Dear Users
Today (4-5 pm in YNiC open plan), there will be a YNiC project proposal given by Matt Speed (Department of Electronics) on imaging the vocal tract.
Title: MRI capture of the vocal tract.
Abstract: Acoustic modelling research in the AudioLab in the Department of Electronics focuses on the synthesis of “natural” human voice through the computer simulation of the acoustic properties of the vocal tract. This study aims to generate accurate 3D acoustic models of the vocal tract based on MRI scans of trained speakers/singers for particular phonemes/vowel-based utterances. From this data, new detailed 3D models of the vocal tract will be produced, leading to more accurate simulation of its acoustic characteristics, and the potential for new developments in the field of speech synthesis. A particular challenge is the recording of the audio speech signal during the MRI capturing process, with a view to providing a benchmark against which sound synthesised using the resulting 3D acoustic model can be compared. The preferred approach has not yet been determined, but will be discussed during the project presentation (18th March). Additionally we hope to use our same subjects to capture different styles of singing with a view to identifying and characterising “ringing” qualities of vocal sounds. This is a secondary study that we will brief our subjects on and carry out if time is remaining once appropriate source data for our primary study has been obtained.
Everyone is welcome to attend.
Best wishes Rebecca