Constructing a support system for self-learning playing the piano at the beginning stage

Alma Mater Studiorum University of Bologna, August 22-26 2006 Constructing a support system for self-learning playing the piano at the beginning stage Tamaki Kitamura Dept. of Media Informatics, Ryukoku Univ., Otsu, Shiga, Japan tamaki_0916@hotmail.com Masanobu Miura Dept. of Media Informatics, Ryukoku Univ., Otsu, Shiga, Japan miura@rins.ryukoku.ac.jp ABSTRACT This paper aims at realizing a support system for selflearning playing the piano at the beginning stage. The system is conceptually designed as a computer-based interactive system, and it is expected that using the system nonpianists of wide age ranges are able to learn playing the piano without experts' instructions. Designed here is an original procedure, strongly dependent to a famous textbook "Methode Rose", and the ways of producing exercises described in "Bayer" are interpreted manually to be included. The proposed system will be implemented by employing this designed procedure. Developed here is a subsystem, which automatically generates basic exercises of learning observing records of his/her weak points. Other facilities of it are: a) analyzing his/her performance recorded using a MIDI sequencer, b) extracting inaccuracies and/or deviations concerning timing, MIDI-velocity and duration, and c) showing obtained defects of users to a PCdisplay together with optimum exercises corresponding to their errors. Keywords Piano, MIDI, Self-learning, Methode Rose, In: M. Baroni, A. R. Addessi, R. Caterina, M. Costa (2006) Proceedings of the 9th International Conference on Music Perception & Cognition (ICMPC9), Bologna/Italy, August 22-26 2006. 2006 The Society for Music Perception & Cognition (SMPC) and European Society for the Cognitive Sciences of Music (ESCOM). Copyright of the content of an individual paper is held by the primary (first-named) author of that paper. All rights reserved. No paper from this proceedings may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information retrieval systems, without permission in writing from the paper's primary author. No other part of this proceedings may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information retrieval system, without permission in writing from SMPC and ESCOM. INTRODUCTION People who have desired to obtain skill of playing the piano have been increasing in Japan. To acquire the ability of playing the piano is kind of a dream for non-pianists who want to play it of wide age ranges. They often, however, give up practicing playing the piano even if they have special interests to be able to play. Possible reasons of this are a) they cannot choose appropriate exercises of practicing playing the piano by themselves, b) it is difficult for them to be instructed directly from experts because of various kinds of restricted resources, c) there is no appropriate circumstances of self-learning playing the piano, and so forth. LEARNING PLAYING THE PIANO IN GENERAL On learning playing the piano under experts there is somewhat a common procedure among experts. Especially, the introduction procedure of it is almost common among experts. On this stage, students may learn basic issues, such as basic movements of fingers, making sure the position of keys on piano, corresponding to musical scores. There are two famous textbooks often used on conventional education of playing the piano, one of them is called as "Methode Rose", written by Ernest Van de Velde, and, the other one is called as "Beyer", written by Ferdinand Beyer. Most of experts employ these textbooks as a manual on teaching students who are at the beginning stage. After this stage, students are usually scheduled to obtain advanced skills of finger movements and performances such as those concerning musical expression with the use of etudes such as "Burgmüller" and "Czerny" under experts' instructions. As you know, on learning playing the piano, there are various kinds of stages, having various levels of difficulties to master, according to the skill of students. In this study, the first stage of learning playing the piano is much important than other stages. It is said that the first stage is relatively ISBN 88-7395-155-4 2006 ICMPC 258

important for students, because the experiences of this stage will effects to students of other stages and to make them have strong interests is expected at this first stage. In other words, if students are failed to learn playing the piano at the first stage, he/she will never try to play it. It is often happened that students have difficulties in practicing piano at the first stage on self-learning it. AIMS This paper aims at realizing a support system for selflearning playing the piano at the beginning stage. The system is conceptually designed as a computer-based interactive system, and it is expected that using the system nonpianists of wide age ranges are able to learn playing the piano without experts' instructions. OUTLINE OF THE PROPOSED SYSTEM Proposed here is an original procedure, in which conventional procedures depicted in the textbook Methode Rose and Bayer effect strongly. Outline of the proposed procedure is listed in Appendix A. Our proposed system employs this procedure, and a subsystem of proposed system is constructed here, which can generate appropriate exercises to beginners automatically. An interactive facility which detect lack of students and point out it if any is tried to implement. IMPLEMENTATION Desirable tasks for beginners Here explains the subsystem of generating exercises automatically for students. Exercises concerning finger movements among 2, 3, 4, or 5 degree, which correspond to the basic step of proposed procedure from 1-2 to 1-5 respectively, are automatically generated. Let us consider generating automatically exercises of two degree. It is desirable that students should concentrate on training finger movements for only 2 degree and no other characteristics should be included. To realize these aspects, each module of automatic generating should generate various kinds of tasks having just single characteristics, not having several characteristics. Other restrictions of this stage is: a) more than six degrees are out of focus, because of avoiding finger crossing on playing with a single hand, b) just a single key (C-major at here),is employed, because of avoiding stroking black keys on piano, c) tempo of relatively slow, such as from 60bpm to 80bpm, d) length of generated task is limited no longer than 5 bars. musical scores should be presented to beginners so as not to be boring training for them. Several musical scores within one bar are prepared as to construct exercises for beginners. Each musical score has good corresponding to a training feature. These short musical score are named as partial task, it will organize exercises named total task, that means complete exercise. Nature of total tasks and partial tasks Here explains the way to produce partial tasks. It is assumed here that all exercises have binary beats, such as 4/4, 2/4, the length of all notes involved in partial tasks should be same as quarter note, and total task is constructed as a concatenation of partial musical scores. Restrictions for note allocations of partial tasks concerning practicing three degrees are as follows: The degree between highest and lowest note on a partial task should be within three degree, Same note repetition in a partial task should be avoided, The last note in each partial task should be same or having three degree compared to the first note in the partial task, Three degree should be appeared in conjunctive three notes in a partial task. Because five patters are possible to each quarter note in a partial task (C, D, E, F, and G, when the key is C), so 5x5x5x5 = 625 patterns are logically available in case there is no constrains in a partial task. There are, however, some kinds of constrains to each task are set and the number of available partial task are to be limited. Table 1 shows the available partial tasks corresponding to practicing patterns of each degree. Examples of partial tasks for exercises of practicing three degrees are shown in Fig. 1. Table 1. List of number of available partial tasks corresponding to each exercise degree # of partial tasks 2 9 3 19 4 13 5 5 Exercises automatically generated should have strong correspondences to single feature. Exercises having several features of training are thought to be inappropriate for beginners, because if exercises have several loads it is difficult to analyze which loads contained to the exercise is fatal for beginners. It is also desirable that various kinds of ISBN 88-7395-155-4 2006 ICMPC 259

Figure 2. All available partial tasks for practicing five degree. Figure 1. Examples of available partial tasks for practicing three degree. Algorithm for generating total tasks Here explains about generating total tasks for practicing five degree. Constrains for allocating notes on partial tasks for five degree are somewhat complex compared to those for two, three and four degrees, because in a partial task you are allowed to allocate just four notes, so to realize five degree among notes you should observe the note allocated on the neighborhood partial tasks. Constrains of allocating four notes in each partial task concerning practicing five degrees are shown as follows: Same note repetition in a partial task should be avoided, The degree between highest and lowest note in a partial task should be four or five degree, If the degree between highest and lowest note in a partial task is four, the fist note in the following partial task should have five degree from the lowest note. If the degree between highest and lowest note in a partial task is five, conjunctive two notes should have five degree among them. Example of outputs Automatic generators of total tasks for practicing two, three four or five degrees are constructed as a subsystem of proposed self-learning system. At first, user can select one of the numbers of degree and by clicking a button he/she will obtain a total task as a SMF (Standard MIDI File) format. Figure 5 shows example of outputs. Though all generated tasks have 5 bars, by changing the parameters concerning length of exercises you will get more than or shorter than five bars. This automatic generator accepts user s profile which will be calculated based on the results of errors for each partial task. User s profile will be taken into consideration to the possible state transitions. All exercises generated automatically, shown in Fig.5, are generated with blind respect to user s profile. Figure 2 shows all possible partial tasks for practicing five degrees, and Figure 3 shows outline of automatic generation for total task of five degrees as state transition model. Figure 4 shows algorithms corresponding to Fig.3. S.:Start Figure 3. Automaton of automatic generating for total tasks of practicing five degree. ISBN 88-7395-155-4 2006 ICMPC 260

S A, S B, S C,S D,S E are corresponding to those in Fig.2. Figure 4. Algorithm of automatic generation of total tasks for practicing five degree. (a) An example of total task for practicing two degree. (b) An example of total task for practicing three degree. (c) An example of total task for practicing four degree. (d) An example of total task for practicing five degree. the piano. Appropriate tasks concerning his/her weak points should be given to beginners combined with information for lacks of them. As a previous study, a computer system analyses the performance of playing HANON recorded by MIDI sequencer and based on the analysis the system produce appropriate tasks for students, after that it is obviously confirmed that students can override the weak points by practicing exercises automatically generated by the system compared to other students practicing exercises written in HANON. The system is named MANON [2]. MANON analyses the performance result and finds the weak points concerning corresponding finger. But in this study, it is assumed that weak points should be associated with characteristics of note sequences, such as partial tasks concerning two, three, four, or five degrees. Proposed method for presenting weak points to students is that musical scores that students are not good at playing should be given to students as an objective to override. In this study a relative weights representing how they are not good at playing are given to each partial task, and on producing total tasks the weights are taken into consideration. The relative weights are automatically calculated by practicing total tasks as a background module, and from a point relative weights are employed to generate total tasks under certain strategies of automatic generation. Giving appropriate tasks to beginners The strategies of automatic generation for total tasks depend on beginners and teachers. If you want to give partial tasks on which a student feels difficulties and fail to play many times, you can often include the partial task to total task, and if you think that such hard task should be avoided at the beginning stage, the partial task is not employed and using other partial tasks you can generate total tasks that will be felt easy by the student. In general, the number of strategies for giving exercises are huge as well as the number of teachers of playing the piano, proposed system may deal with various kinds of strategies, and by accumulating pattern of presenting total tasks combined with students results, a self-learning system will be expected to realize using the scheme of proposed system. Figure 5. Examples of outputs. INTERACTIVE APPLICATIONS Estimating weak points One of a desirable function of proposed system is estimating weak point of beginners by observing errors of stroking key on piano, errors from an ideal timing, velocity or durations. Estimating the appropriateness of piano performance is discussed at Akinaga s study[1]. It is thought that estimated weak points should be presented to beginners and beginners should make sure of their weak points on playing CONCLUSIONS AND FUTURE WORKS In this study a procedure of practicing playing the piano is proposed by referring conventional procedure of usually employed textbooks, a support system for self-learning is proposed and a subsystem of it is realized by implementing the proposed procedure. Examples of exercises for beginners at the first stage are automatically generated by the subsystem. Future facilities for estimating lacks of playing the piano are discussed. Future works are to realize the estimation facility, to give students appropriate tasks observing weak points of him/her, and to construct an integration system of self-learning playing the piano. ISBN 88-7395-155-4 2006 ICMPC 261

ACKNOWLEDGMENTS This study is partly supported by Grant-in-Aid for Scientific Research (No. 16700154), MEXT, High-tech Research Center promoted by Ryukoku University, and by The Scientific Fund of Casio Computer Co. Ltd.. REFERENCES [1] Seiko AKINAGA, Masanobu MIURA, Norio EMURA and Masuzo YANAGIDA, Toward realizing automatic evaluation of playing scales on the piano, Proceedings of ICMPC9 (in press). [2] Norio EMURA, Masanobu MIURA and Masuzo YANAGIDA, A Support System for Self Education to Improve Basic Skill in Playing the Piano, Technical meeting of musical acoustics, Acoustic Society of Japan, MA2005-6, (2005) (in Japanese). APPENDIX A Proposed procedure for learning playing the piano is shown below. Step 1. The basic training within five degrees without using black keys, for right and left hand. 1-1-R: confirming the position of C,D,E,F,G on the keyboard of piano 1-2-R: playing musical scores concerning two degree. 1-3-R: playing musical scores concerning three degree. 1-4-R: playing musical scores concerning four degree. 1-5-R: playing musical scores concerning five degree. 1-6-R: playing musical scores concerning two, three, four, and five degrees. 1-7-R: playing quarter notes and half note, and learning beats of two-four time and quadruple. 1-n-L (n:1,2,..,5) will be conducted after finishing 1-n-R. Step 2. Playing excerpts using both hands within five degrees. 2-1 Playing basic excerpts for both hands within five degrees. Firstly students should play it of independently. After trained for each hands, the excerpts will be played by both hands simultaneously. 2-2 Basic musical expressions, such as slur, staccato, basic articulations, are taken into consideration on playing the excerpt. Step 3. Introducing black keys and rest notes. 3-1 Playing musical scores concerning more than five degrees for each note. Stretching fingers is required. 3-2 Playing note sequences concerning black keys. 3-3 Playing musical scores having rest notes. Step 4. Leaning basic techniques 4-1 Playing musical scores including eighth-notes. 4-2 Practicing finger movements for playing notes having various kinds of degrees using identical two fingers. 4-3 Learning and playing dotted notes. 4-4 Learning and playing accidental notations, such as sharps and flats. 4-5 Learning and playing irregular metes, such as three-eights, six-eights, and so forth. Step 5. Practicing moving of wrists 5-1 Learning playing music excerpts having wide range of note heights, by mobbing hand-wrists using right hands. 5-2 learning same excerpts as 5-1 using left hands 5-3 learning same excerpts using both hands. Step 6. Playing chords and varying finger spans 6-1 Learning playing chords 6-2 Practicing spreading and shortening finger spans Step 7. Learning applicable techniques. 7-1 Learning dynamics of volume, forte, piano, and so forth. 7-2 Playing musical excerpts having much dynamics. 7-3 Practicing various kinds of fingerings to given musical scores. 7-4 Interchange usage of fingerings. ISBN 88-7395-155-4 2006 ICMPC 262