[il-talk] Fw: [Medianews] Get Ready to Speak to Your Smart Phone

Don Gillmore don.gillmore at gmail.com
Tue Jan 25 13:55:08 UTC 2011


FYI!
----- Original Message ----- 
From: "George Antunes" <geantunes2 at comcast.net>
To: <medianews at etskywarn.net>
Sent: Monday, January 24, 2011 7:19 PM
Subject: [Medianews] Get Ready to Speak to Your Smart Phone


JANUARY 24, 2011, 10:09 A.M. ET

Get Ready to Speak to Your Phone—and Be Understood

By BEN ROONEY
Wall Street Journal

http://online.wsj.com/article/SB10001424052748704881304576093853363154900.html


Twice in three days now I have been demonstrated one of those
technologies that we've been eagerly awaiting, but which has proved
disastrous every time any company has tried it: voice-controlled
computing. Most people's experience of it is those awful automated
telephone systems that can't tell the difference between Aintree and
Braintree. Or else you may have used early versions of programs like
Dragon Naturally Speaking, which was anything but natural.

But voice as input now stands on the verge of becoming mainstream.
Google has made huge strides in this area already. Its Android OS has
speech recognition baked in, and if you live in the U.S. you can have
Google Voice transcriptions of your voice mails—though reports of their
accuracy are mixed. There are companies like Spinvox that have a
speech-to-text service for voice messages. Ford, and other motor
manufacturers, offer speech recognition in modern vehicles.

And of course Dragon Naturally Speaking, now up to version 11, has made
such leaps forward that it is unrecognizable from those early products.

As any smartphone user knows, while typing on your phone is possible, it
just isn't a pleasant experience. Predictive text helps a bit, and there
are any number of apps that offer alternative keyboards, or
enhancements. But even so, we are using an user interface that was
designed more than 100 years ago to prevent typewriter keys from jamming.

Which takes us back to those two companies. The first was Nuance—one of
those really big companies that is responsible for software you use
without knowing who's behind it (does your phone have predictive text?
Could be it's a Nuance product. Nuance owns Dragon Naturally
Speaking.)—and at the other end of the scale, Novauris, a small,
U.K.-based company.

Both are offering speech-recognition software that runs on your smartphone.

Novauris is offering a handy little London guide, Speak&Go London. The
app (for iPhone and Android) helps users get from one station to another
on London public transport simply by speaking their journey into the
phone. So far, so ho-hum. Two things are nice: first, the natural
language input. "How do I get from Paddington to The British Museum?"
spoken utterly naturally, and second—and this is the really nice bit—the
speech-recognition part is done on the phone, so it is incredibly fast.
And this is real speech recognition, not just named dialing.

The only downside is that the routing information is done over the air,
so they are in the hands of Transport for London on that. The speech
recognition done on the phone is pretty impressive. The dictionary of
possible London Underground stations and London landmarks isn't a very
large one, so it is well suited to an app. You couldn't do general
recognition on a phone. Yet.

Nuance, meanwhile, announced it is throwing open its speech-recognition
technology to everyone. This means any developer anywhere can integrate
speech recognition into its app. And speech recognition doesn't have to
mean search. In fact, Nuance is offering three services: dictation
(speech to text), search, and text to speech.

Of course, we have yet to see how good their service is, how expensive
it is and how fast. But opening up every, and any, app to speech has the
potential to transform this technology from curious to ubiquitous and
opens up a range of possibilities.

Search is an obvious one—and already implemented in Amazon's app. But
voice recognition could be added to games, to translation services, even
to really basic things like filling in a form, or updating statuses. The
only limit to its use becomes the ingenuity and creativity of
developers. And that has not been found wanting.

Read more:
http://online.wsj.com/article/SB10001424052748704881304576093853363154900.html#ixzz1C0NcbPwG

-- 
========================================
George Antunes, Political Science Dept
University of Houston; Houston, TX 77204
Voice: 713-743-3923  Fax: 713-743-3927
Mail: antunes at uh dot edu

_______________________________________________
Medianews mailing list
Medianews at etskywarn.net
http://lists.etskywarn.net/mailman/listinfo/medianews 





More information about the IL-Talk mailing list