Dr. Lonce Wyse

Associate Professor
Communications and New Media Programme
National University of Singapore

I also direct the IDMI Arts & Creativity Lab

Communications and New Media Programme
Faculty of Arts & Social Sciences
Block AS6 #03-07      (Need Directions?)
11 Computing Dr.
Singapore 117416
Tel: (65)6516-7277

lonce.wyse (at) nus.edu.sg
lonce.wyse (at) zwhome.org
 

In Brief

My BS is in Math & Computer Science, and my Ph.D. is in Cognitive and Neural Systems (Boston University, '94).
I think in sound. I directed the development of the FlexEffex and the ASound systems for building and interacting with real-time sound models. In 1999, Mindmaker Inc. acquired our FlexEffex technology, and a subsidiary was opened here in Singapore where I worked as the R&D director for two years. I rejoined the Institute for Infocomm Research in November, 2001 where I worked on applying techniques from the worlds of graphical models (NN's, HMM's etc.) to sound analysis and sound model construction. The ultimate objective is to take sound examples and automatically identify and/or construct sound models - a kind of specialized semantic analysis. Recently my focus has shifted to studying and teaching about the use of sound in the  arts.

My CV is available here.
Here are some family photos...

    Sound modeling, representation, automatic identification,
    Neural Nets and Graphical Models,
    Sound synthesis, techniques, control strategies, etc.,
    Auditory perception.

    The Networked Ensemble

     Listening Strategies for New Media
   
     Audio Game; Singapore/MIT collaboration through GAMBIT
                CNN story on the game

Synthesis:
  Sound Modeling and Synthesis with Java
           Go directly to my Experimental Interactive Sound page.

  Personal view of the history of "Free Music"
        Music that admits any and all sound for its material domain.

Previous:
    FlexFX sound modeling and realtime synthesis.
    Coordinating Special Interest Groups and seminars on

•  Independent Component Analysis (ICA)
•  Graphical Models
• Sound classification
  

     A straight-forward method to construct clean-sounding audio signals from spectrams in real-time:
        (Long version:)  Real-Time Signal Estimation From Modified Short-Time Fourier Transform Magnitude Spectra  IEEE Transactions on Audio, Speech, and Language Processing
        (Short version:) Real-Time Iterative Spectrum Inversion with Look-Ahead 2006 IEEE International Conference on Multimedia and Expo
    Generative Sound Models   Multimedia Modeling 2005.   An overview of issues and applications for models that generate sound.
   A Sound Modeling and Synthesis System Designed for Maximum Usability.  International Computer Music Conference 2003.
   Application of a Content-based Percussive Sound Synthesizer to Packet Loss Recovery in Music Streaming. (Lonce Wyse, Wang Ye, Zhu Xinglei, (2003) , ACM Multimedia)
   Free Music and the Discipline of Sound. (Lonce Wyse, 2003. Organized Sound, 8:3, Cambridge University Press).
   Description for Non-Speech, Non-Note-Music Audio.  Slides from a talk at an MPEG-7 plenary meeting in Tokyo, 1998.
   Toward Content-Based Audio Indexing and Retrieval and a New Speaker Discrimination Technique (Lonce Wyse and Stephen Smoliar, 1997) - A paper I wrote at KRDL with the venerable Stephen Smoliar on a technique for finding the boundaries between sections of audio, including between different speakers.

More publications ...

   Sonic Arts and Sciences  Fall 2003-present, National University of Singapore, University Scholars Program.

    Saul/Jordan Neural Net (Coded in Java, but if you need a C++ interface, just ask).
        This is a feedforward net that can be run backwards or sideways (that is, run like a bidirectional "associative" net).

    Soda Zoo - Physics, dynamics, sound, all rolled into one suite of cool applets (don't forget to turn the sound on with the checkbox at the top of the page!)
    SoundToys - A very nice collection of pointers to artists, web sounds, Flash works,  and noises.
    Luigi Russolo's Intonorumori - The year, 1910 - the action - the introduction of "noise" into music.
   Jamifi - Design your own sounds and ringtones.