Stray Dog Strut

I’ve been playing about with the Speech capabilities of my Mac – kinda relevant considering my last post. Remember I told you before about playing Chess using voice commands?lol Well, it turns out you can do so much more.

By turning Speakable Items on, and as a personal preference I’ve set my Mac to always listen (rather than holding a key when you speak as before), I can now control a variety of applications and carry out tasks using nothing but the beautiful sound of my own vocal chords, lol.

You can assign new commands to applications or webpages simply by saying “make this speakable”, or you can delve deeper by using AppleScript to make complex vocal commands for say, Photoshop shortcuts. I even read an example online of making a script to have the computer respond if you say a phrase such as “hello computer” – like a verbal version of the well-known “Hello World” program.

As my Mac is now always listening, I’ve got it set to carry out an instruction only when it hears a keyword. I’ve left this on the default ‘Computer’ but saying things like “Computer open Safari” sounds a bit daft, maybe I should call him ‘Mac’?lol

It’s not perfect – everytime I tried to open Pages it opened Stickies instead. But saying “Computer open Mail, Computer open my blog, Computer switch to Mail” and so on, and seeing it all happen is pretty cool!lol

I was reading one of Falling’s old emails today – sort of a nostalgia trip – and she was talking about how difficult it can be for deaf people in lectures and discussions at university because, while having a note-taker is been helpful, it means the person can’t join in with the conversation as by the time they’ve read the notes, the conversation has moved on.

Anyway, at one point she said “You know, it would be so great if there was this voice decoder thing when people spoke into it and it shows text for what they are talking about. I somehow think that’s really far into the future though…”

Now that got me thinking – surely if this is the so-called Digital Age and if we already have voice-recognition software for our computers for dictating letters etc, surely it can’t be that difficult to expand that software to a situation as described above? Ive done some digging around – and unfortunately I could not find a consumer solution available as yet – but I did come across some interesting reading.

In 1998, St Mary’s University in Nova Scotia conducted a pilot project of an automated captioning system called Liberated Learning. This system enabled a lecturer’s words to appear on a screen as they speak – the lecturer would typically wear a small microphone to pick up his/her speech. After the lecture, it also provides a transcript that students can use.

In 2001, St Mary’s asked for the help of IBM in improving the system further. The system came to be known as ViaScribe (as it utilises components of their ViaVoice technology). As far as I can tell, this system is still under R&D by the Liberated Learning Consortium and seems available only to institutions directly. Disapointingly, out of all the institutions involved in the project, the majority are American/Canadian, with the University of Southampton the only UK partner. Hopefully though, with enough publicity, this will be rolled out to our universities in my lifetime!

Another interesting project I came across was Sphinx. Sphinx is a speech-recognition system written in the Java programming language. from the documentation, it appears to be more of a research tool, rather than an end-user product, and although I did some Java at university, it’s all a bit over my head. However, I beleive Java is a very flexible platform and with many Java projects being open source, I think we should see real development in projects like this.

Well, that’s my two cents…

Edinburgh is not shown as the capital city of Scotland on Google Earth – despite London, Douglas and Reykjavik etc all being shown (Capital of England, Isle of Man and Iceland respectively).

We are not amused.