Exploring cross-cultural music similarity through the lens of computational musicology

Instructors: Kaustuv Kanti Ganguli and Vidya Rangasayee

Course overview: This intensive week-long course offers a deep dive into the interdisciplinary field of Computational Musicology, through a combination of lectures, hands-on sessions, and interactive discussions. Students will explore the intersection of music, technology, and cognition; gaining practical skills and theoretical insights to analyze and understand music from diverse cultural traditions.

Target audience: We do not expect students to have any pre-requisite knowledge. A knack for listening to music and an enthusiasm for exploring new perspectives are sufficient. Learning music or having basic programming experience is a plus. Anyone and everyone is welcome.

Day / Module 1: Introduction to Computational Musicology
● Lesson 1: Overview of Computational Musicology
● Lesson 2: Historical perspectives and music information retrieval (MIR) tasks
● Lesson 3: Theoretical foundations: music theory and cognitive science
● Learning Outcomes: Understand the scope and significance of Computational Musicology. Identify critical theoretical frameworks and historical developments in the field.

Day / Module 2: Exploring non-Eurocentric traditions
● Lesson 1: Introduction to non-Western music traditions
● Lesson 2: Indian raga and Turkish makam systems
● Lesson 3: Computational approaches to analyzing global music traditions
● Learning Outcomes: Appreciate the diversity of global music traditions. Explore computational techniques for analyzing and understanding non-Eurocentric musical structures.

Day / Module 3: Signal processing for music analysis
● Lesson 1: Fundamentals of digital signal processing (DSP)
● Lesson 2: Spectral analysis and time-frequency representations
● Lesson 3: Music similarity and its manifestation into perception
● Learning Outcomes: Gain proficiency in extracting and analyzing music signals. Learn techniques for feature extraction and visualization.

Day / Module 4: Machine learning in music & evaluation methodology
● Lesson 1: Introduction to machine learning for music
● Lesson 2: Classification, regression algorithms, and pattern recognition
● Lesson 3: Music cognition and behavioral methods from psycho-musicology
● Learning Outcomes: Understand the application of machine learning algorithms in music analysis and generation. Understand perception and cognition aspects and evaluation of the techniques in subjective and objective domains.

Day 5: Discussion and presentations
● Session 1: Student presentations
● Session 2: Final wrap-up and long-term project ideation and brainstorming

Assignment: Non-obligatory integrative project
Students will work collaboratively to design and implement a computational musicology project integrating concepts and techniques covered throughout the course. An open-ended project will gradually unfold with each day’s content, and students will be free to improvise and choose individual or group projects.

Learning Outcomes: Apply theoretical knowledge and practical skills to a real-world project. Demonstrate proficiency in individual expertise and interpretation of results from complementary perspectives.

Suggested readings
          1. FMP notebooks: https://www.audiolabs-erlangen.de/resources/MIR/FMP/C0/C0.html
          2. Music Information Retrieval notebooks: https://musicinformationretrieval.com/