Research Projects

Speaker-controlled variability in connected discourse: acoustic-phonetic characteristics and impact on speech perception (2008-2011)
This project investigates why certain speakers are easier to understand than others. Speech production is highly variable both across and within speakers. This is partly due to differences in the vocal tract anatomy and partly under the control of the speaker. This project examines whether clearer speakers are more extreme in their articulations (as measured from the acoustic properties of their speech) or whether they are more consistent in their production of speech sounds. In order to better model natural communication, the speech to be analysed is recorded using a new task aimed at eliciting spontaneous dialogue with specific keywords. The first study investigates whether 'inherent' speaker clarity is consistent across different types of discourse and whether speaker clarity is more closely correlated with cross-category differences or within-category consistency in production. The second study investigates whether clearer speakers show a greater degree of adaptation to the needs of listeners. This study has implications for models of speech perception. Understanding what makes a 'clear speaker' will also be informative for applications requiring clear communication, such as teaching, speech and language therapy, and the selection of voices for clinical testing and for speech technology applications.

Some other pages on our site you may enjoy ...

BROWSE - Simple tool for browsing audio files

BROWSE is a program for browsing audio recordings. With BROWSE you can drag and drop audio files in a range of formats onto the display to see and hear their contents. BROWSE allows you to zoom, scroll and save the audio files to other formats.

ESECTION - Speech signal cross-sections

ESection is a free program for calculating and displaying spectral and other related analyses of sections of a speech signal. It can be used to demonstrate the different spectral properties of elements of speech. It can also calculate an LPC spectrum, autocorrelation and cepstrum analyses, and can display the signal as a waveform or as a spectrogram. It automatically finds formant and fundamental frequency values.

ESYNTH - Harmonic analysis/synthesis teaching tool

ESynth is a free program designed to explain the harmonic analysis and synthesis of signals. With ESynth you can create signals by adding together individual sinusoidal waveforms (sinewaves) and study the resulting waveform and spectrum. You can also perform an analysis of an input waveform, to see how a given sound can be represented in terms of a sum of sinewaves.

FAROSON - The Auditory Lighthouse

FAROSON is a free program for displaying a real-time scrolling coloured pattern from speech sounds. The aim is to construct a pattern that reflects our subjective sensations of loudness, pitch and timbre. The program may be useful in teaching about the nature of sound sensation.

WASP - Record & display speech signals

WASP is a simple program to record speech signals and to display a spectrogram and pitch track.

Rachel Baker