Improving your diction is one of the curious spin-off effects of using a productivity tool like Dragon. It’s so much better to hear your voice playback and get tips on where your pronunciation could use work from a machine than a someone.
Users are sometimes skeptical about the software’s ability to decipher their accents if English is their second language. The voice recognition programme has released effective versions for Italian, French, German and Dutch. While the Dutch version would signal that an Afrikaans variation is entirely possible, it seems that Nuance hasn’t deemed the market big enough to justify the research and development.
The bottom line is that it’s not as much about understanding your accent as it is about maintaining uniformity when saying words. The software has regional adjustments when creating a voice profile in order to give it the best chance of recognising dialect differences. If you think about the United Kingdom’s dialects, the variations of spoken English between Wales, Ireland, Scotland and England are already enough to contend with, so why wouldn’t it be able to handle anything you throw at it?
All of this is to say, it will misrecognise your voice from time to time, even if you think you speak the purest form of whatever language you consider your mother tongue. Everyone mumbles, slurs their words or runs them into one another depending on how long you’ve been dictating, what you had alongside lunch and whether you should have booked off sick or not. There are many factors affecting our voices, ranging from environment to stress levels, so you’ve got to try and find a consistency in order to give Dragon the best chance of interpreting your soundwaves and converting them to text.
The fact this happens almost instantaneously, depending on your processor speed and accuracy vs speed slider, is no small miracle. There are a lot of factors that can contribute to effective transcription. On a good voice day, you may still get errors and if you’re not part of the problem when it comes to the factors listed above, it could come down to how you say that word or phrase.
It’s one thing to get corrected by a Grammar Nazi or someone condescending enough to make you feel less than. However, it’s a whole different thing to be misheard by machine learning, a software programme and computer that is simply there to absorb and spit out the best possible guess. That came as ‘film’ and ‘form’ for this author with the software struggling to differentiate between the two soundwaves.
Sometimes, if you’re not pronouncing something definitely or clearly enough, Dragon will let you know it needs work. While diction isn’t the main objective, it’s a nice plus and better to hear from your blank-faced computer than someone who’s been meaning to say something.