Musical Shape

Musical shape analysis focuses on identifying and quantifying the overall contour and phrasing of musical performances, aiming to automate aspects of music education and performance assessment. Current research employs machine learning approaches, such as Siamese residual neural networks, to classify musical shapes from datasets of performed pieces, achieving high accuracy in identifying various phrasing patterns. This work has implications for both automated music evaluation and a deeper understanding of how musical structure is perceived and processed, potentially informing music theory and cognitive science.