INTRODUCING THE INTERPRETATION SWITCHER INTERFACE

TO MUSIC EDUCATION

Verena Konz and Meinard M¨uller

Saarland University and MPI Informatik, Campus E1 4, 66123 Saarbr¨ucken, Germany

Keywords:

Music information retrieval, Music education, Music synchronization, Alignment, User interface.

Abstract:

In the ﬁeld of music information retrieval (MIR), great efforts have been directed towards the development

of technologies and interfaces that allow users to access and explore music on an unprecedented scale. On

the other hand, musicians and music teachers are often still skeptical about the beneﬁts of computer-based

methods in music education. In this paper, we report on an experiment conducted at the University of Music

Saarbr¨ucken with the goal to introduce a novel MIR user interface, referred to as Interpretation Switcher, to

music education and to get feedback from music experts. To this end, we asked nine music students to analyze

different performances of the same piece of music according to a well designed questionnaire, using the novel

switching functionality of our interface. Doing so, we not only tested and evaluated our interface in a setting

of practical relevance, but also indicated the potential of MIR methods in music education.

1 INTRODUCTION

Computers have become an indispensable tool for

storing, processing, and generating music. Even

though computer-based methods and interfaces are

ubiquitously used for music synthesis, there is still

a reluctance in using computers for music analysis

and music education. Research in computer-assisted

music education already started at the end of the

1960s mainly in the USA and the United Kingdom

(Brown, 2007; Kiraly, 2003; Smith, 2009; Stevens,

1991; Guetl and Parncutt, 2008). For example,

computers have served as a tool for creative music-

making. Furthermore, the method of Computer-

Assisted-Instruction (CAI), where students are taught

a particular skill by a computer, has been applied

in areas like music theory or aural training. Vari-

ous studies have been conducted to investigate the

effect of CAI-based methods within music educa-

tion (Smith, 2009; Kiraly, 2003), and the usefulness

of such methods seems to be a controversial issue.

In the ﬁeld of music information retrieval (MIR),

the development of technologies and interfaces for

music exploration and analysis has been an active

research area (Damm et al., 2008; Dixon and Wid-

mer, 2005; Goto, 2003; M¨uller, 2007; Sonic Visu-

aliser, 2009). However, these technologies and inter-

faces are often evaluated in the own lab environment,

where people are familiar with computers. Though,

for building up MIR systems of practical relevance

one needs broader feedback, in particular from mu-

sic experts. Hence, for a user-centered analysis, it is

necessary to conduct “real” user studies that ensure a

natural setting (Lesaffre et al., 2008).

There are many applications in the context of mu-

sic education that may beneﬁt from the above men-

tioned MIR-based technologies and interfaces. How-

ever, musicians and music teachers are often still re-

luctant in using novel computer-assisted methods and

novel MIR interfaces in their lessons. Furthermore,

many of the available MIR interfaces are still too

complicated lacking the necessary user-friendliness

and robustness to be operable by non-experts. Under

such circumstances, it remains a challenge to raise the

interest of music educators for using, testing, and par-

ticipating in the development of novel MIR interfaces

and for discussing possible application scenarios.

In this paper, we report on an investigation with

the objective of introducing a novel MIR interface to

music education. In collaboration with the Univer-

sity of Music Saarbr¨ucken we conducted an experi-

ment consisting of several steps. First, nine piano stu-

dents were recorded playing the same piece of music,

the ﬁrst movement of Beethoven’s Path´etique Sonata

Op. 13, on the same piano and under the same record-

ing conditions. In the next step, the nine audio record-

ings were temporally aligned and integrated in a user

interface referred to as Interpretation Switcher (Fre-

135

Konz V. and Müller M. (2010).

INTRODUCING THE INTERPRETATION SWITCHER INTERFACE TO MUSIC EDUCATION.

In Proceedings of the 2nd International Conference on Computer Supported Education, pages 135-140

DOI: 10.5220/0002765201350140

 SciTePress

merey et al., 2007; M¨uller, 2007), which allows for

synchronous playback of the different performances.

Upon using this interface, the music students were

then asked to analyze the anonymised performances

according to a well-designed questionnaire.

There are a number of achievements of our ex-

periment. Firstly, we tested and evaluated our inter-

face in a setting of practical relevance, thus indicat-

ing the potential of MIR methods in music education.

Secondly, we generated royalty free music record-

ings without any copyright restrictions, which can be

used freely for research purposes. Thirdly, using a

Yamaha Disklavier for our experiments, we also ob-

tained MIDI data (which was actually not used in the

investigation described in this paper) along with au-

dio recordings. Such MIDI-audio pairs can be used as

ground truth material for various MIR tasks (M¨uller,

2007). Finally, we generated many different inter-

pretations of the same piece, which yields valuable

data for tasks such as automated performance analy-

sis (Widmer et al., 2003).

The paper is organized as follows. In Sect. 2, we

present in detail the setup of the conducted experi-

ment describing the piece of music, the recording con-

ditions, the interface, and the questionnaire. Then, in

Sect. 3, we discuss the results of our experiment re-

sponding to the performance evaluation and to the in-

terface evaluation. Finally, in Sect. 4, we sketch fur-

ther application scenarios and indicate future work.

2 EXPERIMENTAL SETUP

2.1 Piece of Music

For our experiment, we chose the ﬁrst movement of

Beethoven’s Path´etique Sonata Op. 13. This piece of

music appeared to be a good choice for various rea-

sons. Firstly, the Path´etique Sonata is a musicologi-

cally outstanding work, for which numerous detailed

descriptions and scientiﬁc literature exists. Secondly,

being a very popular and famous work, the Path´etique

belongs to the standard repertoire of many pianists.

Hence, there are numerous audio recordings for this

piece. The third and most important reason for choos-

ing the Path´etique is that it is very rich in contrast

concerning tempo as well as dynamics.

To make the latter point clear, we describe the

Path´etique’s exposition in more detail. Beginning

with the slow introductory theme marked Grave

(measures (abbreviated mm.) 1-10, see Fig. 1 (a)),

the work starts very dramatically. The introduc-

tion is characterized by its contrasts in dynamics—

fortissimo passages are followed by subito piano and

(a)

(b)

(c)

(d)

Figure 1: First movement of Beethoven’s Path´etique Sonata

Op. 13 (score obtained from (Mutopia Project, 2009)).

(a) Beginning of the introduction (Section A, mm. 1 ff.)

(b) First theme (mm. 11 ff.) (c) Second theme (mm. 51

ff.) (d) Section B (mm. 89 ff.)

vice versa. This contrast in dynamics is underlined by

contrasts in rhythmics, articulation, and atmosphere.

Ending with the chromatic run, the introduction leads

in the ﬁrst theme (mm. 11-27, see Fig. 1 (b)) of the

sonata which is characterized by the tremolo in oc-

tavos in the left hand giving it a dramatic touch. In

contrast to the dramatic ﬁrst theme, the second theme

(mm. 51-88, see Fig. 1 (c)) sounds more playful. It is

based on the call and response principle and is char-

acterized by a play with articulation.

As described above, one can ﬁnd many contrast-

ing elements in the exposition of the ﬁrst movement

of the Path´etique: Contrasts in the shaping of the two

themes, contrasts in dynamics, and contrasts in at-

mospheres. In addition, there is an abrupt change in

tempo at the beginning of the ﬁrst theme (mm. 11),

where the introductory Grave leads in the actual expo-

sition marked Allegro di molto e con brio. Because of

its musical richness, the Path´etique offers the pianist a

wide range of possibilities for shaping the piece with

respect to dynamics, tempo, and agogics. Therefore,

this piece is very well suited in view of the musical

evaluation within our experiment.

2.2 Performance and Recording Setup

The recordings of our experiment were performed at

the University of Music Saarbr¨ucken. Nine students

of different study paths from the piano class of Prof.

CSEDU 2010 - 2nd International Conference on Computer Supported Education

136

Figure 2: Instance of the Interpretation Switcher plugin of

the SyncPlayer for synchronous playback of different audio

recordings of the same piece of music. In this example,

nine different recordings of the exposition of Beethoven’s

Path´etique Sonata are opened.

Thomas Duis were asked to play the ﬁrst movement

of the Path´etique. Being in several training states,

they were on different performance levels. All stu-

dents played on the same instrument under the same

recording conditions on two different days (Friday,

06.02.2009 and Monday, 09.02.2009). In the record-

ing sessions, only the performer, the technical staff,

and the scientiﬁc investigators were present in the

room—the other performers were not allowed to lis-

ten to their fellow students. Using two microphones,

we did not achieve the quality of a recording studio.

However, we obtained audio recordings of a sufﬁcient

quality in view of our experiments. Furthermore, us-

ing a Yamaha Disklavier, we also generated MIDI

data along with the audio recordings. Actually, the

MIDI ﬁles were not used in our experiments, but as

mentioned before they are useful for later projects.

2.3 MIR User Interface

The SyncPlayer system is an advanced audio player

for multimodal presentation, browsing, and retrieval

of music data (Fremerey et al., 2007). One of the

available plugins for the SyncPlayer, referred to as

Interpretation Switcher, is the MIR interface used in

our experiment, see Fig. 2. It allows the user to select

several recordings of the same piece, which have pre-

viously been synchronized (M¨uller, 2007). Each of

the selected recordings is represented by a slider bar

indicating the current playback position with respect

to the recording’s particular time scale. The audio

recording that is currently used for playback, in the

following referred to as reference recording, is repre-

sented by a red marker. The slider of the reference

recording moves at constant speed while the sliders

of the other recordings move according to the relative

tempo variations with respect to the reference. The

reference recording may be changed at any time sim-

ply by clicking on the respective marker located on

the left of each slider. The playback of the new ref-

erence recording then starts at the time position that

musically corresponds to the last playback position of

the former reference. One can also jump to any po-

sition within any of the recordings by directly select-

ing a position of the respective slider, which automat-

ically triggers a switch of the reference to the respec-

tive recording. A similar functionality is provided by

the Sonic Visualiser, a system for viewing and analyz-

ing the contents of music audio ﬁles (Sonic Visualiser,

2009). Here, the MATCH plugin allows for tempo-

rally aligning two recordings and then for switching

from one to the other (Dixon and Widmer, 2005).

Note that the SyncPlayer and the Sonic Visualiser

provide many more functionalities comprising plug-

ins for inter- and intra-document browsing and re-

trieval as well as data visualization and analysis. In

our experiment, we restrict ourself basically to the

switching functionality with the motivation to keep

the interface as simple and intuitive as possible to

avoid any rejections from the users. As Fig. 2 shows,

the Interpretation Switcher looks like a standard au-

dio player with the only difference that more than one

slider control bar is available. After a short explana-

tion of the main switching functionality, none of the

students reported on difﬁculties in using our interface.

2.4 Survey and Questionnaire

Subsequent to the last recording session, the nine dif-

ferent performances were aligned and integrated in

our Interpretation Switcher. Then, we conducted our

survey in the evening of the second recording day

(Monday, 09.02.2009). Eight music students partic-

ipated in the survey, seven of whom were also among

the nine performers. The different interpretations

were anonymised within the interface and the partici-

pants listened to the recordings for the ﬁrst time.

Each participant was provided with a computer

running the Interpretation Switcher and with ear-

phones. After a short introduction of the interface’s

switching functionality, the participants received a

questionnaire having one hour for answering the

questions. This questionnaire consisted of two main

parts. In the ﬁrst part, the students had to listen, to

compare, and to rate the nine different interpretations

with respect to various performance aspects. Here,

the questions were designed in such a way that the

INTRODUCING THE INTERPRETATION SWITCHER INTERFACE TO MUSIC EDUCATION

137

students naturally started to use the switching func-

tionality of the interface, thus getting familiar with

the Interpretation Switcher in a concrete application

of musical relevance. In the second part, they were

then asked to give feedback on the usefulness and op-

erability of the interface itself.

The questions of the ﬁrst part of the question-

naire referred to different sections of the ﬁrst move-

ment of the Path´etique. As a kind of warming up, we

started with a short section (Section A), which only

consisted of the ﬁrst three measures, see Fig. 1 (a).

This section was cut out from the nine aligned per-

formances and presented to the students by the Inter-

pretation Switcher interface. Even being rather short,

Section A already offers the pianists a wide range of

interpretation so that the comparison of the different

performances constitutes a musically interesting task.

In the ﬁrst question (A1), the participants had to rate

the nine different interpretations of Section A with re-

spect to the three musical aspects dynamics, articu-

lation, and agogics. Here, the rating scale ranged be-

tween 1 and 10, where 1 means poor and 10 excellent.

In addition, they had to rate their total impression of

this section’s performances using the same scale. Af-

terwards, in question A2, they had to identify their

own interpretation (if applicable) only by means of

Section A. Then, the performances of Section A were

closed and a different section (Section B) was pre-

sented by the Interpretation Switcher to them. Here,

Section B consisted of the technically more involved

mm. 89-100, see Fig. 1 (d). The students then had to

answer corresponding questions (B1, B2).

At the beginning of the questionnaire, the students

were confronted with different performances of rela-

tively short sections. Here, only switching between

the performances was required to properly answer the

questions—jumping back and forth within a perfor-

mance was not necessary. In this way, the students

became familiar with the basic switching functional-

ity of the interface. In the next stage, they were pre-

sented with the nine performances of the entire ex-

position. They now had to rate their total impres-

sion of the ﬁrst theme (mm. 11 ff., see Fig. 1 (b)),

of the second theme (mm. 51 ff., see Fig. 1 (c)), and

of the entire exposition (questions E1, E2 and E5).

Here the new challenge concerning the use of the In-

terpretation Switcher was not only to switch between

the different performances but also to ﬁnd the corre-

sponding entry points of the two themes within the

recordings. Another task (E3), was to order the nine

different interpretations with respect to the tempo (be-

ginning with the slowest, ending with the fastest) in

the second theme. With this task the students had to

constantly switch between and jump within the per-

formances, being forced to use the functionality of the

interface extensively. In question E4, again, they had

to identify their own performance (if applicable) now

having the entire exposition at their disposal.

After ﬁnishing the questions on music aspects, in

the second part of the questionnaire the participants

were asked to evaluate the Interpretation Switcher in-

terface. Here, the idea was to let the participants ﬁrst

use the interface in an application scenario to gather

practical experience without knowing about the ﬁnal

interface evaluation. In the ﬁrst question (S1), they

should rate the user-friendliness and the degree of us-

ability of the Interpretation Switcher on the above de-

scribed scale from 1 to 10. We then wanted to know

if there were any problems while using the interface

(S2). Furthermore, the students were asked to com-

ment on possible improvements and to propose ad-

ditional functionalities they would have liked when

working on the ﬁrst part of the questionnaire (S3). In

a last question (S4), they should sketch possible appli-

cation scenarios where they could imagine to use MIR

user interfaces such as the Interpretation Switcher.

3 EVALUATION

3.1 Performance Evaluation

In the ﬁrst part of the questionnaire, the partici-

pants had to analyze and compare the different per-

formances against each other. Table 1 presents the

results of question A1, where they had to rate the

nine different performances of Section A with regard

to dynamics, articulation, agogics, and in total. The

ﬁrst row of Table 1 shows the number of the respec-

tive performance; the values of each column corre-

spond to the respective performance. The second row

shows the ratings with regard to dynamics averaged

over the eight participants. For example, the ﬁrst per-

formance was rated with a score of µ = 6.63 on av-

erage. The third row shows the standard deviation,

which is σ = 1.19 for the ﬁrst performance. The fol-

lowing rows of Table 1 are to be read in the same fash-

ion. For example, the participants rated the sixth per-

formance on average with µ = 6.88 (σ = 1.36) with

respect to articulation, whereas the overall impression

of this performance amounts to µ = 6.38 (σ=1.51).

As we can see, the eighth performance was ranked

highest with respect to dynamics (µ = 6.75), whereas

the second one with respect to articulation (µ = 7.00).

The overall rankings for Section A are relatively close

together, which may show that the section was too

short for giving a well-founded evaluation or that it

was played similarly by all students.

CSEDU 2010 - 2nd International Conference on Computer Supported Education

138

Table 1: Evaluation results of question A1 (Fig. 1 (a)). The

average ratings µ along with the standard deviations σ are

shown for the nine performances of Section A with regard

to various musical aspects.

1 2 3 4 5 6 7 8 9

Dynamics µ 6.63 6.25 6.38 6.13 5.75 6.38 5.38 6.75 6.63

σ 1.19 1.49 1.69 1.36 1.67 1.69 1.77 1.49 1.60

Articulation µ 6.13 7.00 6.25 6.38 6.13 6.88 5.38 5.88 6.00

σ 1.81 1.69 1.83 1.30 1.64 1.36 2.20 1.36 1.41

Agogics µ 6.13 6.50 5.13 6.38 5.63 6.50 4.75 5.75 6.13

σ 1.73 1.69 2.53 1.92 1.41 1.41 2.12 1.28 0.99

Overall µ 6.25 6.25 6.00 6.25 5.88 6.38 5.25 6.13 6.00

σ 1.58 1.67 2.14 1.28 1.25 1.51 1.58 1.55 1.20

Table 2: Evaluation results of question B1 (Fig. 1 (d)).

1 2 3 4 5 6 7 8 9

Dynamics µ 5.38 6.38 7.13 6.38 5.63 7.13 5.88 4.75 4.13

σ 1.41 1.69 1.55 1.69 1.41 1.13 1.36 2.12 1.36

Articulation µ 5.38 6.00 6.25 6.38 4.88 6.25 5.13 4.63 4.75

σ 1.85 1.20 1.04 1.60 1.89 1.58 1.89 1.69 2.19

Agogics µ 5.13 6.00 6.88 6.75 5.13 6.75 5.38 4.88 4.00

σ 2.42 0.76 0.99 1.39 1.81 1.28 1.51 1.64 1.51

Overall µ 5.25 5.88 6.75 6.63 5.00 6.63 5.63 5.00 4.25

σ 1.75 0.99 1.49 1.41 1.51 1.41 1.69 1.51 1.39

Analogously, Table 2 shows the results of ques-

tion B1. Here, the best performance concerning the

overall impression is the third one (µ = 6.75), whereas

the worst performanceis the ninth one (µ = 4.25). Ac-

tually, the ninth performance was ranked worst with

respect to all musical aspects. The reason for the

poor rating is that the performing student struggled

signiﬁcantly with the technically more involved Sec-

tion B, thus neglecting the musical shaping. This

may also explain, why the given scores between the

performances differ to a much larger degree for Sec-

tion B than for Section A. Finally, Table 3 presents

the results of questions E1, E2 and E5, where only

the overall impression had to be rated. Here, the ﬁrst,

third, and second performances were rated best with

regard to the ﬁrst theme (µ = 6.88), the second theme

(µ = 6.75), and the entire exposition (µ = 7.00), re-

spectively. Again, the ninth performance was rated

worst with regard to all three categories. Interestingly,

there does not exist a clear winner performance con-

cerning all different musical aspects and themes.

3.2 Interface Evaluation

In the second part of the questionnaire, the students

were asked about the operability and usefulness of the

Interpretation Switcher. As mentioned before, none

of them had serious problems in using the interface,

which is also reﬂected by a high average rating of

µ = 7.63 given for the user-friendliness of the inter-

Table 3: Evaluation results of question E1 (ﬁrst theme,

Fig. 1 (b)), question E2 (second theme, Fig. 1 (c)), and ques-

tion E5 (entire exposition).

1 2 3 4 5 6 7 8 9

1. Theme µ 6.88 6.25 5.88 5.63 4.88 5.25 5.50 5.25 4.75

σ 1.13 1.39 1.25 2.13 2.30 2.19 2.45 2.12 1.83

2. Theme µ 6.25 6.38 6.75 5.63 5.00 5.25 6.38 5.13 4.50

σ 1.83 1.06 1.39 1.30 1.60 1.04 1.85 1.96 1.60

Exposition µ 6.50 7.00 6.88 5.63 4.75 5.13 5.50 5.63 4.38

σ 1.41 1.51 1.46 1.85 1.58 1.81 2.00 1.77 1.60

face. Only one of the participants gave a low score

of 4 explaining a relatively large standard deviation

of σ = 2.07. As it turned out, the reason for this was

that the student was pressured for time and not really

in the mood of participating in our experiment. Ac-

tually, this student also admitted that she has had no

time for properly practicing the piece, resulting in per-

formance number nine with the lowest score, see Ta-

ble 3. Most of the other participants emphasized that

they found the handling and functioning of the Inter-

pretation Switcher very intuitive, even music students

who have had only little experience with computers.

Furthermore, most students found the Interpretation

Switcher very useful for tasks such as performance

analysis, music comparison, and other analysis tasks.

Here, the average rating amounted to µ = 7.13 with

standard deviation σ = 1.89.

After the general rating, the students were also

asked to freely comment on problems, possible im-

provements, additional functionalities, and possible

application scenarios (S2, S3, S4). At this point they

all conﬁrmed that they have had no problems while

using the Interpretation Switcher interface. However,

two students noted that the interface could have re-

acted faster while switching between the respective

performances. One student would have appreciated

to have an additional functionality for displaying the

musical score during playback. Also user-deﬁned

auxiliary markers that can be freely ﬁxed, adjusted,

and removed along the various slider control bars

should be introduced for additional orientation and

navigation purposes. All but one of them afﬁrmed that

they could imagine to use the Interpretation Switcher

within their studies or even for private use. In partic-

ular, they said that the interface may be useful in the

context of special seminars, where the comparison of

different performances play an important role. One

student was enthusiastic about the features offered by

the Interpretation Switcher. He usually records his pi-

ano lessons in order to listen to and to study his own

playing afterwards. Here, he would signiﬁcantly ben-

eﬁt from novel switching and navigation functionali-

ties for comparing and analyzing the recorded audio

material. Also, the Interpretation Switcher could be

INTRODUCING THE INTERPRETATION SWITCHER INTERFACE TO MUSIC EDUCATION

139

very useful for compactly documenting the learning

progress over a longer period in time. For exam-

ple, it could synchronously present the various per-

formances of a speciﬁc musical section recorded in

different piano lessons over the semester.

4 CONCLUSIONS

In this paper, we presented a ﬁrst experiment con-

ducted at the University of Music Saarbr¨ucken with

the main objective of introducing MIR user inter-

faces with novel switching and navigation function-

alities to music teachers and students. Even though

this group tends to be skeptical about using computer-

based methods in music education, most participants

afﬁrmed the usefulness of our interface for compar-

ing and analyzing performances or simply for music

listening and enjoyment. Testing and evaluating our

interface within a concrete application of practical rel-

evance, we not only made a new group of prospective

users acquainted with MIR methods but also obtained

valuable feedback from music experts.

The presented experiment only constitutes the be-

ginning of a planned collaboration with music educa-

tors and students, who are usually not aware of the

developments in music information retrieval. For the

future, we plan to conduct similar experiments on a

larger scale. One further idea is to participate regu-

larly in the lessons of piano students to record their

playing. We then plan to process (segment, classify,

synchronize) the audio material automatically and to

suitably integrate it in our Interpretation Switcher to

document and analyze the students’ learning process.

Finally, we plan to develop and combine various

additional functionalities. For example, as mentioned

by one of the participants, an additional sheet mu-

sic interface for presenting the musical score while

playing back associated audio material would be help-

ful. Actually, such functionalities have been pre-

sented in (Damm et al., 2008). Furthermore, we will

integrate additional functionalities for inter- and intra-

document music browsing including the possibility of

setting user-deﬁned auxiliary markers as well as pre-

computed markers that reﬂect the musical form of the

piece (Fremerey et al., 2007; Goto, 2003). In intro-

ducing novel functionalities, one main challenge will

be to keep the operability of the interface as intuitive

as possible to avoid rejections from the users’ side.

ACKNOWLEDGEMENTS

The research was funded by the Cluster of Excellence

on Multimodal Computing and Interaction at Saar-

land University. We thank Wolfgang Bogler and the

piano class of Prof. Thomas Duis at the University of

Music Saarbr¨ucken for their support.

REFERENCES

Brown, A. (2007). Computers in Music Education. Rout-

ledge.

Damm, D., Fremerey, C., Kurth, F., M¨uller, M., and

Clausen, M. (2008). Multimodal presentation and

browsing of music. In Proc. ICMI, Chania, Crete,

Greece, pages 205–208.

Dixon, S. and Widmer, G. (2005). Match: A music align-

ment tool chest. In Proc. ISMIR, London, United

Kingdom, pages 492–497.

Fremerey, C., Kurth, F., M¨uller, M., and Clausen, M.

(2007). A demonstration of the SyncPlayer system.

In Proc. ISMIR, Vienna, Austria, pages 131–132.

Goto, M. (2003). SmartMusicKIOSK: Music listening sta-

tion with chorus-search function. In Proc. ACM Sym-

posium on User Interface Software and Technology,

pages 31–40.

Guetl, C. and Parncutt, R. (2008). An interactive tool for

training and testing musical auditory skills. In Proc.

ED-MEDIA, Vienna, Austria, pages 5229–5237.

Kiraly, Z. (2003). Solfeggio 1: a vertical ear training in-

struction assisted by the computer. International Jour-

nal of Music Education, 40(1):41–58.

Lesaffre, M., Leman, M., De Baets, B., De Meyer, H., De

Voogdt, L., and Martens, J.-P. (2008). How potential

users of music search and retrieval systems describe

the semantic quality of music. Journal of the Ameri-

can Society For Information Science and Technology,

59(5):1–13.

M¨uller, M. (2007). Information Retrieval for Music and

Motion. Springer.

Mutopia Project. Music free to download, print out, perform

and distribute.

http://www.mutopiaproject.org

Retrieved 12.05.2009.

Smith, K. H. (2009). The effect of computer-assisted in-

struction and ﬁeld independence on the development

of rhythm sight-reading skills of middle school instru-

mental students. International Journal of Music Edu-

cation, 27(1):59–68.

Sonic Visualiser (2009).

http://www.sonicvisualiser.

org/

, Retrieved 12.05.2009.

Stevens, R. S. (1991). The best of both worlds: An eclectic

approach to the use of computer technology in music

education. International Journal of Music Education,

17(1):24–36.

Widmer, G., Dixon, S., Goebl, W., Pampalk, E., and Tobu-

dic, A. (2003). In search of the Horowitz factor. AI

Magazine, 24(3):111–130.

CSEDU 2010 - 2nd International Conference on Computer Supported Education

140