Instructor: Juan Pablo Bello

Office: Room 626, 6th floor, 35 W 4th street

Office hours: Wednesdays 2-5pm

phone: (212) 998.5736 (ext 85736)      

eml: jp[lastname] [at] nyu [dot] edu


MPATE-GE 2623 Music Information Retrieval

____________________________________________________________________________________________


Course Information:


Meetings (Fall 2011): Mondays 4.55 - 6.35pm

Studio F, 8th floor Ed. building.


Overview: This course gives a comprehensive overview of research on the multi-disciplinary field of Music Information Retrieval (MIR). MIR uses knowledge from areas as diverse as signal processing, machine learning, information and music theory. The course will explore how this knowledge can be used for the development of novel methodologies for browsing and retrieval on large music collections, a hot topic given recent advances in online music distribution and searching. Emphasis would be given to audio signal processing techniques.


Goals: Students will undergo advanced training in techniques for automatic music analysis and retrieval. The knowledge that they will acquire, will be relevant for future careers in the music generation, processing, recording, reproduction and distribution industries, with special emphasis on music-oriented on-line services. They will read and understand the literature describing state-of-the-art methods for MIR and gain hands-on experience on the implementation and application of standard and advanced methods by means of MATLAB assignments and a final project.


Pre-requisites: Basic background in digital signal processing and programming is desirable.

____________________________________________________________________________________________


Calendar and Lecture notes:


Lecture notes will be added and/or updated (as pdf files) as the course progresses, sometimes just before the corresponding lecture. Dates are tentative and subject to change


09.12    Introduction / Time-frequency representations

09.19    Time-frequency representations (cont’d)

09.26    Novelty: onset detection (suggested reading)

09.30    (Friday 4-6pm) MBus/MTech meet and greet

10.03    Novelty: onset detection (cont’d)

10.10    Columbus day

10.17    Periodicity: pitch and beat tracking

10.24    Periodicity: pitch and beat tracking (cont’d)

10.28    (Friday 10am-2pm) MBus/MTech Innovation day - Proposal presentations

10.31    Low-level features: timbre analysis - Project proposals (10%)

11.07    Low-level features: timbre analysis (cont’d)

11.14    Harmony: alignment, chord and key recognition

11.21    Harmony: alignment, chord and key recognition (cont’d)

11.28    Structure: form analysis, segmentation

12.05    Sound classification: genre, artist and instrument ID

12.12    Sound classification: genre, artist and instrument ID (cont’d)

12.19    Project demonstrations - Final Report (30%)


The Instructor will provide individual guidance during office hours and by email.


Additionally, two tutors will be available to assist with MATLAB and other practical issues:

Tae Min Cho (tmc323@nyu.edu), Tuesdays 3-5pm, Room 623

Jon Forsyth (jpf211@nyu.edu), Thursdays 3-5pm, Room 623



[Back to top]

____________________________________________________________________________________________


Assessment:


Assignments: 50% (best 5 out of 6) -- see details below.

Projects: 40% -- see details below

    - Proposal + presentation: 10%

    - Final project + presentation: 30%

Class Participation: 10% (readings + discussions, attendance, interest and enthusiasm)


[Back to top]

____________________________________________________________________________________________


Assignments:


There are 6 assignments to be distributed during the semester. All assignments are due one week after they have been distributed or announced. Penalties will apply to delays.


Instructions for the submission of assignments (PLEASE READ CAREFULLY):

  1. 1.Go to the course’s blackboard page (available under “Blackboard Classes” on your home page of the home.nyu.edu site).

  2. 2.Select “Assignments” on the left navigation menu. Under the link of the current assignment, click on the “View/Complete Assignment” button. An “Upload Assignment” panel will open. Attach the solutions to the assignment (see file naming conventions below), make comments if necessary and submit. Please note that saving your files do not submit your answers to the instructor. For more detailed instructions, please visit this Blackboard Tutorial on the use of the Assignment tool.

  3. 3.All files corresponding to the assignment must be uploaded to the blackboard page before 11.55pm on the due date (a penalty will be applied for every day of delay - You should send me an email explaining the reasons for the delay). After 2 days of delay, your assignment will no longer be accepted.

  4. 4.The solutions to every assignment should contain 2 files ONLY: the Matlab source file (i.e. m-file containing the code) and a pdf document containing answers, plots and the discussion of results.

  5. 5.The m-file should include all the code needed to solve the assignment in an ordered fashion. Comment this file THOROUGHLY. I expect to be able to read and run this code and get all results/plots in one go, without needing to change parameters or modify the code in any way. Please be aware that the conciseness. efficiency and presentation of the code will be taken into consideration when evaluating.

  6. 6.Be concise in your answers and comments. Go straight to the point.

  7. 7.Documents should be clearly identified as YourLastNameAssignment# (e.g. Bello3.pdf and, Bello1.m). Additionally, please add the files into a zip file of the same name (e.g. Bello3.zip, Bello1.zip).

  8. 8.Clearly indicate your name in the body of every file (at the top of your pdf document, and as a comment on your m-file).


Keep checking this space regularly for assignments and due dates.


Assignment # 1: Assignment_01_FA11.pdf - due date: monday 09/26

Assignment # 2: Assignment_02_FA11.pdf - due date: friday 10/07

Assignment # 3: Assignment_03_FA11.pdf - due date: monday 11/07

Assignment # 4: Assignment_04_FA11.pdf - due date: monday 11/21

Assignment # 5: Assignment_05_FA11.pdf - due date: monday 12/05


[Back to top]

____________________________________________________________________________________________


Projects and presentations:


  1. • Projects should be done in groups of 2 students each.

  2. • When selecting a topic, students are encouraged to explore applications of the course contents to problems in, e.g. music composition, performance, analysis, dissemination or retrieval.

  3. • The project consists of proposing and implementing a solution to the chosen problem (preferably in MATLAB); writing a report discussing the specifics of the problem, the approach taken, and its results; and presenting the highlights of your work to the class. The solution can be a re-implementation of recent MIR work, typically including a combination of signal processing and machine learning challenges. In all cases, the project should go beyond the  materials covered on the assignments.

  4. • This semester, MM students will collaborate with students from the “Entrepreneurship for the Music Industry” course in the design and implementation of the project proposal, and will jointly present the outcomes of this collaboration at the Innovation Day on Friday 10/28.

  5. • During office hours, the instructor will provide guidance and advise about topic selection, and the execution of the project.


Important dates:


10.28  (Friday 10am-2pm) Proposal Presentations during the Innovation Day: 5’ Talk, 10’ Q&A

10.31  Project Proposals (10%, 4 pages or less): this document should include a project title and clearly explain the proposal by introducing: context, problem, proposed algorithm(s), evaluation method,  brief workplan and bibliographical references. The document should also name the group members and briefly discuss how the work will be divided amongst them.

12.19  Project demonstration and final report (30%, both a written report of no more than 10 pages and source code should be submitted).


Instructions for Final Project Submission and Presentation (PLEASE READ CAREFULLY):

  1. 1.Presentations slots are strictly 15 minutes long: 10 minutes for the presentation and 5 minutes for questions. This time includes changeover time (the time it takes you to set up everything). You should bring the presentation on a laptop and try the projector and sound on the Studio beforehand. This is specially important for those of you who want to use any additional piece of equipment for the demonstration.

  2. 2.Attendance and participation on all presentations is mandatory (ask questions, make comments).

  3. 3.The final project report should not exceed a max. number of 10 pages, and should be written like a conference paper. Structure should be more or less as follows: introduction (including motivation), theoretical background, your approach, implementation, experimental part (including a detailed explanation of the evaluation method), discussion, conclusions and future work.

  4. 4.Students are encouraged to develop their project using Matlab, but are welcomed to use other languages as they see fit (C++, Java, Max/MSP, SuperCollider, etc).

  5. 5.During the project due date you will be expected to demonstrate your software.


Examples of previous projects:

Most projects focus on the re-implementation of existing techniques, e.g.:

  1. -Genre Classification as in the work by Tzanetakis and Cook

  2. -Mood Classification as in the work by Liu, Lu and Zhang

  3. -Chord Estimation as in the work by Bello and Pickens

  4. -Playlist Generation as in the work by Flexer, Schnitzer, Gasser and Widmer

  5. -Event-synchronous music analysis/synthesis as in the work by Jehan

  6. -Chorus section detection as in the work by Goto

  7. -Musaicing as in the work by Casey


A few projects develop original work, sometimes leading to conference publications:

  1. -Chord change detection (see here)

  2. -Automated Rhythmic Transformations (see here)

  3. -Characterization of perceived spaciousness (see here))


[Back to top]

____________________________________________________________________________________________


Recommended Books:


  1. -Zölzer, U. (Ed.). “DAFx: Digital Audio Effects”. John Wiley & Sons (2003)

  2. -Klapuri, A. and Davy, M. (Eds.) “Signal Processing Methods for Music Transcription”. Springer (2006)

  3. -Müller, M. “Information Retrieval for Music and Motion”. Springer (2007)

  4. -Smith, J.O. “Mathematics of the Discrete Fourier Transform (DFT)”. 2nd Edition, W3K Publishing (2007)

  5. -Witten, I. and Frank, E. “Data Mining: Practical Machine Learning Tools and Techniques”. Morgan Kaufmann (2005)

  6. -Further reading will be recommended as the course progresses.



[Back to top]

____________________________________________________________________________________________


Online resources:


Tools:

  1. -Matlab documentation, tutorials, examples: http://www.mathworks.com/access/helpdesk/help/techdoc/matlab.html

  2. -Signal Processing Toolbox documentation, tutorials, examples: http://www.mathworks.com/access/helpdesk/help/toolbox/signal/

  3. -Matlab file exchange: http://www.mathworks.com/matlabcentral/fileexchange/loadCategory.do

  4. -MATLAB array manipulation tips and tricks by Peter J. Acklam

  5. -Data Mining Software: http://www.cs.waikato.ac.nz/ml/weka/

  6. -Sonic Visualizer: http://www.sonicvisualiser.org/


Research Resources:

  1. -ISMIR home (online proceedings, mailing list): http://www.ismir.net

  2. -All ISMIR papers: http://www.ismir.net/all-papers.html

  3. -MIR-related PhD theses: http://www.pampalk.at/mir-phds/

  4. -MIR Evaluation eXchange (MIREX): http://www.music-ir.org/mirexwiki/index.php/Main_Page

  5. -Million Song Dataset: http://labrosa.ee.columbia.edu/millionsong/

  6. -Listing of available datasets (outdated): http://grh.mur.at/sites/default/files/mir_datasets_0.html

  7. -Survey of software tools used by the community (outdated): http://www.music-ir.org/evaluation/tools.html


[Back to top]  

Music Information Retrieval