‘Hello? …Is This Thing On? …’

By Kelmon

I’m pretty much always in awe of many of the blogs where authors manage to come up with new content on a daily basis. Personally, I’m not a big writer of articles (a couple recently, which is pretty rare) and have pretty much exhausted the topics that I could think of. So, when it came to wanting to add some new content to the site and keep it from getting a little stale, I was a little stuck. Still, what’s the best idea in these situations? From the blogs that I read it seems to be to write about your experiences, so here goes…

Last night I started talking to my computer…

For a while now I’ve been phoning home to my parents and friends in the UK via Skype using the Mac OS X client, a pair of walkman headphones and my laptop’s integrated microphone. Anyway, last night I was trying to see if my microphone would work better if I calibrated it and went into the System Preferences looking for anything that might help improve Skype conversations. What I came across, however, was a feature of the OS that I’d not come across before and, while it entertained me for 1/2 an hour, drove my wife nuts.

What I came across was the Speech Recognition software built into OS X. It’s not full speech recognition of the likes of Dragon NaturallySpeaking whereby you can dictate into the likes of Word but the Speech Recognition software will allow you to do stuff like move around in a document, open email, switch applications, get the phone number of a person in your Address Book, etc. To be honest, I found it all to be kinda neat (to the point that I’d use the word “neat” in conversation) but not without flaws.

The principle problem with the software is that it’s not terribly accurate, or it just hates my accent. The little setup screen allows you to calibrate the volume of the microphone and test the software with some test phrases, which all went pretty well. However, when it came time to try and use it quite often it either wouldn’t understand me or, worse, would misunderstand me completely. I must have spent 15-minutes working my way through the “Tell me a joke” command that basically generates a series of Knock, Knock “gags”. The software would understand my “Tell me a joke” request but hardly ever understood me saying “Who’s there” and the other parts of the joke. For those interested, the only joke I did manage to get to the end of was the following:

“Knock, knock”

“Who’s there”

“Thermos”

“Thermos who?”

“Thermos be a better way”

Yeah, no shit, Sherlock…

The software does tell you which commands are understood, which is nice if you’ve lived all your life under a rock and don’t know how a Knock, Knock joke works…

Similar issues were experienced when changing applications. For some odd reason the software found me totally comprehensible when asking it to “Switch to OmniGraffle Professional” but kept interpreting any requests to Open/Change to/Quit iCal as Open/Change to/Quit iChat. I tried faking an American accent to see if that would help things but the only American accent I know is from the likes of Mississippi/Tennessee, so I don’t think that really improved things…

The other principle problem was not with the software itself but because I felt a bit of a tit issuing voice commands to my computer. This is not to say that talking to the computer isn’t an entirely alien concept as I frequently swear at the likes of MS Word when it starts playing silly-buggers. But still, despite the fact that I’d grown-up watching the likes of Star Trek and marveling at how great it was to be able to talk to the ship’s computer rather than using the keyboard on my old Commodore 64, in practice it feels kinda embarrassing. This, perhaps, wasn’t helped by my wife telling me that she’d divorce me if I didn’t stop saying “Who’s there” in assorted styles/accents.

Obviously, it’s easy to see the benefits of such a system for the physically disabled but I’m wondering whether there is ever likely to be a time when I’d prefer to use such an interface instead of the usual keyboard/mouse. I’ve spoken a few times on why, in my opinion, the keyboard/mouse combination is awful for games and have noted that Nintendo’s Revolution controller looks like a step in the right direction, but I’ve not extended this to “normal” human/computer interaction before. While vocal communication might be a while away, if only for the fact that the English language is so phenomenally complicated, perhaps other forms of interaction should be considered. For example, I still think that the “hand as physical pointer” interface that Tom Cruise employs in Minority Report to move “windows” around a screen shows a lot of promise. Yeah, touch-screen styluses go some way to providing this kinda interface but they’re just not as cool…