, although the relative number of children who conformed to the modal pattern differed among the studies. Significant site by sex interactions were found in the timing of the landmarks associated with the pubertal growth spurt and with the prepubertal 'midspurt'.