Research/Technology

The Research/Technology track is designed for students interested in musically-related research or developing new music technology. Students in this track study with a departmental advisor, and if applicable to a specific research topic, with another member of the broader Johns Hopkins academic community. Research/Technology track students work closely with practicing composer and performers in developing new computer music systems. Areas of research may include psychoacoustics, perception, hardware or software synthesis and control techniques, algorithmic composition and performance systems, and other related topics.

The Hall-O-Deck (Virtual Concert Hall) project and multiprocessor DSP strategies
Utilizing hardware provided by the Intel grant, The Computer Music Department is developing a system with which a performer can rehearse in a practice room, and his instrument's acoustic signal will be processed by NT workstations in such a way that the acoustics of a specific concert hall will be replicated. This is accomplished by the performer playing into a microphone, applying digital signal processing to the signal, and playing back signals into the practice room (through small multimedia speakers) which replicate the acoustics of a desired hall. Initial experiments with "off-the-shelf" processing equipment have been promising. We will, however, need to create models for specific concert halls and develop our own signal processing algorithms.

Intel Grant Projects
The Computer Music Department recently received a grant from the Intel Corporation to pursue several research projects including: The PARIS/Ensoniq/Intelligent Devices Project; Creating a realtime, interactive, networked composition/performance system; and Developing the Virtual Concert Hall.
See project proposals.
See Peabody News.
See Intel Press Releases.

Other Research

• Comparison of pitch detection algorithms
• Transmission of MIDI over UDP/IP for distance education
• Multimodal input capacities
• Report for the Nakajima Heiwa Foundations (JAIST 1998) PDF
• Computer-Aided Perception Enhancement (CAPE) project
• Zeta violin latencies and techniques
• Real-time software synthesis for psychoacoustic experiments
• Latency of audio and MIDI data over LANs
• Vibrato characterisitcs of voice and instruments
• Implementation of exemplar-based learning model for music cognition
• Controlling local CD-device using browser plug-ins (CDLink)
• Converting Levy Sheet Music Collection to MIDI
• Exemplar-based conducting gesture learning
• Music style recognition using lazy learning
• Learning expressive performance by examples
• Timbral recognition using lazy learning
• Chorale harmonization using lazy learning
• Pitch estimation using transients
• Musical signal analysis using silicon cochlear
• Evaluation of Java APIs for music applications
• The Hall-O-Deck project and multiprocessor DSP strategies
• Running Pure Data on a network of multiprocessor Pentium computers
• Musical performance experiments using Datagloves
• Improving k-nearest neighbor search using computational geometry
• Implementing genetic algorithms with floating-point chromosone representation on multiprocessor computers
• Locating positions on CDs by measure numbers
• Library music distribution system based on non-proprietary hardware
• A Study of Early Music on CD-ROM

Researchers: Lilit Yoo, Ichiro Fujinaga

Target: ICMC99 (abstract) (paper PDF)

Researchers: John Young, Ichiro Fujinaga

Target: ICMC99 (abstract) (paper PDF)

Researchers: Dave Sullivan, Ichiro Fujinaga

Target: ICMC99 (abstract)

Zeta violin converts analog sound to MIDI via multi-channel IVL PitchRider hardware.The problem is the delay in calculating the pitch. The purpose of this study is to measure the latency which depends on many factors such as pitch, articulation, dynamics, etc. Also new violin techniques will be investigated to minimize the latency.

Researchers: Lilit Yoo, Ichiro Fujinaga, Geoffrey Wright

Targets: SEAMUS98 (abstract) paper read at SEAMUS98, April 18, 1998 (paper PDF)

Introduction to using real-time software synthesis, Super Collider, MPS, and Pd as valuable tools for psychoacoustic experiments. Small example experiments will be conducted.

Researchers: Stephan Moore, David Sullivan, Ichiro Fujinaga

Targets: ICMPC98 ( abstract) (notes) (paper HTML) (paper PDF)

Researchers: Tony Willert, Ichiro Fujinaga

Target: ICMC

Do violin players slightly flat when playing vibrato?

Researchers: Lilit Yoo, Ichiro Fujinaga

Target: ICMC98 (abstract)

The effect of vibrato on response time in determining the pitch relationship of violin tones

Can the vibrato hide bad intonation?

Researchers: Lilit Yoo, Stephan Moore, David Sullivan, Ichiro Fujinaga

Target: ICMPC98 (abstract) (paper PDF)

The exemplar-based learning model is proposed as an alternative approach to modeling many aspects of music cognition.

Researcher: Ichiro Fujinaga

Target: ICMPC98 (abstract) (paper PDF)

Using Optical Music Recognition system, convert the collection of sheet music to computer-readable format including MIDI, which then can be heard via web clients.

Researcher: Ichiro Fujinaga

NSF grant (1999-2001) abstract, part of Digital Libraries Initiative - Phase 2

How do we recognize timbre? How well can we recognize using just the steady-state portion of musical timbre? Can machines do this in realtime?

Researcher: Ichiro Fujinaga

Toward realtime recognition of acoustic musical instruments
Target: ICMC99 (abstract) (paper PDF)