Art and science of singing voice synthesis

• Main Author: Wee, Jade Hean Yu
• Other Authors: Soon Ing Yann
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering; DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/67535
• Institution: Nanyang Technological University

This report aims to recount the process of the author’s Final Year Project (FYP) of implementing a vocoder on the iOS mobile operating system. The vocoder will be implemented as a mobile application, called MicV, as it strives to be easily accessible and readily available to users. Unity is used to develop the vocoder effect on to the iOS platform. The author will reference the source-filter theory to simulate the speech production. Linear predictive coding will be used to obtain the characteristics of an input speech signal, where these characteristics are used to filter a separate input signal. This report includes research and gathering of information, implementation of autocorrelation and Levinson-Durbin recursion, development of mobile application, challenges and limitations occurred as well as future improvements for the application. There have been existing similar applications already available in the market, such as Auto-Tune, but not much has been done on the vocoder effect. Hence, this report also serves as an exploratory project to implement a vocoder effect onto a mobile application. Subsequently, the auhor also tried out different variations of the vocoder effect which includes implementing a virtual piano keyboard, to simulate a “talking keyboard effect”.