<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>The piano education industry occupies a huge market. However, the automatic piano performance evaluation function has shortcomings in the existing piano education. In this paper, we first design a piano performance evaluation model based on the continuous discretization method for piano performance audio and improve the piano performance feature filtering and feature extraction methods through cubic spline interpolation and support vector regression, respectively. Secondly, a piano performance evaluation system based on music theory knowledge is studied to evaluate the user’s performance ability based on the rating of rhythm, main melody, and musicality performance of different levels of score performance audio and to achieve the purpose of assisting piano teaching. The results show that the difference between the scoring based on the continuous, discrete model and the actual level value of the player is less than 0.05, and the output of the evaluation system can achieve consistency; in the comparison experiments, the accuracy of three traditional evaluation models, BP, RNN, and LSTM, is 45%, 42%, and 75%, respectively, while the accuracy of the continuous, discrete model is 78%, with an average performance improvement of 23%. The continuous, discrete piano performance evaluation model proposed in this paper is more effective and objective in evaluating users’ piano performance ability and provides a guiding reference for performance evaluation in the piano education industry.</p>
</abstract>

The piano education industry occupies a huge market. However, the automatic piano performance evaluation function has shortcomings in the existing piano education. In this paper, we first design a piano performance evaluation model based on the continuous discretization method for piano performance audio and improve the piano performance feature filtering and feature extraction methods through cubic spline interpolation and support vector regression, respectively. Secondly, a piano performance evaluation system based on music theory knowledge is studied to evaluate the user’s performance ability based on the rating of rhythm, main melody, and musicality performance of different levels of score performance audio and to achieve the purpose of assisting piano teaching. The results show that the difference between the scoring based on the continuous, discrete model and the actual level value of the player is less than 0.05, and the output of the evaluation system can achieve consistency; in the comparison experiments, the accuracy of three traditional evaluation models, BP, RNN, and LSTM, is 45%, 42%, and 75%, respectively, while the accuracy of the continuous, discrete model is 78%, with an average performance improvement of 23%. The continuous, discrete piano performance evaluation model proposed in this paper is more effective and objective in evaluating users’ piano performance ability and provides a guiding reference for performance evaluation in the piano education industry.

Quantitative Evaluation of Piano Performance Technique and Style Based on Continuous Discretization Method

Department of Instrumental Music Education, School of Music Education

Conservatory of Music, Guangxi University of Arts

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

The piano education industry occupies a huge market. However, the automatic piano performance evaluation function has shortcomings in the existing piano...

Quantitative Evaluation of Piano Performance Technique and Style Based on Continuous Discretization Method

Publié en ligne: 26 juil. 2023

Pages: -

Reçu: 07 oct. 2022

Accepté: 03 févr. 2023

DOI: https://doi.org/10.2478/amns.2023.2.00093

Mots clés
Piano performance, Continuous discretization, Spline interpolation, Vector regression, Feature extraction

© 2023 Dandan Mao et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Quantitative Evaluation of Piano Performance Technique and Style Based on Continuous Discretization Method

Publié en ligne: 26 juil. 2023

Pages: -

Reçu: 07 oct. 2022

Accepté: 03 févr. 2023

DOI: https://doi.org/10.2478/amns.2023.2.00093

Mots clésPiano performance, Continuous discretization, Spline interpolation, Vector regression, Feature extraction

© 2023 Dandan Mao et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Mots clés
Piano performance, Continuous discretization, Spline interpolation, Vector regression, Feature extraction