There are four main aspects to consider when trying to get the best results from your speech-recognition software:

  • the environment,
  • the equipment (see ‘system requirements’),
  • knowledge of how the software works (see above),  and
  • awareness of its limitations.

Firstly, it is best to work in a fairly quiet environment. However it is possible to use speech recognition software in an open plan office environment—and many people do so successfully—as long as noise levels are not excessively high, and a high-quality noise-cancelling microphone is used (see recommended microphones below).

It is important to be aware of the software’s limitations—having realistic expectations will help you to deal with the frustrations of working by voice. For people who need to limit their physical use of the computer, be aware that it is generally not possible to do all computing tasks by voice. For example, logging on must be done manually as the speech recognition software cannot be used until you are logged on to the computer. The audio check, which should be done on a regular basis, cannot be completed entirely by voice.

There are also some tasks that are tedious and inefficient to do by voice, for example, spatial tasks in graphics applications.

It can be challenging for new users to speak whilst thinking about what they want to write next, particularly when dictating complex documents. People differ in the ease with which they can adapt to speech-operated computing—it is often more difficult for people who were very competent typists and are now unable to use their hands for computing.

Patience is definitely necessary when using speech-recognition software, as correcting recognition errors can sometimes be a stop-start process that interrupts your flow of thought, particularly for new users.

Unlike keyboarding, system error can occur up to 5% of the time. In other words, even if you do everything right, the software will still occasionally misrecognise what you say. Proofreading is also different for documents written by dictation. As there will be no spelling errors, some people find it is necessary to proofread more actively, reading for meaning rather than checking for words that look misspelt.

How to speak

The quality of sound input is obviously important in getting the best out of speech- operated computing. While the microphone and soundcard are a large determinant of this, speaking style is also a factor. Some new users find it challenging to exercise the necessary control over their speech—avoiding extraneous sounds such as ums and ahs, thinking about what you want to say before you speak, remembering to turn the microphone off when speaking to oneself etc. Turning the microphone off when not dictating is necessary to prevent the microphone from picking up background noise.

Speaking clearly and naturally is important: in fact, the software will recognise continuous speech better than individual words (this is particularly true of the most recent version 9). However, running words together or speaking in an informal, conversational style will lower recognition accuracy. When voicing commands, it is necessary to pause before the command and then say it as a phrase (without pausing midway through). Performing an audio check every day is recommended to ensure sound quality remains high.

Remember also that the software generally listens for words rather than letters, so it is necessary to instruct the software that you are spelling by saying e.g. “type c”. If you say “c” it is likely to be interpreted as the word ‘see’ or ‘sea’ rather than the letter ‘c’. People are often surprised that more technical and complicated words are often recognised better than simple words— this is because many small, common words like “for”, “all”, “or”, “more” sound similar. Technical words are usually more distinctive sounding and longer, making recognition easier.

While you’re here…

While you’re here, how about helping us out with a donation?

We’re a really small organisation doing a really big job. We give people with RSI across Australia the info they need to get the right treatments, navigate the worker’s compensation system and better manage daily life.
Every little bit helps – so make a donation here to help us as Australia’s only RSI support organisation.

Skip to content