Speech
While many people like to see reminder messages, others prefer to hear them. In this work package, we look at ways of designing auditory reminders so that they are easy to understand, even if your hearing is not what it used to be, and even if there is a lot of background noise. Such reminders could be anything from a spoken message to a short tune.
Our work focuses on two questions:
- How well can people understand and learn different types of auditory reminders?
- To what extent do auditory reminders distract people from what they are doing?
To address these question, we are running a series of three studies. In each study, we ask people to act on auditory reminders that they hear while they are doing something else - memorising a sequence of numbers or words, reading text, or walking. We're particularly interested in walking because falls are a big problem for older people living on their own, and we would not want a reminder to startle a person so much that they trip and fall.
As part of the first research question, Karl Isaac is looking at the effect of different types of background noise on the intelligibility of computer-generated messages [AMT paper PDF]. Karl is planning to extend his work to other perturbations that can make speech more difficult to understand, such as reverberation due to room acoustics.
Researchers
Karl Isaac (PhD Student)
Karl is a student in the Centre for Speech Technology Research at the University of Edinburgh. Karl holds a BSc in Software Engineering and an MSc in Mobile Computing. Karl's research will focus on the synthetic speech aspects of the MMH project. In his spare time, Karl likes spending time with his four children, preferably somewhere warm and sunny.
Dr Maria Wolters (Researcher Co-Investigator)
Maria is a senior research fellow at the Centre for Speech Technology Research at the University of Edinburgh. The goal of her research is to improve the accessibility and functionality of voice-based interaction. Maria studied at the University of Bonn, where she attained an MSc in Computer Science in 1997 and a PhD in Communication Research and Phonetics in 2001. She went on to join the University of Newcastle and Queen Margaret University as a clinical phonetician, before moving to the University of Edinburgh in late 2004. She is currently a research fellow on the MATCH project. Maria is married with a son and daughter, whom she homes will one day be trilingual in German, English, and Scottish Gaelic.
Prof Steve Renals (Co-Investigator)
Steve is director of the Centre for Speech Technology Research at the University of Edinburgh with over 150 peer-reviewed publications. His research interests are signal-based approaches to human communication, in particular spoken language processing and machine learning approaches to modelling multimodal data. Steve is an associate editor of IEEE Signal Processing Letters and a former member of the IEEE Technical Committee on Machine Learning and Signal Processing. He is one of the Edinburgh MATCH project primary investigators.
Publications
