Because the voice is part of the human body, vocalists feel a personal identification with their instrument as a part of themselves, rather than an outside entity that is manipulated to produce music. A bassoon or a piano can be seen as technology, but the larynx of the singer is part of the human anatomy. The nature of the voice lesson is an intimate relationship between the teacher and the student and the tradition of singing has been passed down by word of mouth from teacher to student.

The purpose of this study is to observe and measure the impact of technology during an eight-week series of voice lessons. The technology was an integral part of the lesson format, but was not the primary method of instruction. The technology functioned as a supplement to hands-on teaching. I suggest which technologies are feasible and provide research-based strategies for incorporating these technologies. Six students received eight 45 minute voice lessons.

To what extent did the presence and use of technology impact the teacher’s ability to provide a viable voice lesson and the participant’s attitudes toward the learning process? All research questions were addressed by the analysis of weekly logs, observations, and survey questions in the form of Likert-type responses.

Some teachers of voice have traditionally shown a bias against scientific method in the use of the voice lesson. This aversion is enhanced with the presence of strange, untested technologies that find their way into the modern voice lesson. In 1951, McLean supported a more spiritual approach to the teaching of singing with this statement:

Even though some research exists in the use of technology and voice science, a need for a study that incorporates the technologies into voice lessons still exists. Voice pedagogues (i. e. Reid, 1984; Rubin, 1988) call for a way to incorporate voice science and technology into the training of voice without sacrificing basic technique. Titze, (1986) surmised that the value of all the ". . . charts, graphs, gadgets, and gismos in the studio" (p. 22) will not be solved until research is undertaken from the standpoint of someone trained in voice education rather than voice science. Cleveland (1994) reflects the growing acceptance with these comments:

Moore (1937/91) provides an early view of the scientific study of the voice in his history of laryngeal investigation. Development of the laryngoscope began in 1807 with Buzzoni, but the first "real success" (p. 267) was by the singing teacher Manuel Garcia, who used a dental mirror to view the larynx of his students. Von Leden (1990) provides a first-hand account of voice science in the middle part of the 20th century in his report on the evolution of the discipline. In 1994, Cleveland concluded that the preceding 25 years had been the most productive period for the study of the singing voice. Brewer (1989) constructed a descriptive matrix to reflect voice research that shows the interrelation of the unsolved problems, academic disciplines, and research tools pertinent to the profession.

Higgins (1991) notes the lack of reliable research in the area due to poor research design, lack of treatment time, lack of expertise of experimenters, poor quality of treatment, and lack of internal validity of experiments. Reasons for the poor research methods include the rapid change in technology, the delay of acceptance in the classroom, a traditionally narrow view of instruction, reluctance to extend the research by applying new technology to old problems, and the lack of qualified researchers. He suggests that future researchers follow the action research paradigm. Berz and Bowman (1995) point out the debate over the validity of feasibility and effectiveness in research studies that compare traditional teaching and computerized instruction. They suggest claims by these researchers could be due to a novelty effect or media advocacy as a bias for the investigators.

These considerations have been addressed in the design of the present study, which occurs in a naturalistic voice lesson setting.

Rudolph, Richmond, Mash, and Williams (1997) suggest specific strategies for adaptation of technology to the National Standards. Williams and Webster (1996) produced a compendium of applications of technology to music. Central to the philosophy behind this book is the "Systems Perspective" (Reese & Davis, 1998), in which people who use computers and the tasks they perform are considered more important than the software and hardware used.

Three studies exist concerning the auto-accompaniment software SmartMusic (formerly named Vivace) used in the present experiment. All of the studies centered on instrumental music. Ouren (1997) documented the effect of Vivace on the playing skills, musicality, and attitude of eight middle school students. Tseng (1996) investigated qualitatively the interaction of 10 college flute students with the Vivace system. Sheldon, Reese, & Grashel (unpublished) investigated differences in performance quality among three groups of instrumental music education undergraduates who received either no accompaniment, live accompaniment, or digital accompaniment. Wu (1997) explored the impact of Karaoke, a technology with some common characteristics.

Miller and Schutte (1990) discuss the role of feedback from spectral analysis as applied to the singing voice. Rossiter and Howard (1996) considered real-time visual feedback for voice development in prospective professional voice users in their development of a computer-based biofeedback device. Welch, Howard, and Rush (1989) used real-time feedback to develop a computer-based system of providing feedback for pitch detection. Ester (1994) developed a HyperCard stack called Hyper Vocal Anatomy to teach laryngeal anatomy to undergraduate music majors. Freeman, Syder, and Nicolson (1996) designed a multimedia tutorial for students of voice therapy.

The present study has grown out of two previous studies. In 1995, I completed a study that explored the various avenues for research in voice that were available on the Internet. In addition to Internet exploration, I used a series of interviews and on-line research to determine the attitudes of voice users toward Internet resources. In 1997, I completed a report of the extent to which attitudes of pre-service music teachers were affected by an Internet-based presentation of a voice relaxation process known as the "McClosky technique for vocal relaxation." Materials from the 1997 study have been incorporated into the present research. Some of the techniques used for the spectral analysis portion of the research have been adapted from Miller & Doing’s (1996) research.

Despite the research mentioned here, studies specifically addressing singing and voice production are not prominent enough to make broad generalizations or impact the teaching profession.

The experiment took place in two studios at the University of Illinous. One studio is equipped with an electronic keyboard and a computer that has the SmartMusic auto-accompaniment system installed. The other studio has a computer with sound analysis software and an Electroglottograph (EGG), a device to measure the opening and closing of the glottal folds.

A series of lessons were produced. Each of these lessons was supplemented with at least one of the technologies highlighted in this study. Students were able to access the Web documents through the Internet at any time. The students were also able to access the Smartmusic system outside of lesson time. Spectral analysis was not available outside of lessons.

Data were collected in three ways. First, each participant completed a weekly questionnaire regarding their reactions to the process. This questionnaire was submitted via email. Second, the participants completed four on-line Web forms that contained questions relating to the effectiveness of the various technologies. Most questions were Likert-type responses and provided quantitative data for statistical analysis. Third, I kept logs of my reactions to the process and my observations of the students.

Several sources of bias need to be considered. Members of the participant group all received instruction from the same teacher, who is the experimenter in this case. Care must be taken to observe whether changes in measurable phenomena are due to the presence of technology or by the effect of the instructor. Another source of bias considered was the "novelty effect" (the tendency for research subjects to react favorably to a treatment because the treatment is novel) and the "Hawthorne effect" (the tendency of subjects to react favorably because they know they are being observed).

Results of the study are reported in this paper in the following order. First, reactions of the participants and teachers to the individual technologies and their sub-components are reported. The three technologies highlighted for study are the Internet, Smartmusic accompaniment software, and spectral analysis software. Next, the effectiveness of each of the technologies are compared relative to one another. Finally, conclusions of a more general nature (e.g., the participants’ attitudes toward technology in general) are presented.

For the purposes of this project, the internet is defined as a combination of the World Wide Web and electronic mail. Web pages designed for informational purposes were used in lessons to illustrate points and used outside lessons as a reminder of the topic of a particular lesson. Examples of what was highlighted in lessons include proper posture, breathing techniques, and the McClosky technique for voice relaxation (see Repp, 1997). The Web and email were also used as the primary source of communications, including data collection.

The informational function of the internet achieved mixed results. I found Web pages awkward to use in lessons because of the loss of eye contact necessary for me to change pages during the lesson. Use of the pages also split student attention, as the student was required to pay attention to the computer screen and the teacher. The ability of the students to access the pages outside of lesson time was well-received, but the pages were not accessed extensively.

The use of the Web as a communication tool was effective. Problems included some students not checking their email regularly. Use of Web forms saved time in transferring data to database, but students did not always fill out forms before the lesson started. Computer glitches also forced some student to complete the forms more than once.

Spectral analysis took place two times during semester. This equipment was not available outside of lesson time. Use and interpretation of the spectral data required a relatively high degree of training and specialized equipment. The spectrographic analysis was divided into three parts. The first reading, taken through a microphone using a freeware software package called Spectrogram, produced an image with frequency on the vertical axis, time on the horizontal axis, and amplitude represented by color changes on the waveform (see Figure 1). This time-based spectrogram was used to demonstrate the spectral makeup of voice by showing changes over time, such as formant differences, as the singer changed vowel sounds.

Figure 1.

The second spectral reading, using Voce Vista software and hardware, uses a microphone to take a "snapshot" of the voice. In Figure 2, the light blue line shows a theoretical [a] (as in father) vowel, while the yellow line shows the actual sung vowel (Miller & Doing, 1996). The spectral readings were effective in demonstrating to the student the need to increase the spectral energy in the region around 3000 Hz known as the "singer’s formant."

Figure 2.

The third part of the spectral analysis process was acquired using the electroglottograph (EGG) included with the Voce Vista hardware. Two electrodes were placed on either side of the student’s larynx. The EGG measures the opening and closing of the vocal folds analogous to the changes in a slight current passing through the electrodes. Despite the long research tradition associated with the EGG, the process was difficult to administer and led to questionable results.

One of the most exciting uses of technology in the applied lesson format that has become feasible recently is the use of software as an accompanist. Since a piano accompaniment is standard in most vocal performances, teachers have been forced either to play the accompaniment for the student–a process that distracts the teacher from listening to the performance–or the student has been required to hire an accompanist, potentially leading to financial hardship.

I used three parts of the Smartmusic software package in lessons: the accompaniment feature, the tuner, and warm-up feature. The accompaniment feature was well received by students. The accompaniments work well in lessons and alleviate the need for a lesson accompanist. Unlike playing with a tape accompaniment, the intelligent software is able to react to the nuances of the singer to some extent. The ability to change keys on command is particularly effective. Students were able to learn the system and use it on their own without significant difficulty. Some problems occurred with the accompaniment in the performance situation. The software did not always register the student entrances, so it was necessary to start one performance a second time. The need to switch the software’s key disks led to delays within the performance.

I used the tuner function of the Smartmusic system with great effectiveness. The software features a tuner that displays the name of the pitch the student is singing along with a pitch wheel indicating whether the pitch is sharp or flat. The tuner was surprisingly well-received by the students and turned out to be one of the most effective technologies available both in and out of lessons.

The third part of the Smartmusic package utilized in the research was the warm-up feature. The software plays individual pitches or chords that ascend or descend in response to taps on a foot pedal. The software allowed the students without piano skills to practice warms and exercises without having to finger chords on the piano. As the teacher, however, I preferred to use the piano keyboard, so I could more easily monitor the pitch reached by the singer.

At the end of the semester I asked each student to fill out an on-line questionnaire that ranked the relative effectiveness of each of the technologies used in the process … both in the lesson and outside the lesson, where appropriate. Results below represent the average response on a seven point Likert-type scale with "1" being the most effective and "7" the least effective.

The above technology-related measures used both inside and outside of lessons were summed to determine whether students preferred the technologies used in the lesson or on their own. A paired samples t-test showed that students preferred using the technologies outside of class time by an average of .25 points on a seven-point scale (t=2.67, df=5, p=.045)

Students were also asked about their attitude toward educational technology and the use of educational technology for the use of teaching voice at the beginning and end of the semester. A paired samples t-test was performed on the data to determine if reports of attitude had changed over the semester. Attitude toward educational technology worsened by .5 on a 7-point scale (t=3.16, df=5, p=.025). Attitude toward the use of educational technology used for purposes of teaching voice was not statistically significant (t=1.46, df=5, p=.21) .

The fact that the participants’ attitude toward technology worsened over time was disappointing. I believe that the deterioration could have been due to several unrelated sources of bias. The attitudes toward educational technology were unrealistically high at the beginning of the semester (2.0 on a seven-point scale), possibly due to the "novelty effect." Also, because the final survey took place at the end of the semester, attitudes in general may not have been positive, given that it was late November in Central Illinois. Unfortunately, no control group data was taken for comparisons.

The fact that students found using technologies outside the lesson situation more effective than inside is worthy of note. Teachers should make access to technology available to students outside of class as much as possible.

Use of the SmartMusic software was found to be the most feasible of all the technologies surveyed. The software is relatively inexpensive, easy to use, and effective from the viewpoint of both the student and the teacher. Instructors using the software should make sure to utilize the tuner and warm-up features that were found to be particularly effective in the present study.

Use of the Internet was found to be feasible with some limitations. Use of Web pages in the lesson was cumbersome from the teacher’s point of view and relatively ineffective form the student’s point of view. Use of the Web outside of lessons was better-received and provided a way to communicate with the student outside of lesson time. Use of email and Web forms for data collection and communication were effective, but the use of these technologies is commonplace at the University where the study took place. Teachers at other institutions should judge their students' use of email before relying on it for the sole method of communication. Because the Web and email are so engrained in the lives of these particular students, the "novelty effect"–possibly a factor in the present experiment–did not affect this portion of the experiment as much as the newer technologies. Thus, the Web may not have been considered favorable due to bias.

Use of spectral analysis software was found to be unfeasible. From the instructor's perspective, the analysis calls for hardware not readily available and expertise beyond what can be reasonable expected of a typical voice teacher. Although the students found the initial experience with the spectrometer to be extremely positive (probably due to "novelty effect"), the attitude toward the technology decreased as the novelty wore off and, by the end of the semester, the technology did not rank high in student responses. From the experimenter point of view, the technology lacked reliability, because I could record a great variety of responses from a single individual, and lacked validity, because the readings of the analysis did not always reflect the changes I noted from recordings of the student.

In conclusion, the saturation of technology into the voice lesson was feasible and extremely effective. Even the relatively ineffective technologies received high responses from the participants. From a teacher’s point of view, this group of students was the best I have ever taught … even better than students who pay for lessons. As proof of the high level of interest and participation, not one student missed a single lesson all semester.

This study was the first portion of a two-semester project. During the Spring of 1999 I will repeat the process, with the exception that participants will receive differing levels of technology to act as internal control groups. I will also work to limit the sources of bias that occurred in Phase I.