September 30, 2007

Speech Recognition for Automation

Filed under: Scripting — Marcus Tettmar @ 4:52 pm

The last time I tried speech recognition was about 15 years ago and back then it wasn’t very good. I suppose it was early days for the technology but I pretty much gave up on it. This morning, however, I was reading Scott Hanselman’s post about speech recognition in Vista and wanted to give it a go myself. It just so happened that I was in PC world today so I picked up a decent microphone, which I have been meaning to get for a while anyway – for product demos and Skype etc.

So when I got back I hooked up the microphone to my vista box and started playing with speech recognition. My goodness it is pretty amazing! In fact I am writing this post using dictation.

One really cool thing is that I can start Macro Scheduler macros using voice recognition by speaking the keyboard shortcuts assigned to them. This could be really handy in situations where voice recognition is needed for controlling complicated processes. E.g. Hands-busy medical applications, or for helping people with disabilities. Macros can be written to control those processes and assigned to shortcuts which can be triggered by voice commands. And all this is capable using Windows Vista without extra software (except Macro Scheduler of course).

QuickButtons makes it easier because you just have to say the text of a button rather than the shortcut key sequence. With QuickButtons you can create an AppBar with buttons assigned to keystroke macros, applications or Macro Scheduler macros. The AppBar can remain visible above or below the application you are working in. Just speak the caption of a button and that button is “clicked” and the macro is executed.

So not only have I learnt how cool Vista’s speech recognition is today, I’ve also discovered how it can be used with our very own Macro Scheduler and QuickButtons to great effect. Dedicated voice recognition software can cost thousands. Yet with Vista and Macro Scheduler and/or QuickButtons you can make voice recognition do anything for far less. Awesome.