Can Speech Recognition Find Its Voice in Computing?


[qi:006] Microsoft (s msft) once again is touting its speech-recognition technology, predicting on its site this week that “talking to a computer may soon be as natural as using a mouse.” But while voice is a natural fit for mobile phones and some other platforms, when it comes to traditional computing — using a laptop, desktop or even a netbook — the use-case scenarios for speech recognition are more limited. It will take quite an effort to convince users to talk to their laptops instead of typing on them.

The mouse has nearly become an extra appendage for many of us; touchpads and touchscreens are simple and intuitive; and larger screens minimize the need to drill down through menus. The use of voice as a navigational tool will increase as smartphone usage continues to ramp up, and it will likely be used in specific situations on smaller devices like netbooks — when a user’s hands are otherwise occupied, for instance, or in rare on-the-go scenarios. For traditional computers and laptops, though, I don’t see it gaining much traction among mainstream consumers.

Don’t get me wrong: Speech-recognition software has come a long way in the last few years. Once prohibitively error-prone and inconsistent, the technology now is the foundation of impressive offerings from players such as MacSpeech, Nuance (s nuan) and Vlingo. Microsoft seems especially well-positioned in the space. The company picked up startup Tellme in a 2007 deal reportedly worth more than $800 million, and it is wisely using cloud-based technology to build a system that leverages an enormous amount of data to determine what users are saying.

Picture 3

For $79, get this report and more from GigaOM Pro. Subscribe today!

The Redmond gang has helped push speech recognition into the mainstream with Ford (s f) Sync, the in-dash recognition and search system it powers. And Microsoft is aggressively rolling out voice functionality across its offerings including Bing for Mobile, Exchange Server 2010, Windows 7 and Windows Mobile.

Voice is a natural fit for drivers whose hands are busy and whose eyes are on the road. It also will be absolutely crucial in mobile (GigaOM Pro, sub. required), where small keypads and screens make touch navigation difficult — or, if you’re behind the wheel, dangerous. And consumers are likely to love features like Microsoft’s Voice Mail Preview, which uses speech-to-text to deliver voicemail information to a user’s inbox.

Use of speech-recognition offerings should ramp up in the next few years as phones continue to grow more sophisticated. When it comes to usage on desktops and laptops, though, any real uptake is a long way off — if it happens at all.


Reid Williams

I remember stark disappointment at the performance and reliability of speech recognition software I experimented with 15 years ago in college. That has transformed into wonder at my speech-recognizing Google mobile app. That led to a search for more information about the current state of the technology. Only then did I realize that my Macbook Pro is enabled with speech recognition for certain commands. I played with it for an hour, but I guess old habits die hard because I haven’t used it since and stick to my mouse and keyboard.

I do think, however, that it’s just a matter of acculturation. For younger people who grow up on it, it will become second nature and they’ll laugh about their grandparents clunky, old, manual technology.

Stacey Higginbotham

I actually use speech recognition software in typing up my posts, and think it will have a hard time, not because it isn’t a good interface, but because there’s a steep learning curve in figuring out how to talk to a computer so it knows what you want it to do. For example, when I ask my computer to Open Firefox I need to tell the software whether I want it to actually open Firefox or if I want it to type “open Firefox” or even “open fire fox.” Plus, there are hundreds of commands to master verbally for tasks like cutting and pasting links into a post for example. However, once that stuff is learned, it’s pretty awesome.

Comments are closed.