Research at CSC

 Alvise Vidolin
Alvise Vidolin

The '70s: making the computer sound

Research on voice and speech synthesis
development of digital audio system
data link  IBM 370/145 - IBM S/7
Digital synthesis of musical instruments
Software for computer aided composition

CSC System

 Spectrum


The '80s: the computer as a musical instrument

Real-time sound synthesis
Development of the 4i System

4i
 4i
Two details of the 4i System

The '90s: Expressive Information Processing

Analysis of musical gesture
Models of musical performance

The 2000s: space e multimodality

Space as a compositional and theatrical/narrative parameter
Development of technologies for the musical control of space
Multimediality and multimodal interaction in performance


Elbalmer
The Elbalmer, Giorgio Battistelli
 Medea
Medea, Adriano Guarnieri

A look to the past

The Preservation of Interactive Multimedia Installations (2013): learn more

Several masterpieces are at risk of oblivion
Digitalization:
Crossmodal digitalization (sound synthesis from 2D or 3D model of the carrier)
Protocols for digitalization
Compensation of original equalization by means of virtual analog synthesis
Decoding (dbx, Dolby)
Metatada and contextual information extractions:
Automatic optical detection of tape collations (splices)
Binary tree automatic detection for re-mediation processes using Music Information Retrieval techniques (similarity trees, etc.)
Restoration and reinterpretation to offer novel forms of enjoyment of masterpieces from the past
Frequency domain methods:
Spectral Attenuation; Psychoacoustic model
By signal model:
Extended Kalman Filter (EKF); Sequential modelling through particle filtering; High level modelling and inference for audio signals using bayesian atomic decompositions; Full bayesian models for time-domain polyphonic audio; MCMC methods for continuous time ARMA models
By source model:
Variational approximation methods for audio source separation and restoration; Non-negative Matrix Factorization (NMF); Intended/ancillary source separation algorithms for non–harmonic signals using bayesian harmonic models