Error detection —— the ability to verify the correspondence between music-as-heard and music-as-written —— is the sine qua non for musicians. A performer in the process of learning a new piece of music plays the score, simultaneously questioning the accuracy of the rendition. A private lesson teacher, even when focusing on a student’s technique or musicality, listens for correct pitches and rhythms. And conductors, having already studied the score to be rehearsed, must detect any differences between two versions of that score: the mental ear’s rendition, and the ensemble’s. Error detection is a more commonly exercised aural skill for most musicians than melodic, harmonic, or rhythmic dictation. These latter activities, however, remain the focus of ear training courses, even though the relationship between these activities and achievement in error detection has not been proven conclusively.²

To practice error detection requires two items: a musical score and an incorrect rendition of that score. The World Wide Web is an ideal medium for guided practice in error detection because sound files are separate from graphic files. Thereby, a sound file may or may not render correctly an associated graphic file. However, both graphic and sound files begin as the same score in a music processor. A screen shot of the excerpt is transformed into a GIF file and incorporated into HTML as an image source. To create a sound file, export a score as a MIDI file, transform it into a movie, and play the embedded source using QuickTime.

This paper introduces three web-based models for practicing error detection. While creating these models, simplicity and pedagogical functionality were the guiding principles. For example, each model seeks to maximize instruction while using the fewest number of linked pages. And, each page contains a minimal amount of graphics, sound, color, and text, so that downloading time does not interrupt a student’s focus and flow while completing an exercise. Another consideration is the physical dimension of each notated musical example: a graphic file must fit onto an average-size screen, so that it is unnecessary to scroll in order to view the complete score. Naturally, this requirement limits the length of musical excerpts.

A pedagogical goal of all three models is to provide instructional feedback when wrong answers are selected. Feedback must assist the user without giving away the correct answer. Thereby, all three models include "help" pages, as well as the opportunity to hear and see the difference between the correct answer and any incorrect ones. These "help" pages complement the objective to increase a user’s ability to associate musical sound with its visual representation.

The first model contains seven linked pages. Page 1 presents a score and its opening pitches with the instructions "Study the score." Beneath the score are three renditions of it; one is correct and two contain errors. The objective is to identify the correct rendition of the score. Example 1 outlines a design in which rendition #2 is correct, indicated by the straight line down the middle. When rendition #2 is selected as the answer, text on the linked page identifies it as a correct answer and prompts a return to the index of error detection exercises.

When either rendition #1 or #3 is selected, text on the linked page informs the user of a poor choice. On example 1, the double-arrow lines connecting page 1 to renditions #1 and #3 indicate that when either wrong rendition is selected, an option is to go back to page 1 and try again. Alternatively, the user may go forward in search of help.

The help pages present the selected wrong rendition in both audio and visual formats, so that the user can associate abstract sound with its concrete notation. Beneath the score of the wrong rendition appears the desired score originally shown on page 1. Text instructs the user to compare the two scores and return to page 1 to try again.

Example 1. Diagram for "Study the Score"

The second model, the objective of which is to improve the user’s ability to hear harmonic progressions, contains eight pages. The first page presents a musical rendition of a bass line along with its score. Beneath the score is the Roman numeral analysis of a progression suitable for the given bass line. After listening to the bass line, a user is instructed to hear the progression in his/her mental ear and to identify which one of four renditions corresponds to the given progression.

Example 2 places page 1 in the center; rendition #3 is the correct answer from which the user may return to the error detection index. When a wrong answer is selected, the user may immediately return to page 1, as suggested by the double-arrow lines, or proceed to an associated help page. The help pages in this model are similar to those in the previous one, in that they allow a user to compare the selected incorrect answer with the desired answer.

Example 2 Hear the Harmonies.

Example 3 diagrams an error detection exercise in which the user locates and corrects discrepancies between a score and a rendition of that score. Page 1 displays a musical score that is a clickable map. After listening to the musical excerpt, the user discerns which pitches and/or rhythms are notated incorrectly, and clicks on the notes which are heard as wrong.

Clicking on any note links either to "Yes" or "No." The former affirms that the user has identified a notational error, while the latter a correct note. On example 3, the double arrows connecting Page 1 to both "Yes" and "No" indicate the option to return and detect other errors that may be present. Alternatively, a user can progress to the next step in the exercise at either Page 2.1 or 2.2. Both pages are accessible via page 1, as well.

The text on page 2.1 confirms that the user has correctly identified the errors, while that on page 2.2 assists the user in finding some, if not all, of the errors. Both pages duplicate the graphic and sound files from page 1. Now, however, the score displays an "X" above or below each notated error, confirming the location of all wrong notes.

With knowledge of which notes are notated incorrectly, the user is prompted to visualize the correct notation as shown on page 3. In addition, the final page again displays the score from page 1 with all its errors; but this time, an associated sound file provides a rendition of the flawed score, in order to reinforce the correspondence between music-as-heard and music-as-seen.

Example 3 Find and Fix the Errors.

_________________________________

1. Online access: www.utexas.edu/courses/mus411/edintro.html

2. Byo, James L. (1997). The Effects of Texture and Number of Parts on the Ability of Music Majors to Detect Performance Errors. Journal of Research in Music Education, 45, 51-66.