IBM Gives a Mobile Voice to Developers


Imagine not having to go through the machinations to hear a voice mail
that you likely don’t need to hear, or not having to scroll for a contact off a smartphone directory listing. Instead you just give a short voice command to read a text version of the voice message or the contact name and get it in just seconds.

That’s the beauty speech recognition holds for mobile devices within the
next year, and such innovative capabilities are what IBM (NYSE: IBM)
believes could be among the biggest technology transformation to date.

It’s also why Big Blue announced today it is offering its speech
recognition engine to clients and partners, developing both consumer and
enterprise applications for smartphones and mobile Internet devices.

By providing it for collaborative development, IBM said it was taking a
new strategy that allows developers to work directly with IBM Research to
augment, improve and build new products with the technology as the base.

“It’s an evolving strategy for our research division, rather than trying
to push our technology mainly through the product and services divisions,”
according to a press statement.

The speech research group is also focusing on forming partnerships to take the technology to market. These partnerships will be licensing or co-development agreements.

“From unified messaging to analytics, the mobile Internet will be a huge
IT transformation driven by end user needs and good user interfaces,” David
Nahamoo, CTO, speech technology, IBM Research, told

Given that form factor and device hardware features, such as touch screens, aren’t going to be compelling purchase factors down the road, experts say the mobile Internet will involve applications such as speech recognition.

In taking a collaborative approach, IBM hopes to spur advanced
applications that make it easier to do everything from online search to
enhanced data services.

“The value of speech for mobile workers will really expand now that the
infrastructure is in place. It [speech recognition] just goes hand in hand
with mobility and will have a big impact in the coming years,” added
Nahamoo, who leads IBM’s Super Human Speech initiative that began back in

Right now there’s already speech recognition taking place on handsets.
One is Yahoo’s (NASDAQ:YHOO) oneSearch application launched last year, which provides users with a Yellow Pages search functionality.

That application, which licenses IBM’s speech recognition engine, is just
the start of the mobile Internet transformation, Nahamoo said.

“It’s not going to take more than a year, as many devices already have
some embedded voice technology already,” Nahamoo said. “The market is happening.”

Speech recognition, by all accounts, is one of those technologies that’s
been worked on and advancing for decades. That’s because the challenges have
been many. One, the issue of the background noise, is still a hurdle, though
other issues, such as grammar and punctuation, have undergone great
improvements, according to IBM.

“The speech models used to develop the speech recognition engine have
improved considerably, thus increasing the accuracy of the recognition,”
explained Nahamoo, noting that IBM speech software now lets users
interact with car navigation and entertainment systems and is helping to
improve customer service efforts.

One analyst said speech will soon equal text as a mobile application.

“Improved performance and tools for interpreting the resulting text are
driving rapid growth, for example, of speech as an interface to mobile
phones,” said Bill Meisel an analyst at research firm TMA Associates, in a press statement.

Meisel pointed to the Yahoo search model as a prime example of how speech
will alter mobile device capabilities.

“Just say what you want and get it,” he said. “The transcription of
voice mail and call-center conversations to text makes it possible to quickly
search and analyze the result, creating new growth and flexibility in
unified communications and customer-service analytics,” he added.

News Around the Web