In the present study, education majors minoring in music education (n = 24) and music performance majors (n =14) read and performed the original version and melodically altered versions of a simple melody in a given tempo. Eye movements during music reading and piano performances were recorded. Errorless trials were analyzed to explore the adjustments of visual processing in successful performances. The temporal length of the eye–hand span (time between gaze and the performed note) was typically around one second or less. A measure of gaze activity indicated that performers generally inspected two quarter-note areas between two metrical beat onsets. The performance majors operated with shorter fixation durations and applied larger eye–hand spans as well as greater gaze activity than education majors. The latter two measures were generally affected by unexpected melodic alterations and simple rhythmic patterns. The study manifests both the flexibility and limitations of the mechanisms of visual processing in temporally controlled music reading, addressing some of the everyday conceptions about sight-reading by means of systematic research.

A music reader has to “look ahead” from the notes currently being played—this has usually been called the Eye-Hand Span. Given the restrictions on processing time due to tempo and meter, the Early Attraction Hypothesis suggests that sight readers are likely to locally increase the span of looking ahead in the face of complex upcoming symbols (or symbol relationships). We argue that such stimulus-driven effects on looking ahead are best studied using a measure of Eye-Time Span (ETS) which redefines looking ahead as the metrical distance between the position of a fixation in the score and another position that corresponds to the point of metrical time at fixation onset. In two experiments of temporally controlled sight reading, musicians read simple stepwise melodies that were interspersed with larger intervallic skips, supposed to create points of higher melodic complexity (and visual salience) at the notes following the skips. The results support both Early Attraction (lengthening of looking ahead) and Distant Attraction (lengthening of incoming saccades) in the face of relative melodic complexity. Notably, such effects also occurred on the notes preceding the nominally complex ones. The results suggest that saccadic control in music reading depends on temporal restrictions as well as on local variations in stimulus complexity.

In this study the effects of skill development on the eye movements of beginning adult sight-readers were examined, focusing on changes in the allocation of visual attention within metrical units as well as in the processing of larger melodic intervals. The participants were future elementary school teachers, taking part in a 9-monthlong music training period. During this period, 15 novice sight-readers' development was observed in three measurements, with 15 amateur musicians functioning as a comparison group. The novices' allocation of fixation time within metrical units gradually approached a pattern demonstrated by the amateurs in which increased sensitivity to metrical divisions was evinced by larger average fixation times on the latter halves of bars. Concerning larger melodic skips in otherwise stepwise melodic contexts, an analysis of fixation times suggested that the novices' visual processing of skips did not proceed in terms of note comparison across the skip but rather through a direct identification of the notational symbols involved. Skill development was seen, then, as increasing fluency of this identification process. These and similar findings may lead to a better understanding of the problems encountered by novice sight-readers and thus to advancements in the pedagogy of music reading.

This article addresses the silent reading of music notation, combining eye-movement measures with a semantic analysis of readers' verbal descriptions of the notated music. A group of musical novices (n = 16) and two groups of musical amateurs (less experienced n = 11 and more experienced n = 10) participated in three separate measurement sessions during a nine-month-long university music course designed for future primary-school teachers. In each session they viewed a notated folk song for 30 s and then described what they had seen. Greater musical experience was found to be connected with shorter fixation durations, more linear scanning of the notated music, and more accurate and integrative verbal descriptions. Repeated measurements during the course were connected to more accurate and integrative descriptions as well as the use of longer saccades for all participants. A cluster analysis of the results revealed three separate silent-reading styles: elementary processors, lacking in both accuracy and integration in their verbal reports; accurate analyzers, producing accurate but non-integrative descriptions; and accurate integrators who, besides accurate and integrative descriptions, showed a tendency of using shorter fixation times than the two other groups. The results are discussed with a view to visual processes in other domains. In addition, a question is raised whether educational programs tend to unduly neglect considering music notation as a source of information beyond its role as a performance aid.