Music theorists, musicologists, performers, composers, and conductors are frequently faced with real-world challenges that can typically only be solved through manual labor. Such problems include: “by ear” transcription of a melody embedded in a multi-part performance; Roman numeral and functional harmonic analysis; segmentation and labeling of a piece into distinct tonal (or key) areas; labeling and identification of cadences, structural boundaries, or “repeated” musical fragments; annotation of lyrics, chord symbols, and more. We apply heuristic and machine learning approaches to understand or extract these higher-level musical features from symbolic musical data, with the aim of automating or facilitating these real-world tasks.

Publications

• Condit-Schultz, N. & Ju, Y. & Fujinaga, I. (2018). "A Flexible Approach to Automated Harmonic Analysis: Multiple Annotations of Chorales by Bach and Prætorius," in Proceedings of the International Society of Music Information Retrieval (Paris, France).
• Garfinkle, D., Arthur, C., Schubert, P., Cumming, J., & Fujinaga, I. (2017). "PatternFinder: Content-based Music Retrieval with music21," in Proceedings of the 4th International Digital Libraries for Musicology workshop (Shanghai, China): 5–8.
• Ju, Y., Condit-Schultz, N., Arthur, C., & Fujinaga, I. (2017). “Non-chord Tone Identification Using Deep Neural Networks,” in Proceedings of the 4th International Digital Libraries for Musicology workshop (Shanghai, China): 13–16.
• Condit-Schultz, N. (2016). “The Musical Corpus of Flow: A Digital Corpus of Rap Transcriptions,” Empirical Musicology Review, 11(2): 124–146.
• Arthur, C. (2016). “A Corpus Approach to the Classification of Non-chord Tones Across Genres,” in Proceedings of the 14th International Conference for Music Perception and Cognition (San Francisco, USA): 74–76.