A Machine Learning Approach to Automatic Sound Quality Assessment in Violin Learning

Ramirez-Melendez R, Giraldo S, Waddell G, Williamon A. A Machine Learning Approach to Automatic Sound Quality Assessment in Violin Learning. In: AES Convention 153. Audio Engineering Society; 2022. Paper 40. Available from: https://aes.org/publications/elibrary-page/?id=21923

@inproceedings{RamirezMelendez2022_21923,
  author = {Ramirez-Melendez, Rafael and Giraldo, Sergio and Waddell, George and Williamon, Aaron},
  title = {{A Machine Learning Approach to Automatic Sound Quality Assessment in Violin Learning}},
  booktitle = {AES Convention 153},
  note = {Paper 40},
  year = {2022},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=21923}
}

TY  - CPAPER
TI  - A Machine Learning Approach to Automatic Sound Quality Assessment in Violin Learning
AU  - Ramirez-Melendez, Rafael
AU  - Giraldo, Sergio
AU  - Waddell, George
AU  - Williamon, Aaron
T2  - AES Convention 153
M1  - Paper 40
PY  - 2022
DA  - 2022/10/06
UR  - https://aes.org/publications/elibrary-page/?id=21923
PB  - Audio Engineering Society
LA  - en
AB  - The automatic assessment of music performance has become an area of increasing interest due to the growing number of technology-enhanced music learning systems. In most of these systems, the assessment of musical performance is based on pitch and onset accuracy, but very few pay attention to other important aspects of perfor-mance, such as sound quality or timbre. This is particularly true in violin (and other string and wind instruments) education, where the quality of timbre plays a significant role in the assessment of musical performances. However, obtaining quantifiable criteria for the assessment of timbre quality is challenging, as it relies on consensus among the subjective interpretations of experts. We present an approach to assess the quality of timbre in violin performances using machine learning techniques. We collected audio recordings of several tone qualities. We processed the audio recordings to extract acoustic features for training tone-quality models. Feature information for discriminating different timbre qualities were investigated. Initial computational models were trained using machine learning techniques with selected audio features to provide feedback on tone quality. A real-time feedback system designed for pedagogical use was implemented in which users can train their own timbre models to assess and receive feedback on their performances.
ER  -