CSC484A/CSC589A/MUS490/MUS590

(Music Information Retrieval )

Taught by:  George Tzanetakis
(http://www.cs.uvic.ca/~gtzan)
gtzan@cs.uvic.ca

dog gramophone      


News
Outline
Assignments
Projects
Software
Resources
Feedback
Bibliography
Grades

Assignment 3
Due date: By email before December 8.


Assignment 1
Due date: Thursday Oct. 19, 2006 in class

Assignments 2
Due date:  November 16th in class

Assignment(s) 2 will be individualized. In all cases it involves
surveying a topic related to the project of each group and
writing a short (4 page minimum) paper using a standard
conference paper format.

Please use the ISMIR 2006 paper templates for preparing your
assignment:
http://ismir2006.ismir.net/info_authors.html

All your assignments should follow the standard paper
conventions of abstract, introduction, etc and

In all cases try to find as much relevant sources as you
can.  Through the UVic library website you can access
all the papers of IEEE, ACM and the AES (Audio Engineering
Society). In addition online resources such as Google
Scholar might be helpful. The ISMIR online proceedings are also a great source of information:
http://www.ismir.net/#_Proceedings

Also feel free to email me for
advice regarding possible relevant papers.

Try to make your surveys/overviews as informative and
specific as possible. The goal is to write a document
that me and your fellow students can read in order
to understand in depth the particular topic. In addition
to general information try to be as specific as possible
about practical issues such as programming languages,
environments, implementation details etc.


Here are the individual assignments:

1) Lacey Antoniou: Overview of audio plugin formats with
VST plugins as the canonical example. Trace the motivation,
history and evolution of VST (and possibly other plugin
formats). Provide short examples of how one can write
a simple plugin and how it is used by other pieces of software.

2) Tony Antoniou: Survey different pitch detection approaches
tracing their historical evolution. Provide at least 1 example
of a time-domain, frequency domain and perceptually motivated
pitch detector. Try to characterize the different approaches
in terms of accuracy, computational complexity, and difficulty
of implementation.

3) Keith Chan: Explore the literature on automatic chord/key
detection from audio signals, music-score alignment, and
pitch-based representations such as chroma or pitch histograms.
Try to formulate common building blocks between these
different problems and characterize them in terms of
performance and complexity.

4) Aaron Hilton: Do an overview of psychoacoustically motivated
front-ends. A good starting point would be the documentation
of the Auditory Toolbox by Malcolm and the corresponding
references for each algorithm. Describe correlograms and
how they can be used for tasks such as pitch detection.

5) Randy Jones: Write a historic overview of the evolution of
recording techniques starting initially from just faithful
reproduction of a live performance to the highly edited
and mastered recordings of today.

6) Dale Lyons: Do a literature survey of singing voice detection,
singer identification, and synthesis of the singing voice. Try to
identify common building blocks and approaches and suggest
directions for future work.

7) Nathan McDonald: Try to trace the evolution of modifications and special effects to the trumpet sound (for example when were mutes first used ?). Provide examples of music notation from actual pieces that uses different mutes/special effects. You should also provide a short history of the trumpet as an instrument. What are some of the current directions in trumpet technology (for example the silent mute by Yamaha).

8) Terence Nathan Survey different pitch detection approaches
tracing their historical evolution. Provide at least 1 example
of a time-domain, frequency domain and perceptually motivated
pitch detector. Try to characterize the different approaches
in terms of accuracy, computational complexity, and difficulty
of implementation.

9) David Sprague: Explore the literature on music or sound visualizations. Try to go as far back in time as you can and try to include some of the recent trends such as VJing. What are some
of the tools used ? What are the programming languages,frameworks and environments that one could use.

10) Graham Percival: Provide a historical overview of automatically
generated music. A good starting point is the Iliac Suite of Hiller
and Isacson. Another important stop is the Experiment in Music
Intelligence project by David Cope.

11) Adam Tindale: Survey work in automatic analysis and transcription of percussive music going as far back as possible. Describe current approaches, datasets and open problems.

12) Ryan Willoughby : Provide a summary overview
of the New Interfaces for Musical Expression
(NIME) conference: http://www.nime.org/
Select 2-3 papers from each year that you find particularly
interesting and  describe why.

13) Onat Yazir: Do a literature survey of automatic music genre/style classification. What are some of the common building blocks used ? What are some of the challenges ? Suggest directions for future work.