Saturday, January 7, 2017

"CES 2017: The Year of Voice Recognition"

From IEEE Spectrum:
The beeping, flashing, pulsating glory of the world’s largest consumer electronics trade show has returned to Las Vegas. The first batch of new products and services went on display at CES on Tuesday, and startups and industry giants will debut more gadgets and technologies throughout the week.

Just a few of the curious wares spotted by IEEE Spectrum editors last night include a battery-powered scarf that filters air pollution, a hairbrush that uses sound waves to analyze dryness and frizz, a smart cane that detects falls, and a connected cat feeder that avoids overfeeding by recognizing felines by implanted microchips. Also, a US $120 camera that lets you stare at the inside of your refrigerator, should you ever choose to do that (assuming the milk isn’t blocking the view).

Major technology companies have also begun to make their announcements about new products they will launch in 2017. Qualcomm released its newest chip, the Snapdragon 835, which, rumor has it, could turn up in Samsung Galaxy 8 smartphones later this year. Huawei said its newest Honor smartphone, called the 6X, which boasts a battery life of 2.1 days and costs only $250, is now available in the United States. And Faraday Future unveiled its long-awaited self-parking FF 91 electric car, which integrates more than 30 sensors including cameras and a retractable lidar system to navigate into a parking space all on its own.

Looking at deeper trends, several experts said the most meaningful long-term developments will come from the companies scraping away at voice recognition. Once we master it, they believe, voice-recognition capabilities will fundamentally change the way we interact with and build electronics.
This was a strong element of Tuesday’s analysis of the global consumer market by Shawn DuBravac, chief economist, and Steve Koenig, senior director for market research, of the Consumer Technology Association (CTA), which runs CES.  In DuBravac’s opinion, voice-recognition technology has improved enough in the past few years that it is now poised to usher in an era of so-called faceless computing.

In particular, the word error rate for voice-recognition systems dropped from 43 percent in 1995 to just 6.3 percent this year, and is now on par with humans. “We have seen more progress in this technology in the last 30 months than we saw in the first 30 years,” DuBravac said. Another analyst attending CES that I spoke to was Ronan de Renesse, a consumer technology analyst for the business intelligence firm Ovum, who said he was watching a startup called Voicebox, which has worked on voice recognition for partners including Samsung, AT&T, and Toyota.

In addition to redefining the traditional computer interfaces, voice recognition could improve a host of products that are already on the market. CTA estimates total sales of voice-activated digital assistants such as Google Home or Amazon Echo to be about 5 million units to date, and expects that to double to 10 million in 2017. With all of these products, clarity and functionality are key. DuBravac figures there are currently about 1,500 apps (called “skills” in Amazon-speak) that can interact with Alexa, Amazon’s voice-activated personality and says he would not be surprised to see 700 new ones announced just this year at CES....MORE