Haokun Tian

Profile

Project title:

Timbre Tools for the Digital Instrument Maker

Abstract:

This PhD topic aims at building timbre tools for the digital instrument maker, which is much more fundamental and important. In this age of electronic production, where MIDI is very popular, digital instruments are in great demand. While the mainstream digital instruments on the market rely on recording and signal processing, new AI-based audio synthesis technologies are emerging and showing great potential, so we can seize the opportunity to fill this gap by developing AI-based timbre tools. New technologies may enable creators to design timbres without using MIDI, and they are capable of making synthesis more realistic and imaginative at the same time, i.e., synthesizing acoustic instruments with greater fidelity than signal processing methods, while interpolating to obtain more imaginative timbres (e.g. a timbre in between a piano and a violin).

C4DM theme affiliation:

Communication Acoustics, Sound Synthesis

Research

Research Interests:

Machine Learning for Sound Synthesis, Timbre