Well, this is going to be interesting. This is the first blog post that I’m dictating into a piece of speech recognition software. It’s proving to be quite a long and drawn-out affair. Every time I make a mistake I have to go back and correct it. However, every time that happens, the piece of software learns the mistake and is less likely to repeat it in future.
One of the things I like about the idea of speaking into a microphone instead of typing into a keyboard is that it lets me pace up and down while I do it. I quite like pacing up and down because it helps me think.
It also means that I’ll be able to record my lectures and conversations into documents as I speak. I’ve always thought that an easy way to write a book would be to speak into something that would just write down what you said, and then you could just go back and edit it later. Mind you, I’d have to get a lot better at ordering my thoughts. Nobody wants to read my verbal rambling in a book.
Anyway. This is all good practice for the software to get to know my voice a bit better. Ideally, it would be great to just be able to say what I think and then send that as a blog post to the Internet. However, as it stands, this is taking me a lot longer than it would normally take to just type something of this length.
I suspect one of the problems is that I need to learn how to speak more clearly and in less of a New Zealand accent. Mind you, I do have it set to international English, so it should be able to cope. Perhaps it’s just a matter of learning my quirks.
In the meantime, I’ll just keep practising. And if I speak to you in a strange newsreader kind of way, you’ll understand that it’s because I’ve got into the habit. I have to say though, I am very impressed with what it can do so far. In the writing of this blog post, my hands did not touch the keyboard once.
I don’t know if you noticed though, but so far my dictation has left my language a little bit stilted and unnatural. The bridge between spoken language and written language is quite far in my head, so when I try and speak in written language, it comes out as if it’s something I’m just learning to do. Which it is. So I suspect that will get better in time too.
Let’s hope so.

Hi. Has the voice recognition software adjusted to your accent yet? Or are you at least having more success now? I’m considering purchasing the same software and like you I’ve also got a New Zealand accent (tamed down after spending a few years out of the country.) I’m wondering whether the UK or Australian versions would be better.
It’s pretty good. I still have to correct it from time to time, but that’s only when I’m not speaking clearly. I know my accent’s the problem when I think I say ‘pin’ but it comes out ‘pun’ – but generally speaking, as long as you’re giving it a whole string of words rather than word at a time, it gets the meaning from the context pretty well.