CCRMA Open House 2013
The Open House is an excellent opportunity to see the wide range of interdisciplinary research and creative work being done in computer music, human-computer interaction, digital signal processing, psychoacoustics, and more.
12:00 - 5:00 Demos and posters throughout the building
5:30 - 7:30 Reception at CCRMA (click here for directions to the Knoll) Parking on campus is free after 4 PM.
Bing Concert Hall
327 Lasuen St Stanford, CA 94025
Parking is available in the Galvez Lot.
An Automatic Music Score Following SystemZhengshan Shi
Score alignment is the process of automatically following a music performance with respect to the score by tracking all of its audio events. This project constructs an audio-to-score alignment system using chroma features of musical octaves. And based on the Dynamic Time Warping(DTW), the system associates musical events in a score with time segments of an audio signal.
The Black Box
Romain Michon, Myles Borins, David Meisenholder
firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
Black Box is a site based installation that allows users to create unique sounds through physical interaction. The installation consists of a geodesic dome, surround sound speakers, and a custom instrument suspended from the apex of the dome. Audience members entering the space are able to create sound by striking or rubbing the cube, and are able to control a delay system by moving the cube within the space.
BlindControl: Rocking the Keys with a Multi-Touch Interface
BlindControl is an iPad app that allows keyboarders to modify their keyboard sound in a simple and intuitive way using multitouch gestures. Unlike other iPad apps, this app does not require the user to look at it, and is especially designed with professional live musicians in mind. Come to see it in action and play with it yourself!
Sarah Smith, Priyanka Shekar, Jimmy Tobin
email@example.com, firstname.lastname@example.org, email@example.com
Dough-Re-Mi is a new electronic musical instrument for voice manipulation with play dough controllers. It derives from the concept of embodying one's voice within pliable matter and molding it. In performance, a number of instrument-players may form a choir; the resulting music is a polyphony of voices, each under the maker's handwork.
Examining the effects of rhythmic structure on pitch perception
Megha Makam, Takako Fujioka, Jonathan Berger, Craig Heller
We aim to first reproduce the results in a previously published study (Jones et al 2002); their study showed that isochronous auditory stimuli can drive attentional resources to shift. This was reflected by the subjects' evaluation of pitches of tones presented at expected or unexpected moments. Our study will additionally build upon that paradigm to show that non-isochronous auditory rhythms may also drive temporal expectancies and shift attentional resources according to the metrical hierarchy of the auditory rhythm.
Gobang Music Chess
The Gobang Music Chess is a real-time sonified chess board for the game 'goBang.' Each new sound generated is dependent on the board cell triggered by the user. The music is changing realtime according to the user's control.
Interactive performance environment
The real-time interactive performance environment provides components to perform and control sound synthesis and algorithmic music generation by live coding or using human interface device like wiimotes and webcam.
Node, Sockets, OSC, Inclusive Design, Code Generation... Cool!
Interface.js is a fully functional web application offering a simple OSC interface to parse the state of five multi-touch fingers and three separate axes of accelerometer data. The application works on any device that conforms to w3c standards, and it been tested on both android and IOS.
Internet Rooms from Internet Audio
Chris Chafe, John Granzow
Until now, “network room” has been an enclosed space only in metaphor, where for the connected inhabitants the place is a network location to gather and interact (chat rooms, game rooms, etc.). We describe an actual network room for musical interactions in which plausible, room-like reverberation is spawned between endpoints of our network audio software. Each new user becomes a node in a mesh and all sounds entering the mesh are reverberated by the mesh. The medium in which these echoes exist is not, however, air but the Internet and the acoustical properties differ because of the medium's distinct “physical laws.” We focus here on a project which creates a mesh of local area internet rooms, a LAIR for the purpose of a sound installation which highlights the differences of this new acoustical medium. The paper describes a working implementation for distributed reverberation.
An Introduction to Interactive Sound Source Separation
Nicholas J. Bryan
In applications such as audio denoising, music transcription, music remixing, and audio-based forensics, it is desirable to decompose a single-channel recording into its respective sources. One of the most effective methods to do so is based on non-negative matrix factorization and related probabilistic models. Such techniques, however, typically perform poorly when no isolated training data is given and offer no mechanism to improve upon poor results. To overcome these issues, we present a new method of incorporating user-feedback into the source separation process. An initial prototype
user-interface shows the propose method can achieve high-quality separation results compared to prior work and even perform well without training data.
Microtonal music software
The microtonal music notation software is based on some C++ open source solutions providing microtonal scales and notations which can be played back in real-time. It contains an instruments sample bank with different articulations that are independently modifiable in the mixer.
Real-time Key Analysis
Craig Stuart Sapp
Play some music on the keyboard. The computer will listen to what you are playing and try to guess the musical key. Its choice of key will be displayed on the computer screen as well as played back on the synthesizer. When the computer is certain, its performance of the tonic will be loud; when uncertain, the computer will play timidly.
Michael Sturtz, Bill Burnett, Dan Klien, Carr Wilkerson, Andrew Evans
Students work in collaborative groups to learn and apply the design thinking processes to reinvent the theater experience. Small student groups identify, define, needfind, ideate and prototype the elements necessary to create a new artistic genre of live performance that will utilize technology in new ways and embody what is unique to the Silicon Valley / SF Bay Area. This multidisciplinary class leverages different technical and creative disciplines to create an accessible and radical collaborative performance atmosphere.
Romain Michon Research Overview
This poster is a short synopsis of the of the most recent research projects carried out by Romain Michon:
- Faust2android: a Faust architecture for Android.
- The Faust-STK: a set of physical models of musical instruments written in the FAUST programing language.
- The chant-library for OpenMusic: a re-implementation of the program CHANT for OpenMusic.
SAFVY: auto-tagger visualizer
Jorge Herrera, Juhan Nam
Spawning from research on sparse feature learning for music annotation, we developed a real-time visualizer to demonstrate the system working.
Single-Channel Singing Voice Separation
Source separation has been studied for a long time, but the problem of single-channel separation, and the specific case of the singing voice in music recordings has become a subject of research rather recently. Our project explores the capabilities of a simpler approach to source separation, the Probabilistic Latent Component Analysis (PLCA) method. This project evaluates how the performance of the PLCA changes when we confront it to different practical scenarios, varying the level of information about the sources. In particular, we study the possibility of deriving a universal model for singing voice instead of training the model on a given singer.
software BitBending in action
Kurt Werner and Mayank Sanganeria
A hands-on demo of a software-modeled "BitBent" FM synthesizer, built with our C++ library for digital integrated circuit modeling. "Bit Bending" is a particularly fertile technique for circuit bending that deals with short-circuits and manipulations upon digital serial information. Circuit bending, the process of creatively modifying or augmenting sound-producing electronic devices, occupies an increasingly important musical and cultural niche. Though the practice began in the 1960s (and traces roots to Leon Theremin's experiments with radio tubes in the 1920s), it is still understudied.
tulpasynth & Jnana
"tulpasynth" is a collaborative music system that enables a group of people to spontaneously create together by manipulating a physics-based environment on a touchscreen interface.
"Jnana" is an algorithmic accompaniment tool integrated within Ableton Live. It has the ability to generate rhythms and melodies based on the musical material used as input.
Uniform Noise Sequences for Nonlinear System Identification
Francois Germain, Jonathan S. Abel, Philippe Depalle, Marcello M. Wanderley
Noise-based nonlinear system identification techniques using Hammerstein and Wiener forms have found wide application in biological system modeling, and been applied to modeling nonlinear audio processors such as the ring modulator. These methods apply noise to the system, and project the system output onto a set of orthogonal polynomials to reveal parameters of the model. Though Gaussian sequences are invariably used to drive the unknown system, it seems clear that the statistics of the input will affect the model estimate. Motivated by the limited input and output ranges supported by analog systems, the use of input noise sequence having a uniform distribution is explored. Simulations results identifying Hammerstein and Wiener systems show that the uniform and Gaussian distributions perform differently, with the uniform distribution generally producing more accurate noise and harmonic responses.
LECTURES (In Order of Appearance)
Simulation and Evaluation of Acoustic Environments
Michael Vorlaender, Stefan Weinzierl, Hans-Joachim Maempel
Perceptions of sound in rooms is still subject to research. The SEACEN group (www.seacen.tu-berlin.de) is currently working on a coordinated effort to improve the complete signal chain from the numerical modeling, the data acquisition within numerical or real sound fields, the coding and transmission to the electro-acoustic reproduction by binaural technology or by sound field synthesis. A novel approach for the comparative evaluation of real and simulated environments will also enable the evaluation of the plausibility and/or the authenticity of virtual acoustic environments and a contribution to the development of better physical metrics to predict the qualities of concert halls and opera houses.
Designing Performable Laughter
Jieun Oh, Ge Wang
Expressions in paralanguage are inherently indicative of our physiological or affective state despite the apparent crudeness of the signal. In this talk, we focus on laughter expressions through interactive musical synthesis, exploring how social and emotional meaning can arise from a phrase of appropriately-shaped vocalized notes in laughter. We will describe a model for interactive, performable laughter, and evaluate its efficacy.
An Exploration of Tonal Expectation Using Single-Trial EEG Classification
Blair Kaneshiro, Jonathan Berger, Marcos Perreau Guimaraes, Patrick Suppes
We introduce basic methodologies of machine learning and demonstrate its use on single-trial EEG in a tonal expectation paradigm. EEG was recorded while participants listened to chord progressions, in three keys, with cadential resolutions to the tonic, dominant, flatted second, or silence. Single trials of EEG responses to cadential events were classified, with the goal of correctly identifying the stimulus that produced the response. Classification of responses by harmonic function of cadential endings across keys produced classifier accuracies significantly above chance. We additionally show that single-trial classification can be used to identify task-relevant temporal and spatial components of the EEG response. Using only top-performing time ranges or electrodes produced classification rates approaching and even exceeding those obtained from the full response.
Introduction to Faust
Romain Michon, Julius Smith
Faust (Functional AUdio STream) is a programming language that provides a purely functional approach to signal processing while offering a high level of performance. FAUST aims at being complementary to existing audio languages by offering a viable and efficient alternative to C/C++ to develop signal processing libraries, audio plug-ins or standalone applications.
While Faust is primarily being developed at GRAME (France), a community of developers at CCRMA is actively contributing to the development of this language. This talk aims to give a broad overview of Faust.
An Introduction to Interactive Sound Source Separation
Nicholas J. Bryan
In applications such as audio denoising, music transcription, music remixing, and audio-based forensics, it is desirable to decompose a single channel recording into its respective sources. One of the current most effective class of methods to do so is based on non-negative matrix factorization and related latent variable models. Such techniques, however, typically perform poorly when no isolated training data is given and offer no mechanism to improve upon poor results. To overcome these issues, we present a new method of incorporating user-feedback into the source separation process. An initial prototype user-interface shows the proposed method can achieve high-quality separation results compared to prior work and even perform well without training data.
Wicked Problems and other Design Considerations in Composing for Laptop Orchestra
Composing music for ensembles of computer-based instruments such as laptop orchestra is a multi-faceted and challenging endeavor whose parameters and criteria for success are ill-defined. In the design community tasks with these qualities are known as "wicked problems".
This talk will frame composing for computer-based ensembles as a design task, discuss its wicked properties, and show how explicitly addressing the themes of visibility, risk, and embodiment can help guide the composer/designer.