Content Copyright © 2006 Bloor. All Rights Reserved.
In comparison to talking, typing is:
- Slow: you can talk at up to 160 words per minute, whereas you are lucky if you can type at 40 wpm.
- Inaccurate: words tend to be mistyped or misspelled—further reducing the effective speed. This is especially true for people suffering from dyslexia.
- Injurious: continuous typing is a common cause of repetitive strain injuries (RSI).
- Impossible: various impairments make typing impractical; these include: lack of limbs, Parkinson’s disease, injured arms, or RSI.
So why do we still type rather than dictate to our computers? There are some potentially good reasons:
- Talking aloud in a library may be considered anti-social.
- Physically cannot speak.
- Lack of suitable computer software for the computer or language being used.
- The dictation software is not sufficiently usable or accurate to be beneficial.
With the latest version of Nuance Dragon NaturallySpeaking the last two reasons should no longer apply.
Dragon has been the leader in speech recognition dictation software for some years and Version 9 has been improved so that it should be attractive to most PC users.
The recognition algorithms have been improved. According to Nuance the accuracy has improved by 20% over V8 and 50% over V7; this means that the product should give 99% accuracy straight out of the box.
This level of initial accuracy means there is no longer any requirement to train the program to a particular user’s voice. In earlier versions the user had to read several passages of text—The Alice in Wonderland exercise—and many users found this time consuming and frustrating. The accuracy does still improve over time by recognising when the user has made a correction and incorporating this into its database. This will now be most useful for adding specialist jargon used by the dictator.
Nuance have improved the on-line help so that users can quickly become competent in all aspects of the product including voice commands to control the computer and good dictation practices.
Dragon NaturallySpeaking V9 is section 508 certified (the main US legislation on computer accessibility); this means that V9 can be used by people with disabilities; for example everything can be controlled without needing to use a mouse.
Support has been added for Bluetooth headsets so that the user no longer needs to be tethered to the computer. The problem, until now, has been that the bandwidth was too small and this lead to a low quality audio signal and therefore a lower accuracy rate.
Enterprises can now more easily implement Dragon as it is now available on Citrix thin client or via a network install.
The medical and legal professions have been leaders in the use of dictation software. The methods of working, the specialist language, the need for formal and structured output, and the continued use of dictation and transcription services, have made speech recognition a powerful and productive solution. Nuance provides a special version of Dragon for these two professions with extensive specialist vocabularies.
Bloor would recommend that enterprises should have Dragon available on their network and offer it to any employee who requests it; this would bring multiple benefits:
- They will be able to offer jobs to applicants who cannot use a keyboard, for whatever reason, increasing the pool of talent they can tap into and also reducing the possibility of legal action against them.
- Employees who complain of arm or back problems, which may be caused by use of the computer and could be a precursor to RSI, can use Dragon for some of their work. This should reduce the number of sick-days taken by employees because of arm and back problems. It also reduces the chance of long-term sick leave and litigation surrounding cases of RSI.
- Finally, enterprises will find that certain tasks can be tackled faster and more accurately using dictation rather than typing. Having Dragon available will provide the opportunity for employees to recognise these task themselves and improve the overall productivity of the organisation.
Speech recognition is becoming pervasive with its use in call centres, in-car systems, mobile phones etc. It is now time it became a common tool for dictation in the home and the office and, in particular, for anyone who can not easily use a standard keyboard and mouse.