Summary of Computational Music Analysis From First Principles, by Dmitri Tymoczko and Mark Newman
Computational music analysis from first principles
by Dmitri Tymoczko, Mark Newman
First submitted to arxiv on: 30 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Sound (cs.SD); Audio and Speech Processing (eess.AS)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents an innovative approach to annotating musical compositions using coupled hidden Markov models. The authors applied their method to annotate 371 Bach chorales in the Riemenschneider edition, a corpus containing approximately 100,000 notes and 20,000 chords. They provided three separate analyses that achieved progressively greater accuracy by making increasingly strong assumptions about musical syntax. Despite minimal human input, the authors’ method was able to accurately identify both chords and keys with an accuracy of 85% or greater compared to expert human analysis. The resulting annotations are accurate enough for music-theoretical purposes while being free from subjective human judgments. This work has implications for longstanding debates on the objective reality of Western harmonic theory’s structures and specific questions about its syntax. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is all about using special computer models to help with music analysis. They took a huge collection of Bach songs and used these models to figure out what notes and chords were being played. The models got really good at it, even without much human help! This can be useful for people who study music theory because it helps get rid of any personal opinions or biases that might come into play when analyzing music. It’s also interesting because it challenges some long-held ideas about how music works. |
Keywords
* Artificial intelligence * Syntax
