Visual Information Pianists Use for Efficient Score Reading

Abstract

When sight-reading music, pianists have to decode a large number of notes and immediately transform them into finger actions. How do they achieve such fast decoding? Pianists may use geometrical features contained in the musical score, such as the distance between notes, to improve their efficiency in reading them. The aim of this study is to investigate the visual information pianists rely on when reading music. We measured the accuracy of the musical score reading of 16 skilled pianists and investigated its relationship with the geometrical features. When a single note was presented, pianists easily read it when it was located within three ledger lines. When two notes with an octave interval were presented, interestingly, their readable range was extended compared to that of the single note. The pianists were also able to recognize the octave interval correctly even if they misread the height (or pitch) of the target notes. These results suggest that the pianists decoded two notes composing an octave interval as a single "two-tone geometric pattern." Analyzing the characteristics of incorrect responses, we also found that pianists used the geometrical features of the spatial relationship between the note head and the ledger line, and that the cause of the misreading could be categorized into four types: [Type I] Confusion to a neighboring note having the same ledger line configuration; [Type II] Interference from a commonly used height note having the same note name; [Type III] Misunderstanding based on the appearance probability; [Type IV] Combination of the above three. These results all indicate that the pianists' abilities in score reading rely greatly on the correlation between the geometric features and playing action, which the pianists acquired through long-time training.

Keywords: appearance probability; geometrical features; pianist; readable range; sight-reading.

Figure 1

Terminology and example of target notes on G-clef: (A) single-note condition, (B) octave-interval condition, and (C) seventh-interval condition.

Figure 2

Range of target notes: (A) single-note condition, (B) octave-interval condition, and (C) seventh-interval condition.

Figure 3

Flow of the experimental procedure.

Figure 4

The proportion of the participants who gave correct answers to the target notes on the G-clef: (A) single-note, (B) octave-interval, and (C) seventh-interval conditions. The region of the target notes is indicated by the high-contrast area on the keyboard. In the octave (B) and the seventh intervals (C), the note name (given in the lower part of the keyboard) indicates the “higher note” contained in the intervals (black-colored in the score). The white bars indicate the target notes without any accidental marks. The blue- and red-colored bars show the target notes with sharp (♯) or flat (♭), respectively, in the case that the accidental marks mean playing black keys. The blue- and red-colored triangles show the target notes with sharp and flat, respectively, in the case that the accidental marks indicate playing white keys (C♭, B♯, E♯, and F♭). The red dotted line indicates the 80% correct rate. The orange- and green-colored outlines indicate notes within the lower- and upper- limits of the readable notes (that is, where the correct response rate was over 80%), respectively.

Figure 5

The proportion of the participants who gave correct answers to the target notes on the F-clef: (A) single-note, (B) octave-interval, and (C) seventh-interval conditions. The symbols mean the same as in Figure 4, except that the note name represents the “lower note” contained in the intervals (B,C).

Figure 6

Readable range for each type of target note.

Figure 7

Ratios of played key color are indicated for each key color of the presented note, when the participants made incorrect responses under the single-note condition. The horizontal indexes “White”, “Black”, and “White*” represent the presented notes indicating white keys, black keys, and white keys with an accidental mark, respectively.

Figure 8

The number of incorrect responses is indicated for each target note: (A) single-note, (B) octave-interval, and (C) seventh-interval conditions. Classification of incorrect responses are indicated by the color of the bars. The horizontal indexes represent the presented note name, and the vertical indexes represent an octave number of the presented target note. The left and right panels indicate the target notes on the F-clef and the G-clef, respectively. In the case of the octave- and seventh-interval conditions, each note name and octave number corresponds to the lower note contained in the intervals. The orange- and green-colored diamonds represent notes within the lower- and upper- limits of the readable notes in the single note condition, respectively. The yellow background indicates the unreadable range.

Figure 9

The ratio of errors in interval in the octave- and seventh-interval conditions. The amount of error was represented based on the unit of semitone.

Figure 10

Example of the method of categorization of the C6 note. The notes in the left score (A) is on the ledger line that is the same as C6. The notes in the right score (B) above the ledger line that is different from C6.

Figure 11

Example of positional errors to the C note.

Figure 12

Example of a Type III error.