An Overview of Recent Computer Music R&D at UCSB
(as presented at ICMC 2004)

Stephen T. Pope

Media Arts and Technology
UC Santa Barbara

Date: Friday, March 4, 2005
Place: Engineering Sciences Building, Room 1001
Time: 2:00 pm — 3:00 pm

Abstract:
The 30th annual International Computer Music Conference (ICMC) took place last November at the Frost School of Music of the University of Miami in Florida. UCSB was represented in a total of seven events: 2 normal paper presentations, one long-format paper presentation, one panel discussion, and two musical performances. This seminar will give an overview of the ICMC and present abbreviated versions of the paper presentations by UCSB researchers. The concrete paper abstracts are given below. The presentation will also include excerpts from music and video performances by Bob Sturm and Stephen Travis Pope.

Feature Extraction and Database Design for Music Software
- Stephen Travis Pope, Frode Holm, and Alexandre Kouznetsov

Persistent storage and access of sound/music meta-data is an increasingly relevant topic to the developers of multimedia software. This paper focuses on the design of music signal analysis tools and database formats for modern applications. It is partly tutorial in nature, and partly a discussion of design issues. We begin with a high-level overview of the dimensions of music database (MDB) software, and then walk through the common feature extraction techniques. A requirements analysis of several application categories will allow us to carefully determine which features might be most useful for them. This leads us to suggest concrete architectural and design criteria, and to close by introducing several of our recent implemented systems. The authors believe that much current MDB software suffers due to ad-hoc design of analysis systems and feature vectors, which often incorporate only low-level features and are not tuned for the application at hand. Our goal is to advance the state of the art of music meta-data extraction and database design by fostering a better engineering practice in the construction of high-level feature vectors and analysis engines for music software.

Concatenative Sound Synthesis and my New-Found Penchant for Unrestrained Thievery
- Bob Sturm

Concatenative sound synthesis (CSS) is a relatively new technique appropriated for musical use from speech synthesis. A sound or composition can be concatenatively synthesized with audio segments from a database of other sounds using feature matching algorithms. Segments of sounds are matched based on similarity of specified feature vectors. A target sound is reconstructed using sound segments from a specified corpus of sounds. Several researchers have implemented various forms of concatenative sound synthesis, but the creative application of this technique is missing. To explore CSS I have implemented a rather simple algorithm using MATLAB. With this program a recording of Mahler can be synthesized using recordings of accordion polkas; howling monkey sounds can be used to reconstruct President Bush's voice. I have used MATConcat to create many interesting and entertaining sound examples, as well as two computer music compositions. These examples will be presented.

Re-inventing the Orchestra: HCI in music performance
- Dan Overholt

In music composition and performance it has always been important to use a variety of instruments in order to create interesting sonic environments. Historically this led to the development of many different acoustic instruments, but musicians have increasingly been using computers to create music - today, various audio synthesis techniques are used to generate sound; these can be viewed as modern corollaries to the different orchestral instruments. These synthesis algorithms are capable of a much wider range of sound generation than their acoustic counterparts, however the interfaces used to control them are predominantly based on traditional instruments such as the piano (MIDI keyboard) or simply use the computer keyboard/mouse. While such standardized approaches are convenient, they limit the range of musical expression that our new orchestra potentially offers. Instead of losing the compelling expressiveness and live performance capabilities inherent in traditional instruments, we should extend our methods of sonic control to a more intricate level by developing new gestural interfaces for electronic music. I have developed several sensor-based instruments to this end, and will explain the ideas behind their creation and demonstrate how they work.