This is how fast BASIC can be made to speak...

Wednesday 19th May

I suppose I should explain Sundays post. I have been dabbling for a while with the idea of using voice recognition software. Unfortunately most voice recognition software packages cost a few hundred pounds. So I wasn't going to go out of my way to pick one up.

Then came last week, where I find myself taking more boxes from my dad's house. In one of these boxes was some old software I had ignored for a few years including the voice recognition package called ViaVoice 98 which run under windows 95. A bit old, but how much can voice recognition software change in that time?

Still, with high hopes, and not at all sense of foreboding I installed it on my current machine and got down to training it to recognise my voice. I went to the basic training, which included 100 lines of dialogue. I did the medium training, which included another 150 lines of dialogue. In fact, I did the entire training which was a grand total of 476 lines of dialogue.

It only took me two hours.

Then I told the machine to learn my voice, a process that would take approximately 2 to 3 hours. Luckily I watched for the first few minutes, because the system failed pretty much instantly. I pressed it again. And again. And again. And again.

Windows vista doesn't like ViaVoice '98.

After 2 hours I had to give up on the training. But strangely enough the programme would still work without being trained. So I ran it, leading me to dictate what you ended up reading on Sunday.

Not to be too deterred, I searched the web for free voice recognition software because there must be some out there. Something like the fifth link on Google was a link to a forum where one user had posted that dragon was much better than the free voice recognition software on windows vista. My windows already has a voice recognition system in it?

I nearly punched the wall.

So since then I've been trying to teach my machine properly in up to date software how to recognise my voice. It is slow and agonising process, but we'll get there in the end.

For example, I just dictated this entire post using this system. And some words didn't need to be corrected at all.