Quantitative analysis of piano performance proficiency focusing on difference between hands

Area of Science:

• Music Cognition
• Computational Musicology
• Performance Science

Background:

• Quantitative evaluation of piano performance is crucial for music education and computational rendering.
• Existing methods often overlook the difference between hands (DBH), a potentially vital aspect of performance quality.

Purpose of the Study:

• To investigate the significance of the difference between hands (DBH) in determining piano performance proficiency.
• To explore the utility of DBH features in distinguishing expert pianists from amateurs.

Main Methods:

• Recruited 34 expert and 34 amateur pianists to perform musical excerpts on a Yamaha Disklavier.
• Recorded performances in MIDI format and extracted handcrafted temporal and dynamic features for each hand (right hand - RH, left hand - LH).
• Calculated DBH features by subtracting corresponding features between RH and LH, and employed statistical analyses, Principal Component Analysis (PCA), and Support Vector Machine (SVM) for classification.

Main Results:

• DBH was significantly more pronounced in experts than amateurs across various temporal and dynamic features.
• Experts demonstrated distinct RH playing characteristics, including longer key presses and higher velocities compared to amateurs.
• The inclusion of DBH features alongside basic MIDI features improved the F1 score for differentiating experts from amateurs to 93.6%.

Conclusions:

• Simultaneous control and differentiation of both hands is a critical skill differentiating expert pianists.
• DBH features offer valuable insights and should be incorporated into quantitative piano performance evaluation systems.
• This approach enhances the accuracy of distinguishing performance proficiency levels.