The cost of implementing systems such as CART, CPrint, Typewell and automatic speech recognition can be significant, as they require one transcriber to one lecturer to work. Yet speech-recognition products, like Dragon NaturallySpeaking, turn spoken words into text via a computer without being personally attended. Couldn't this be a much more cost-effective solution for the classroom?
Many students with learning disabilities fi nd it much easier to speak their words in preparing a research paper than to write them. And professors use Naturally- Speaking as a time-saving tool to provide feedback to students on papers and exams. While NaturallySpeaking is already well utilized in education, it is neither designed nor advertised as a solution for real-time transcription and captioning for academic lectures.
Voice recognition technology that could work in the academic environment is on the radar screen of educators and technology companies. A network of researchers and educators formed the Liberated Learning Consortium some 25 years ago to spearhead projects such as the development of an effective, automated speech-to-text system that could be used in classrooms to help people with hearing loss. Based at Saint Mary's University in Halifax, Nova Scotia, the consortium involves researchers in universities around the world, as well as corporations like IBM, in the work of programming speech-recognition software that can provide accurate, real-time captioning of live speech.
When using a program like NaturallySpeaking, the user must say words like "period" to end a sentence, or "new paragraph" to begin a new paragraph. Plus, users can easily correct mistakes as they go, helping the program to "learn" to understand their unique speech patterns. But professors often speak in incomplete sentences or engage in discussion with students, factors which create extra difficulties for software programs to deal with. One solution to this has been to employ human "editors" to read the computer's transcription and make corrections on the fly. Once again, however, this reintroduces the cost factor of paying a person to perform this task.
Clearly, software is not ready to be left on its own in the classroom, but given the advancements in speech recognition technology over the past few years, it seems within reach. In the meantime, speech-to-text services are making education accessibility to students with hearing loss. Another system, Elluminate Live!, utilizes a live provider for the closed-captioning capability combined with software for other features. The program connects on-site and distance learners with visual, hearing, mobility and/or cognitive disabilities with audio and visual lesson elements, making it a very accommodating solution for institutions on a tight budget.
