Speech Recognition with Dragon NaturallySpeaking and WINE
Date and Time
Wednesday, April 21, 2010 from 6:30 pm to 9:00 pm
Location
MIT Building E-51, Room 335
Presenters
Susan Cragin
Summary
A demonstration of installing and running Dragon NaturallySpeaking on Linux using Wine
Abstract
One of the most complex programming tasks has been the development of Continuous Speech Recognition (CSR). Companies started to develop it only when government funding was available, in the early 1990s. When that funding was pulled, the industry shrank and consolidated, and now the only two good CSR engines are closed-source and Windows-centric, and owned and developed by Nuance Communications. And only one of those engines, for Dragon NaturallySpeaking, is still being developed.
But CSR has been shown to be an invaluable resource.
- It speeds work and lowers costs, particularly in the medical field, which documentation relies on narrative descriptions using complex medical terms. In the legal field, certain documents can be drafted quickly. In building inspection, notes can be made on-site in narrative form.
- It assists persons who have trouble typing. It also can prevent repetitive strain injuries in the arms and hands.
- It can be an aid to learning, and fosters creativity. This is something that is only starting to be understood.
Linux needs CSR. Developing its own would take years, but Dragon NaturallySpeaking 10 is available now, with blazing speed and accuracy, especially considering the overall sorry state of Linux audio.
Susan Cragin will demonstrate how NaturallySpeaking can be installed using wine, how it works and what it's limitations are, and how it can interact with native Linux programs.
Cragin is a WINE Maintainer of DNS10 Standard.
