MUS423 Research Seminars
The CCRMA Music 423 Research Seminar brings graduate students and supervising faculty together for planning and discussion of original research. Students and faculty meet either in small groups or individually, as appropriate for the research topics and interests of the participants. Research carried out is typically presented at the weekly CCRMA Colloquium (if it is of general interest to the CCRMA community) or at a Special DSP Seminar scheduled for that purpose. In either case, announcements appear on the CCRMA Home Page as Upcoming Events.
Recent DSP Seminars
Joseph Tilbian: "Stride: A Domain-Specific Language for Sound Synthesis, Processing, and Interaction Design"
Date:Thu, 04/06/2017 - 5:15pm - 6:15pmLocation:CCRMA ClassroomAbstract: Stride is a domain-specific language for real-time sound synthesis, processing, and interaction design. Through hardware resource abstraction and separation of semantics from implementation, Stride targets a wide range of computation devices such as micro-controllers, system-on-chips, general purpose computers, and heterogeneous systems. With a novel and unique approach at handling sampling rates as well as clocking and computation domains, Stride prompts the generation of highly optimized target code. The design of the language facilitates incremental learning of its features and is characterized by intuitiveness, usability, and self-documentation.Event Type:DSP SeminarFREEOpen to the PublicAdaptive mixing of noisy and robust beamformers for enhancement, visualization and reproduction of sound fields
Date:Thu, 03/02/2017 - 5:30pm - 7:00pmLocation:CCRMA Classroom [Knoll 217]Event Type:DSP SeminarSpeaker: Symeon Delikaris-Manias , PhD candidate (advisor Ville Pulkki), Department of Signal Processing and Acoustics, Aalto University, Finland
FREEOpen to the PublicNext Generation Sound Synthesis at the Edinburgh Parallel Computing Centre
Date:Tue, 02/28/2017 - 5:30pm - 7:00pmLocation:CCRMA Classroom [Knoll 217]Event Type:DSP Seminar
Abstract: The NESS project (standing for Next Generation Sound Synthesis), funded through a Starting Grant from the European Research Council for five years beginning on January 1, 2012 is an exploratory project, concerned entirely with synthetic sound—and in particular, numerical simulation techniques for physical modelling sound synthesis in parallel hardware. It is a joint project between the Acoustics and Audio Group and the Edinburgh Parallel Computing Centre, both at the University of Edinburgh. The models developed in the course of the project span a large set of systems, including brass, cymbals and gongs, percussions, guitar/fretboard interaction, bowed strings and large 3D room acoustics simulations.
FREEOpen to the PublicAugmented Reality Headphone Reverberation
Date:Tue, 02/21/2017 - 5:30pm - 7:00pmEvent Type:DSP SeminarAbstract: In audio-visual augmented or mixed reality applications, computer-generated audio objects are rendered via acoustically transparent earphones to blend with the physical environment heard naturally by the viewer/listener. This requires binaural artificial reverberation processing to match local environment acoustics, so that synthetic audio objects are not readily discriminable from sounds occurring naturally or reproduced over loudspeakers. Approaches involving the measurement or calculation of binaural room impulse responses in consumer environments are limited by practical obstacles and complexity.
FREEOpen to the PublicDigital Modeling of Expressive Acoustic Instruments
Date:Thu, 01/26/2017 - 5:30pm - 7:00pmLocation:CCRMA Class Room [Knoll 217]Event Type:DSP SeminarAbstract: The increasing demand of expressive virtual instruments has led the research and development of powerful and sophisticated technologies. While samplers and synthesizers have been for decades the cornerstones of virtual music, it’s nevertheless true that limitations of both techniques are well known. The purpose of our research is to find a way to overcome these constraints exploring both the Physical Modeling and the Sample Modeling approach. We have performed an in-depth study of the Physical Modeling techniques, especially the Digital Waveguides Synthesis by Prof. Julius O. Smith.
FREEOpen to the PublicDeep Learning for Audio Applications using TensorFlow
Date:Thu, 10/27/2016 - 5:30pm - 7:00pmLocation:CCRMA Class Room [Knoll 217[Event Type:DSP SeminarAbstract: Deep learning has become an essential part of audio analysis right from information retrieval to synthesis. The tutorial will cover most of the fundamental aspects of training and debugging neural networks. I will go through various tips and tricks of how to approach a problem, what kind of architectures, loss functions to use, how to set-up experiments for audio classification. At the end of the session, students should be able to run and implement some state of the art algorithms in Tensorflow by themselves. Basic programming efficiency is required. It will also cover why and how do they work. Finally we will also explain some of the techniques of decoding why they work and what do these network learn for various audio applications.
FREEOpen to the PublicTuneSplit Semantic Equalization and Remixing
Date:Thu, 04/28/2016 - 6:00pm - 7:00pmLocation:CCRMA Class Room [Knoll 217]Event Type:DSP SeminarTuneSplit Semantic Equalization and Remixing
by Alejandro Koretzky, CEO & Founder at tuneSplitFREEOpen to the PublicReal Time Audio Signal Processing with MATLAB
Date:Thu, 03/31/2016 - 6:00pm - 7:30pmLocation:CCRMA Class Room [Knoll 217]Event Type:DSP Seminar
This talk will present the new MATLAB Audio System Toolbox for streaming live audio in and out of MATLAB/Simulink, along with algorithms and tools for processing, tuning, and analyzing such live audio signals. Low-latency processing is possible via ASIO support under Windows or Core Audio support on a Mac.
FREEOpen to the PublicWave Digital Filter Tutorial—Part II
Date:Mon, 01/25/2016 - 6:30pm - 8:00pmLocation:CCRMA ClassroomEvent Type:DSP Seminar
Come one, come all, to the 2nd of a 3-part tutorial series on everyone's favorite lumped system discretization formalism: Wave Digital Filters (WDFs). WDFs can be used to simulate acoustic systems (e.g., piano hammer, clarinet reed) and classic audio circuitry (e.g., effects pedals, guitar amps). Though powerful, WDFs can be confusing. In Part I of this tutorial, Ph.D. candidate Kurt James Werner explained the basics of WDFs. In Part II, we'll discuss how to handle nonlinearities in WDFs (e.g. the Tube Screamer distortion pedal [3]).Wave Digital Filters are an active research area at CCRMA.Open to the PublicWave Digital Filter Tutorial—Part I
Date:Mon, 01/11/2016 - 7:00pm - 8:30pmLocation:CCRMA ClassroomEvent Type:DSP SeminarCome one, come all, to the first of a three-part tutorial series on everyone's favorite lumped system discretization formalism: Wave Digital Filters (WDFs)! Alfred Fettweis originated WDFs as a filter design technique in 1970; since the early 2000s researchers have extended his approach to simulate lumped acoustic systems (e.g., piano hammers, clarinet reed) and classic audio circuitry (e.g. effect pedals, guitar amps). Though powerful, WDFs can be confusing. In this tutorial, Ph.D. candidate Kurt James Werner will start from scratch, telling you everything you need to know about the sometimes puzzling and notoriously self-contained WDF literature.
Open to the Public
