The 2011 Market Leaders

Article Featured Image

As an interface of choice, speech is being felt—and heard—in nearly every market segment, most notably in mobile phones and cars. Months of research have been poured into the following pages, as we honor the winners, leaders, and Vendor Contenders in six categories. In a nutshell: Acquisitions have winnowed the speech engine field, with the big blow coming when Nuance acquired IBM’s speech portfolio, removing Big Blue as a player. A tough economy has squeezed sales of speech self-service, but signs point to a bullish future. Soaring growth marks speech analytics, as companies scramble for actionable results. Demand for voice security has this sector poised to expand, while the ubiquity of smartphones spells huge potential for mobile voice search. And outsourcing is fueling professional services. One thing’s for sure: In a hotly contested market, any company that made our leaderboard has earned its spot.


The Market
Consolidations and acquisitions have dramatically winnowed the speech engine field. Most significant was Nuance’s acquisition of IBM’s speech portfolio in 2009, which effectively removed IBM from the market. This year, the relationship between the two has grown, as Nuance markets IBM’s Watson—a powerful natural language artificial intelligence technology you may have seen on Jeopardy—in the medical field. The alliance has made Nuance, already a big company, a force to be reckoned with in speech engines and speech technology.

This year, perhaps because consolidation has come with higher pricing, cost has had a big impact on the engine category. Notably absent is Microsoft’s Tellme engine, which won last year. Judges this year felt that Microsoft had not done enough innovating. Moreover, many voiced complaints that its customer service left much to be desired. While Microsoft received high scores in cost (tied only with Novauris) and decent scores in accuracy, neither was enough to put it on the leaderboard.

The Leaders

Storied AT&T has been involved in speech technology since the company was but an aspirational gleam in the eyes of starry-eyed computer scientists. Since then, AT&T has been committed to research and development, which is a big reason the company is among the leaders this year. AT&T boasted a high innovation score, a great ability to customize and implement, and very decent recognition scores.

As for cost, our analysts and consultants were all over the map, giving a mix of high and middling scores. The reason may be the company’s failure to price its recognizer. “It’s free to developers,” says Bill Scholz, a consultant and the president of New Speech. “But they haven’t figured out yet how to price it to people who are deploying it. It’s as though it’s free, though it’s clearly not AT&T’s intent to give it away.” Because of this ambiguity, Scholz says, some are wary of using it for fear of getting slapped with a hefty price tag down the road.

Also on the leaderboard was Nuance Communications. It’s more or less unanimous: Nuance has one of the highest-quality recognition engines on the market. As in years past, Nuance returned perfect scores from every analyst and consultant in our panel this year—an impressive feat. It also did well in customer satisfaction with a solid 4/5 average. In fact, it did very well in most categories. Where its score suffered was cost.

“They’re ridiculously overpriced,” one consultant says. “A few companies…have backed off from using them because they’re too damned expensive—even though they’re good.”

The Winner
Loquendo reclaims the top prize for the third time in this category. Our consultants and analysts gave the company high marks in customization options, accuracy, and customer satisfaction. Particularly strong is Loquendo’s text-to-speech (TTS) engine, which Scholz describes as “the absolute tops.” He and many of our judges were particularly impressed by the recent addition of vocal effect—conveying emotions through vocal inflections—which is relatively new to TTS technology.

Improvements like that one have earned Loquendo a reputation as forward-thinking. Loquendo also has been pushing harder into the automated speech recognition (ASR) marketplace with its engine. This year’s refresh of Loquendo ASR, version 7.9, now includes Arabic, bringing the number of supported languages and dialects to 28. Loquendo also launched a Romanian TTS voice this year, bringing its stable voice fonts to 30.

“Loquendo is becoming more aggressive,” one analyst says. “It’s more of a competitor to Nuance, and Nuance really owns the market.”

Vendor Contender
Our Vendor Contender this year is LumenVox, which claims to be “the most affordable” in the industry. Though such boasts may be dubious—every vendor self-describes as “the leading provider”—LumenVox is, at the very least, one of the most inexpensive ones out there. With “lite” one-port packages beginning at $195, it’s hard to argue on price. Overall, LumenVox sported solid scores across the board—mostly fours with a slight dip in recognition. Where it really shined, of course, was in cost.
—Eric Barkin


SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

Angel Unveils Lexee

Self-service solution enables voice activation on all mobile apps.