|
|
|
||
|
The subject builds on the introductory talk on computational music processing via hands-on work on selected problems: polyphonic music transcription, audio fingerprinting and music generation. There will be short homework for each. We will use Python and music-specific libraries (librosa, music21).
The subject does not require knowledge of machine learning. We assume knowledge of Programming I and II, Algorithms and data structures I (ideally II as well), basic math subjects, and that you took the Computational music processing introductory subject.
The subject will be taught in English.
Last update: Mírovský Jiří, RNDr., Ph.D. (18.04.2024)
|
|
||
|
In the first third of the semester, the student formulate project topics. The “zápocet” is given for completed and briefly presented homework (33 %) and a semestral project: implementation (34 %) and documentation (33 %). Using the topic also for your Individual software project is possible (if the superviser of the software project agrees, and if there is a clear designation of which items are for this subject and which are for NPRG045), but this will require a larger amount of work and/or signing up for a corresponding Bachelor thesis, in accordance with the conditions of NPRG045.
The semestral project must be presented at one of the last sessions in the semester (possibly also as work in progress, especially in case of overlap with NPRG045). It is also compulsory to briefly present at least one homework solution. Last update: Mírovský Jiří, RNDr., Ph.D. (22.03.2024)
|
|
||
|
McFee, B., Matt McVicar, Daniel Faronbi, Iran Roman, Matan Gover, Stefan Balke, Scott Seyfarth, Ayoub Malek, Colin Raffel, Vincent Lostanlen, Benjamin van Niekirk, Dana Lee, Frank Cwitkowitz, Frank Zalkow, Oriol Nieto, Dan Ellis, Jack Mason, Kyungyun Lee, Bea Steers, … Waldir Pimenta. (2023). librosa/librosa: 0.10.1 (0.10.1). Zenodo. https://doi.org/10.5281/zenodo.8252662
Müller, Meinard, and Frank Zalkow. "libfmp: A Python package for fundamentals of music processing." Journal of Open Source Software 6, no. 63 (2021): 3326. https://joss.theoj.org/papers/10.21105/joss.03326.pdf
Lerch, Alexander. An Introduction to Audio Content Analysis: Music Information Retrieval Tasks and Applications. 2nd Edition. New York: Wiley-IEEE Press, 2021. Freely available as slides: https://github.com/alexanderlerch/ACA-Slides and accompanying code: https://github.com/alexanderlerch/pyACA and website: https://www.audiocontentanalysis.org/
Dorien Herremans, Ching-Hua Chuan, and Elaine Chew. 2017. A Functional Taxonomy of Music Generation Systems. ACM Comput. Surv. 50, 5, Article 69 (September 2018), 30 pages. https://doi.org/10.1145/3108242
Morreale, Fabio, Megha Sharma, and I. Wei. "Data Collection in Music Generation Training Sets: A Critical Analysis." ISMIR 2023 (2023). https://researchspace.auckland.ac.nz/bitstream/handle/2292/65322/ISMIR_2023-1.pdf?sequence=1 Last update: Mírovský Jiří, RNDr., Ph.D. (22.03.2024)
|
|
||
|
1. Conditions of completing the subject, project topics. Introduction to first problem: polyphonic music transcription to MIDI. Datasets, evaluation methods. 2. Polyphonic music transcription: deterministic methods, non-negative matrix factorization. 3. Polyphonic music transcription: machine learning. 4. Polyphonic music transcription: presentations, discussion of methods and results. 5. Music fingerprinting: introduction to problem, datasets, evaluation methods. 6. Music fingerprinting: representations without and with machine learning. Indexing. 7. Music fingerprinting: cover song identification. 8. Music fingerprinting: presentations, discussion of methods and results. 9. Music generation: defining subproblems (stylistically defined vs. free, symbolic vs. audio), datasets, automated evaluation options. 10. Music generation: symbolic level. Deterministic and aleatoric approaches, mosaics. 11. Music generation: symbolic level. Machine learning: linearization, tokenization. Audio tokenization and audio generation. 12. Music generation: user interfaces and interactivity. Utilizing existing models. 13. Project presentations. 14. Project presentations.
The subject does not touch digital music and sound production and sound engineering: live coding, working with DAWs, mixing, VST plugin implementation etc. are all out of scope. Last update: Mírovský Jiří, RNDr., Ph.D. (22.03.2024)
|
|
||
|
Designed for 2nd-3rd year Bachelor students or 1st year Masters students (esp. if there is interest in music-related theses). Last update: Mírovský Jiří, RNDr., Ph.D. (18.04.2024)
|