Music Technology Seminar (MUMT621)

Music Information Acquisition, Preservation, and Retrieval


Outline for Winter 2024

ISMIR Submission deadlines: Abstract: 5 April 2024 (topics), Full paper: 12 April 2024

01/09  
01/16

 

01/23

Slide or HTML presentation I (10 min)

  • Cole: The Harmonix Set
  • Jason: Music datasets at Stanford University
  • Kai: WhoSampled.com and cover song datasets
  • Mat: Music visualization tools
  • Lucas: MetaBrainz databases
01/30
  • Classifiers (genre, instrument, mood, performer, composer, rhythm, beat tracking, etc.)

Slide or HTML presentation I

  • Kyrie: Manuscripts on parchment
  • Colin: Multitrack datasets
  • Mila: CompMusic and other non-Western music datasets
  • Hanwen: Early European music datasets
  • Jiawen: MAESTRO and other datasets from Google
02/06

Slide presentation II (12 min)

  • Cole: MusicXML
  • Mat: WAV / WMA
  • Mila: Audio in video formats (MOV, MPEG4, AVI, WMV, etc.)
02/13

Slide-based presentation II 

  • Lucas: Overview of MPEG formats
  • Jason: MEI
  • Kai: MPEG-1
  • Colin: Lyra audio codec
  • Kyrie: Lossless audio formats
  • Jiawen: MIDI format
02/20

Slide-based presentation II

  • Hanwen: Humdrum / **kern

Slide-based presentation III (12 min)

  • Hanwen: Differential privacy
  • Kai: Dynamic Programming
  • Lucas: LSTM
  • Jiawen: Decision Trees (ID3, C4.5, CART)
02/27

Slide-based presentation III 

  • Cole: GAN
  • Jason: Symbolic feature extraction
  • Mat: Audio feature extraction
  • Colin: SVM
  • Kyrie: CNN
03/05
  • Study Week: No class

 

03/12

Slide-based presentation IV (12 min)

  • Mila: HMM
  • Mat: Beat/tempo tracking using audio
  • Kai: Audio source separation
  • Hanwen: Piano music transciption
  • Mila: Genre classification
  • Lucas: Image-based music genre classification

03/19

Slide-based presentation IV 

  • Cole: Lyrics-to-audio alignment
  • Colin: Realtime pitch detection and tracking
  • Jiawen: Instrument classification
  • Kyrie: Handwritten character / music recognition and writer identifications
  • Jason: Mood classification

03/26

Evaluation of MIR: MIREX

Slide-based presentation V (10–12 min)

  • Kyrie: Medieval handwriting identification
  • Cole: Optical font recognition
  • Kai: Source separation using non-negative matrix factorization
  • Jiawen: Piano pedal detection
  • Colin: Recommendation systems based on audio features

Submit a draft of the final project proposal (1–2 pages) plus a partial bibliography

04/02

Slide-based presentation V 

  • Lucas: Multimodal recommendation systems
  • Hanwen: Generative algorithms for Early Music
  • Jason: Symbolic datasets of popular music
  • Mat: Music genre classification using CNN

 

04/09

Final project presentations (8 min)

  • Cole: Engraver Detection
  • Lucas: A Deep Learning Approach to Music Genre Classification using Album Covers
  • Kyrie: Medieval Writer Idnetification
  • Hanwen & Jiawen: Sustain Pedal Detection from Audio Recordings
  • Colin: Exploration of Timbre Transfer Systems and Applications
  • Kai: A Comparative Analysis of Leading Audio Source Separation Approaches
  • Mila: Cultural Classification of Lullabies from Across the Globe
  • Mat: Music Genre Classification Using a Convolutional Neural Network with Mel Spectograms
  • Jason: Stylistic Differences between Anglo-American and Cantonese Popular Music

Submit the final project proposal (1–2 pages) with a full bibliography

Assignments

Assignment #1 (Due 01/23) 4%


Assignment #2 (Due 02/06) 6% Slide-based presentation II (12 min)
  • Slide-based presentation of audio file formats: AIFF, WAV/WMA, Opus (Ogg Vorbis), Lossless formats (FLAC), etc.; MPEG formats: MPEG1 (MP3), MPEG4 (AAC), MPEG7, & MPEG21; symbolic file formats: MEI, MusicXML, MIDI, etc., or audio compression techniques. Submit the slids as a PDF file.
  • Annotated bibliography (short summary and evaluation) as a PDF file (use links, if avialable) (Chicago style, Documentation II: Author-Date References, 17th edition) of the presentation topic.

Assignment #3 (Due 2/20) 10% Slide-based presentation III (12 min)
  • Slide-based presentation of review of topics related to classifiers, e.g.: Feature extraction (audio or symbolic), Feature selection (with Genetic Algorithms) Neural Networks (shallow, deep, CNN, RNN, LSTM, Transformers, etc.), Support Vector Machines, Decision Tree (ID3), Random Forest, Gradient Boosting, Gaussian Mixture Models, Hidden Markov Models, AdaBoost, and Dynamic Programming; include music information retrieval applications. Submit the slids as a PDF file.
  • Written summary (2–3 pages, single spaced, single or double columns, 1" margins) plus an annotated bibliography (Chicago style, Documentation II: Author-Date References,17th edition) of the presentation topic.

Assignment #4 (Due 03/12) 10% Slide-based presentation IV (12 min)
  • Slide-based presentation on topics of transcription, recognition, annotation, and classification. It may include general review of transcription systems (monophonic, polyphonic, separation): Beat-box transcription, singing transcription, beat/tempo tracking using audio, beat/tempo tracking using MIDI, blackboard polyphonic transcription, Goto's works, historical overviews, real-tme pitch tracking, piano music transcription, timbre recognition, genre classification, review of a component of jMIR, etc. Submit the slides as a PDF file.
  • Written summary (2–3 pages) with an extensive bibliography (Chicago style, Documentation II: Author-Date References, 17th edition), which should include practically all significant papers on the subject. You do not need to annotated them but should reference each one of them in the summary paper.

Assignment #5 (Due 03/26) 10% Slide-based presentation V (10–12 min)
  • Slide-based presentation on review of topics related to similarity, query systems, recommendation / playlist systems, or exploratory research on topics related to your final project. Submit the slides as a PDF file.
  • Written summary (2–3 pages) plus an extensive bibliography (Chicago style, Documentation II: Author-Date References: 17th edition) of the presentation topic.

Assignment #6 (Due 03/26) 5%

  • A draft of the final project proposal (1–2 pages), including a tentative title, subgoals, plus a partial bibliography (Chicago style, Documentation II: Author-Date References, 17th edition).

Assignment #7 (Due 04/09) 5%

  • Final project presentation describing what you intend to do (5–8 min). Submit the slids as a PDF file.
  • Final project proposal (1–2 pages) with the title and a full bibliography (Chicago style, Documentation II: Author-Date References, 17th edition).

Final project (Due 05/02) 40%
  • Software project with description (2–3 pages) (Repository on Github) (ISMIR style)

or

  • Research paper (5–10 pages) (ISMIR style)

Improving writing skills for McGill graduate students: Graphos

Reading list

Participants

  • Cole Thierrin
  • Colin Raab
  • Hanwen Zhang
  • Jason Lee
  • Jiawen Mao
  • Kai Mikkelsen
  • Kun Fang
  • Kyrie Boressa
  • Lucas March
  • Mat Vallejo
  • Mila Bertolo
Created: 2004-01-02 Modified: Ichiro Fujinaga <
McGill Crest