Monday, 12 August 2019

Speech Recognition


Before my blogging was rudely interrupted by falling down the stairs, I was working my way through a book called A Brief History of the Future by Charles Joynson. I was looking at the developments he predicted and seeing where we have got to today.

He says that by 2093 speech recognition
was far quicker and more accurate than ever before. This meant that the very last keyboards were recycled and speech became the standard way to communicate with computers. Throat microphones also became common and more sensitive, which meant that they could detect and understand speech without it being audible to others.

Dictation

As early as Star Trek, the crew of the Enterprise talked to their computers, but that was fiction. What is the reality today? The biggest problem with speech recognition is the fact that we all speak differently. My speech can be slurred because I had a stroke. Others have accents of all kinds.

I have used software called Dragon Naturally Speaking for some time, but it required me to read out pieces of text so that it could get used to my voice, and it still makes mistakes. I wonder whether dictation saves me time over typing because I have to go back and correct it.

Already Here

I found a good article on the Cheeky Monkey IT Support website about the future of speech recognition. We already have voice activated kit like Apple’s Siri, Amazon’s Echo and Alexa which are used to control things, but for dictation you need more sophisticated software.

An article on the PCWorld website reviews the 5 top speech recognition software, and Dragon Naturally Speaking comes top, but it’s also very expensive. I was surprised to find out that the latest versions of Microsoft Windows has speech recognition built in. You can turn it on via Control Panel Ease of Access. I’ll be checking that out. Google Chrome also has voice input capability, and so do most phones.

The Future

As for the future, the Cheeky Monkey article suggests it could be used in shops, where the customer can ask if their size is in stock without having to find a member of staff. Another benefit would be in the police, where statements could be taken directly instead of having to type them up, and the same with reports.

I think speech recognition will be in full use a lot sooner than 2093.


Ann Marie Thomas is the author of four medieval history books, a surprisingly cheerful poetry collection about her 2010 stroke, and the science fiction series Flight of the Kestrel. Book one, Intruders, and book two Alien Secrets, are out now. Follow her at http://eepurl.com/bbOsyz

No comments:

Post a Comment