16. Appendix C: Using speech recognition to assist teaching and learning

The capabilities of speech recognition technologies have greatly improved in speed and accuracy of response over the past few years and current applications are relatively inexpensive and capable of very accurate and fast responses on standard computers for normal rates of speech with minimal training of the system to the speaker’s voice or training the user of the system.

Teachers and students have preferences regarding whether and when they find the spoken or written forms of language easier or more useful for teaching and learning. Text to speech applications can automatically change text into speech while speech recognition technology can be used to automatically change speech into text. Speech recognition can be used by academic staff to produce teaching materials and by students for producing coursework and notes.  Speech-recognition can also be used to replace the keyboard for those with a physical disability or to help prevent, or support those who already have, repetitive strain injury (RSI) or work-related upper limb disorders (WRULD).

Speech-recognition can also be used for providing real time text transcription in lectures to provide a text display of what is being spoken as well as a verbatim transcript for later reference. To achieve a similar result without the technology would involve the use of expensive, highly trained real-time speech to text reporters who are in great demand for court reporting and real time subtitling of television programs. Since standard speech recognition applications require the user to dictate punctuation, the Liberated Learning Project (www.liberatedlearning.com) is working with IBM to develop a special speech recognition application that can provide a readable display from the normal speech of lecturers without requiring the dictation of punctuation. Dr Wald is a member of the Liberated Learning Project executive advisory committee and chair of the research and development committee and the application is currently being used at Universities in Canada, Australia, USA and UK. Real time speech to text transcription can assist deaf or hearing impaired students who find it difficult to follow the lecturer through hearing alone as well as students or lecturers whose first language is not English and can also be of benefit when there are poor acoustics (e.g. excessive reverberation or background noise or too quiet speech). The automatic production of a verbatim transcript for later reference can assist students who find it difficult to take notes during a lecture, for example dyslexic students, hearing impaired students who need to watch to lipread or follow a sign-language interpreter, visually impaired students, or students who have a physical disability affecting writing or typing. In addition many students who have no disability or learning difficulty find it difficult to take notes at the same time as listening, watching and thinking. Speech-recognition can also be used to support distance learning by providing automatic speech to text transcription for on-line text chat, Email and video or audio-conferencing.