The Speech Communication Lab (SCL) and related offices are located in the University of Maryland, College Park campus at A.V. Williams Building. SCL, headed by Professor Carol Y. Espy-Wilson, is a part of the Electrical and Computer Engineering (ECE) Department and the Institute for Systems Research (ISR) at the University of Maryland. Members of the lab consist of undergraduate students, graduate students and research scientists.
Research in SCL combines digital signal processing, speech science and machine learning to address issues in speech communication. One major effort is concerned with the development of several components of a speech recognition system based on phonetic features. In particular, there are projects focusing on the signal representation, lexical access and the development of a new paradigm for speech recognition. A second project is focusing on the improvement in quality and intelligibility of speech and the development of tools for speech enhancement. A third project is concerned with articulatory and acoustic modeling. The fourth project concerns speaker identification using a set of acoustic parameters automatically extracted from speech. We have ongoing work on the automatic extraction of such parameters from speech for both speaker and speech recognition purposes. Other research interests involve speech synthesis .
The lab consists of a network of several workstations that are equipped with tools that allow speech analysis, playback and digitization. Software tools include HTK, SPlus, Matlab, Maeda’s vocal tract model, and various word processors. We have also developed VTAR, a Matlab based program for vocal tract acoustic response calculation based on a frequency-domain model.
Sources of support include grants from the government agencies such as the Department of Defense, NSF and NIH, and private agencies such as Speech Technology and Applied Research, the Honda Research Institutes and Hughes Network Systems.