Changes in speech articulation can accompany many biomedical disorders and conditions that affect neurological control of speech. Historically, measuring these changes by acoustic analysis has required either laborious hand measurement or computer programming skill. As part of the output of previous work, we have developed fast automatic computer procedures based on acoustic landmarks for a number of specific studies. These procedures considerably streamline the process of acquiring and analyzing speech production data. In multiple collaborations we have found that the method we have developed to produce these procedures yields significant and useful results in widely separated areas of neurological interest. However, development of these procedures has thus far proceeded in response to the particular needs of individual projects and has not systematically addressed the needs of outside users. In response to PA 08-114, we will expand our library of procedures into a convenient software tool, SpeechMark, for wider use by scientists with or without a background in speech research. SpeechMark's modules will be presented as a menu of specific measures and will be provided in the form of plug-ins ("ports") to existing analysis software tools such as WaveSurfer, the R statistical-analysis system, MATLAB, and Excel. They will offer summary statistics for the measures as well as for additional composite measures that we have found to be reliable indicators of the existence of speech differences between populations or conditions. Further, SpeechMark will be set up to report data from a single audio file, or groups of audio files in a standard table format, for easy input to statistical or other analysis software We plan to develop an organized methodology in, a common Software Development Kit (SDK) as a foundation for every SpeechMark port. This will enable us to maintain existing ports and develop new ones in response to the changing needs of the user community. As envisioned by PA 08-114, the availability of such tools will "jump-start" the use of speech production analysis across a wide range of neurologically important areas of research.

Public Health Relevance

Changes in health status and speaking condition often affect the articulation of speech. Scientific investigation of these changes has been handicapped by lack of access to robust, user-friendly software tools for fast, automatic measurement of speech acoustics. The software development is aimed at making acoustic landmark-detection software, which is focused on articulation, available to a wider community of scientists in the form of add-ons to existing analysis software tools such as WaveSurfer, MATLAB, and Excel.

Agency
National Institute of Health (NIH)
Institute
National Institute on Deafness and Other Communication Disorders (NIDCD)
Type
Small Business Innovation Research Grants (SBIR) - Phase II (R44)
Project #
2R44DC010104-03
Application #
8253437
Study Section
Special Emphasis Panel (ZRG1-BBBP-V (10))
Program Officer
Shekim, Lana O
Project Start
2009-05-11
Project End
2014-12-31
Budget Start
2012-01-13
Budget End
2012-12-31
Support Year
3
Fiscal Year
2012
Total Cost
$508,808
Indirect Cost
Name
Speech Technology/Applied Research Corp.
Department
Type
DUNS #
837257039
City
Bedford
State
MA
Country
United States
Zip Code
01730