Previous research suggests that individuals with a Vocal Pitch Imitation Deficit (VPID, a.k.a. “poor-pitch singers”) experience less vivid auditory images than accurate imitators (Pfordresher & Halpern, 2013), based on self-report. In the present research we sought to test this proposal directly by having accurate and VPID imitators produce or recognize short melodies based on their original form (untransformed), or after mentally transforming the auditory image of the melody. For the production task, group differences were largest during the untransformed imitation task. Importantly, producing mental transformations of the auditory image degraded performance for all participants, but were relatively more disruptive to accurate than to VPID imitators. These findings suggest that VPID is due partly to poor initial imagery formation, as manifested by production of untransformed melodies. By contrast, producing a transformed mental image may rely on working memory ability, which is more equally matched across participants. This interpretation was further supported by correlations with self-reports of auditory imagery and measures of working memory.

We propose a new framework to understand singing accuracy, based on multi-modal imagery associations: the MMIA model. This model is based on recent data suggesting a link between auditory imagery and singing accuracy, evidence for a link between imagery and the functioning of internal models for sensorimotor associations, and the use of imagery in singing pedagogy. By this account, imagery involves automatic associations between different modalities, which in the present context comprise associations between pitch height and the regulation of vocal fold tension. Importantly, these associations are based on probabilistic relationships that may vary with respect to their precision and accuracy. We further describe how this framework may be extended to multi-modal associations at the sequential level, and how these associations develop. The model we propose here constitutes one part of a larger architecture responsible for singing, but at the same time is cast at a general level that can extend to multi-modal associations outside the domain of singing.

The development of singing accuracy, and the relative role of training versus maturation, is a central issue for both music educators and those within music cognition. Although various studies have focused on singing accuracy in different age groups, to date we know of no data sets that maintain the consistency in recruitment, methodology, and measurement that is necessary to make direct comparisons. We report analyses of three data sets that meet these criteria: two groups of children (kindergarten, middle school), and one group of adults (college aged). The data were collected at different times, but used a similar set of tasks and identical scoring procedures. Results indicate considerable improvement in accuracy from kindergarten to late elementary that dramatically reverses such that college students perform at the level of kindergartners. It appears singing accuracy may be related to variables involving singing experience rather than general development, and singing skill could decline over time if not maintained through engagement. A secondary purpose was to explore the efficacy of acoustic scoring for some singing tasks and how well it mimics human judgments of accuracy. The acoustic scoring procedure was highly correlated with expert judgment and could provide a standard approach to scoring that is largely automated. We discuss the potential benefits of a more unified approach to measuring singing accuracy and suggest future research that includes children, adolescents and adults in the sample.

We report two experiments exploring whether matched metrical and motivic structure facilitate the recognition of melodic patterns. Eight tonal melodies were composed from binary (four-note) or ternary (three-note) motivic patterns, and were each presented within a metrical context that either matched or mismatched the pattern. On each trial, participants heard patterns twice and performed a same-different task; in half the trials, one pitch in the second presentation was altered. Performance was analyzed using signal detection analyses of sensitivity and response bias. In Experiment 1, expert listeners showed greater sensitivity to pitch change when metrical context matched motivic pattern structure than when they conflicted (an effect of metrical encoding) and showed no response bias. Novice listeners, however, did not show an effect of metrical encoding, exhibiting lower sensitivity and a bias toward responding “same.” In a second experiment using only novices, each trial contained five presentations of the standard followed by one presentation of the comparison. Sensitivity to changes improved relative to Experiment 1: evidence for metrical encoding – in the form of reduced response bias when meter and motive matched – was found. Results support the metrical encoding hypothesis and suggest that the use of metrical encoding may develop with expertise.

THE TERM "TONE DEAFNESS," COMMONLY APPLIED TO poor-pitch singing, suggests that the cause lies in faulty perception. However, it is also plausible that problems lie in production, memory, and/or sensorimotor integration. We report the results of two experiments on vocal pitch imitation that addressed these possibilities. Participants listened to and then vocally imitated unfamiliar 4-note pitch sequences.Within each experiment, 10-15% of the participants imitated pitch at least one semitone off and were categorized as "poor-pitch singers." Such deviations were reliable across different pitch classes and therefore constitute transpositions. In addition, poor-pitch singers compressed the size of intervals during production. Poor-pitch singers did not differ from good singers in pitch discrimination accuracy, although they appeared to be hindered rather than helped by singing with correct accompaniment. Taken together, findings suggested that poor-pitch singing results from mismapping of pitch onto action, rather than problems specific to perceptual,motor, or memory systems.

Two experiments investigated the perception of melodic and rhythmic accents in musical patterns. Musiclike patterns were created in which recurring melodic and/or rhythmic accents marked higher order periods that, when both accents were present, could differ in terms of period and/or phase according to the construct of joint accent structure (M. R. Jones, 1987). Listeners were asked to indicate the location of accents in these patterns by tapping to tone onsets. Each experiment pursued two main questions. First, are accents, as manipulated, salient to listeners? Second, do listeners track higher order time spans formed by melodic and rhythmic accents in a way that shows a sensitivity to interrelationships between melody and rhythm? Results supported affirmative answers to these questions in analyses of tapping locations and time spans between taps, respectively. Furthermore, results suggested that accents function as temporal landmarks that listeners can use when tracking the time structure of musical patterns, and that the complexity of this time structure arises from higher order time spans marked by different types of accents.