This is the website for the Speech and
Hearing Research Program at the Department of Veterans Affairs
in Martinez, California. Since 1978, we have been doing
research which attempts to understand and address specific
speech and hearing difficulties faced by the large community
of veterans living in our community.
is headed by Dr. Pierre Divenyi, whose research into what
is known as the "Cocktail Party Effect" has shed
light onto the important aspects of the difficulties in communicating
that elderly individuals face on a day-to-day basis.
page will provide interested parties with information about
our past, current, and future research, as well as the people
involved in that research, and other related undertakings
of the lab.
Book Just Out —
Speech Separation by Humans and Machines (P.Divenyi, ed.), Kluwer Academic Publishers, 2004
SPEECH SEPERATION AND COMPREHENSION IN COMPLEX ACOUSTIC ENVIRONMENTS —
MONTREAL 2004 WORKSHOP
November 4 - November 7, 2004
This workshop, with the Air Force Office of Scientific Research as its principal sponsor and attended by over 50 participants, continued the discussion started with the 2003 Montreal workshop. It was both a survey and an analysis of the most important questions and problems related to the "cocktail-party effect" performed by human listeners and by machines. The topics included presentations and discussions on sensing arrays, reverberation, integration of different machine approaches, human-machine collaborative systems, neural bases of auditory and visual scene analysis, and evaluation and corpora. Eleven predoctoral and postdoctoral participants (attending thanks to a National Science Foundation award) presented posters. A final general discussion tackled the question of how to design a device capable of performing separation and interpretation of speech presented in complex environments.
ON SPEECH SEPARATION —
an NSF SPONSORED WORKSHOP
October 31 - November 2, 2003
An overlook on the various
methods makes the observer conclude that at least some of the
limitations of each of the schemes could be overcome by integrating
them into a single approach. Such integration, however, has the
prerequisite that proponents of different approaches get together,
present their methods, theories, and data, discuss these openly,
and attempt to find ways to combine schemes, in order to achieve
the ultimate goal of separating speech signals by computational
means. Given the sheer volume of current work on the diverse methods,
such a meeting of researchers is not only necessary but also extremely