Timbre interpretable representation modelling method integrating self-supervised learning and music theory descriptors

Key Points

• The aim is to develop a modeling method for timbre that integrates self-supervised learning with music theory descriptors.
• Develop a novel timbre representation modeling method.
• Utilize self-supervised learning techniques for data training.
• Integrate music theory descriptors to enhance representation accuracy.
• Demonstrated improved accuracy in timbre representation.
• Highlighted the effectiveness of integrating self-supervised learning.
• Indicated potential benefits for applications in music analysis.