• SummerSeminars

Summary of Summery Seminars

Over the summer of 2010, we will have an informal seminar series. We'll talk about some cool stuff people are doing with music technology, look at the science behind it (mainly from conference and journal articles), and generally have fun while increasing the breadth of our exposure to the field.

Not all talks are original material; presenting a survey of "state of the art" work by the top researchers in the field is perfectly acceptable (including citations to their original papers).

Schedule

• Time: 11am - 12am

  Room: Rankine 407

unscheduled talk offers/ideas

"Musical Genre Classification: can computers recognize classical music?" (Graham, 20 minutes)

Talk abstracts

Physical Modeling: simulating realistic musical instruments with second-order differential equations

Creating realistic instrument sounds is a difficult task. We can record sounds and play them back, of course -- but what if we want to create different sounds? A violinist playing "sweetly" sounds nothing like a recording of a "bold" violin playing at half the volume!

The major focus of sound synthesis over the past ten years has been physical modeling. By imitating the physical behaviour of instruments (waves traveling along a string, in a column of air, or along a surface), we can produce virtual instrumental sounds which are much closer to the real thing.

Drawing on:

• Julius O. Smith III, excerpts from many lecture slides and papers on his fantastic websites about digital audio at CCRMA: [https://ccrma.stanford.edu/~jos/wg.html]

• Sterling, Dong, Bocko, "Representation of solo clarinet music by physical modeling synthesis", ICASSP 2008.
• Larson, "Creating a virtual cello", Music 421 final project, Standford 2003.
• Young, Serafin, "Investigating the performance of a violin physical model: recent real player studies", ICMC 2007.

Musical Robots: bagpipes, clarinet, guitar, and violin

This short talk is mostly a chance to chill and watch movies of cool robots playing real musical instruments. I'll add a few remarks about the mechanics and programming that went into each robot, and I'll try to convince undergraduate students to create a musical robot for their 4th-year team project.

Since this is Scotland, we will begin by examining McBlare, the robotic bagpipe player.

Drawing on:

• Dannenberg, Brown, Zeglin, Lupish, "McBlare, A Robotic Bagpipe Player", NIME 2005.

• Kapur, "A History of robotic musical instruments", ICMC 2005.
• Judge, NICTA/UNSW's 2008 Artemis competition winner (clarinet).
• teamDARE, Artemis 2009 competition winner (guitar).

• Wornle, Chia, Hong, Lee, Lim, "RoboFiddler - A Robotic Violin Player", class project for University of Adelaide mechanical engineering, 2006.

• Kusuda, "Toyota’s violin-playing robot", Industrial Robot 2007.

Visualising Performances

With musical performances being watched and analysed in many different formats (audio, video, motion capture) this presentation is a chance to explore the different possibilities of visualising performance data be it interactive programs with tools for analysis, videos which describe properties of the music or static images which aim to capture information about a whole performance in one picture. Opportunity for an informed discussion on future multi-modal tools for visualisation and analysis will be encouraged!

Drawing on:

• The Performance Worm - Langner and Goebl, Dixon
• Sonic Visualiser - Cannam
• Motion History Images - Davis
• Motiongrams - Jensenius
• Database - Pullinger

Singing Voice Synthesis: the mathematics of Hatsune Miku, the virtual pop idol

Most cyperpunk novels include the notion of computer-generated pop singers (such as William Gibson's Idoru). Well, this is now science fact, not science fiction -- in August 2009, Hatsune Miku (a computer-synthesized singer) gave a "live" concert in Japan. Even more impressively, this synthesis is done with inexpensive consumer software.

This talk examines the digital signal processing which translates a large library of recorded audio samples into the final audio, tracing its development from the initial 2003 PhD thesis in Spain to the latest in Japanese popular music.

Drawing on:

• Loscos, "Issues on Modeling the Singing Voice", PhD thesis, Universitat Pompeu Fabra, Barcelona, 2003.
• Kenmochi, Ohshita, "VOCALOID – Commercial singing synthesizer based on sample concatenation", Interspeech 2007.
• Janer, Bonada, Blaauw, "Performance-driven control for sample-based singing voice synthesis", DAFX 2006.
• Hamasaki, Takeda, Nishimura, "Network Analysis of Massively Collaborative Creation of Multimedia Contents: Case Study of Hatsune Miku videos on Nico Nico Douga", ACM uxTV 2008.

• Nakano, Goto, "VocaListener: A Singing-to-Singing Synthesis System Based on Iterative Parameter Estimation", SMC 2009.

Computer Music Expression: emotional and expressive performances with algorithms

to be added

Drawing on:

• to be added

A Scientific Approach to Science Education

After receiving the 2001 Nobel prize in Physics for his work on the Bose-Einstein condensate, Dr. Wieman turned his attention to the way that universities (particularly physics departments) were teaching students. Using the scientific techniques of repeatable experiments with solid empirical data, his team has identified a number of policies and techniques which vastly improve science education.

I was fortunate to hear him speak, and will give a summary of his work while attempting to use as many of his jokes as I can remember. I will also add a few remarks comparing his team's findings with my experience as a lab demonstrator for two 1st-year courses in this department.

Drawing on:

• Wieman, "Why not try a scientific approach to science education?", http://www.scientificblogging.com/carl_wieman/why_not_try_scientific_approach_science_education

• Dr. Wieman's lecture slides from http://www.cwsei.ubc.ca/resources/other.htm

• Hake, "Interactive-Engagement Versus Traditional Methods: A Six-Thousand-Student Survey of Mechanics Test Data for Introductory Physics Courses", American Journal of Physics 1998.

Computer Music Composition: imitating Chopin (and others)

Computers can beat humans at chess, but can they compose music? Of course they can! ... depending on what you mean by "compose" and "music".

This talk will examine some of the algorithms used in such work. Some projects aim to imitate the style of previous composers, some aim to extend the work of current composers, some include the audience in the music-composition process to create "extreme sight-reading", and one project even changes the style of music based on an EEG of a subject's brain activity!

Drawing on:

• to be added later

Sustainability in F/OSS: developers as a non-renewable resource

The time and energy which developers spend on open-source projects is not an infinite resource. Developer effort can stall due to external demands on their time (such as family, career, or health), but also due to internal factors (such as a loss of motivation or interest). Long-term projects (5+ years old) should try to engage in sustaindable development practices. How can we retain developer interest? How can we prepare for the inevitable loss of developers? How can we train the next generation of developers?

This talk draws upon experiences from GNU/LilyPond (a 14-year old sheet music typesetter), but makes general suggestions (and warnings!) for users, developers, and project leaders.