A Quick Guide To Speech Recognition Software

Speech Recognition Software

As technology progresses, speech recognition software is improving and with apps like Siri for the iPhone, it is becoming both cheaper and more widely available.  With this in mind, we thought we’d write a quick guide to the pros and cons of using speech recognition software, including what it is and isn’t suitable for.

First off, it is important to realise that speech recognition software is pretty much useless for more than one speaker, and so is totally unsuitable for one to one interview transcription, focus group transcription and conference transcription, always use a transcription service for this type of transcription.

Dictation.  The truth is, speech recognition software is not as accurate as a human transcriptionist (and is unlikely to be for a good long while yet). However, with higher end software such as Dragon Naturally Speaking, you are able to ‘train’ your software to recognize your voice by proofreading your work and then updating your software.  The problem is, however, that if you are using speech recognition for anything other than personal notes or emails, you will always have to proofread what has been written thoroughly before using it – especially if you are either a medical or legal professional. If you don’t do this, you will likely find that some of your corrections just never sink in.

Proofreading. Proofreading a document written using speech recognition software means having a keen eye for mistakes, as unlike human transcriptionists, errors made by computer software will not be simple spelling mistakes, but are often words that are totally out of context and do not belong in your documents.

Background Noise.  It is especially important to reduce background noise when using speech recognition software as it is prone to picking up background noise and interpreting it as words.  One of the best ways around this is (obviously) to record in a quiet environment or to buy a noise cancelling microphone.

Layout. Although some software does allow you to build templates, this can be complicated and does not always work according to plan.

Speed.  It is a good idea to speak slowly and pronounce your words as correctly as possible.  It’s also worth thinking about what you want to say in advance in order to avoid having to edit out any ‘er’s’, ‘um’s’ or redundant speech.

Technical Assistance.  If you use speech recognition software for work, it is extremely important that you choose a company that offers great customer care and technical assistance in case something goes wrong; otherwise there is a good chance you’ll end up with a backlog of work with a long wait to fix it.

We think that speech recognition software still has a way to go before it can compete with a human transcriptionist and therefore we strongly recommend that you ensure that your documents are always thoroughly proofread before any kind of professional use or distribution.

If you would like more information about our proofreading, editing or transcription services, why not contact us today?

Here is an example of a simple dictation recorded using speech recognition technology (including tracked changes):

Proofread Speech Recognition Document (with tracked changes)