Jan 23, 2015 - Subscribe. World politics Business & finance Economics Science & technology Culture Blogs Debate Multimedia Print edition. All latest updates.
2/5/2015
Speech recognition: Watch what you say | The Economist
More from The Economist
My Subscription
World politics
Business & finance
Subscribe
Economics
Science & technology
Culture
Speech recognition
Watch what you say
All latest updates
Log in or register
Blogs
Debate
Multimedia
Print edition
Comment (11)
Timekeeper reading list
Email
Reprints & permissions
Print
Better automated acquisition of speech may be more about seeing than hearing Jan 23rd 2015 | SEATTLE | Science and technology
Follow The Economist
Latest updates » The Silk Road trial: Tales from the crypt Democracy in America | Feb 4th, 23:53
The Economist explains: How disease forecasts can go wrong The Economist explains | Feb 4th, 23:50
Tracking aeroplanes: Eyes in the sky Gulliver | Feb 4th, 21:51
"IF HE were proven to be malfunctioning, I wouldn't see how we'd have any choice but disconnection." In the film "2001", Frank Poole, an astronaut played by Gary Lockwood, considers what should be done with HAL, the homicidal computer in charge of the ship. HAL
Ratings agencies and the financial crisis: A fine too far
learns of his human masters’ plan to unplug him by lipreading their conversation through a
Business and finance | Feb 4th, 20:54
window—a strategy that several researchers and companies are getting closer to realising. Their goal is less about spaceshipdriving robots and more about improving the
Ukraine's new bailout: Nothing yet
performance of voicecontrolled helpers such as Apple’s Siri and Microsoft’s Cortana.
Free exchange | Feb 4th, 20:22
No matter how good voicerecognition software becomes, it will always be hostage to its sonic environment. Ask your digital assistant to dial a number in a quiet office and it might
Network neutrality: No more overtaking
hear the right numbers. Try again near a busy road or at a noisy party and you will probably
Business and finance | Feb 4th, 20:20
be disappointed. If only your phone could simply read your lips. Ahmad Hassanat, an artificialintelligence researcher at Mu’tah University, in Jordan, has
Portrait of Italy: Diversity, not disunity
been trying to teach a computer program to do just that. Previous attempts to get computers
Culture | Feb 4th, 18:50
to lipread have focused, understandably enough, on the shape and movement of the lips as they produce phonemes (individual sounds like "b", "ng" or "th"). Such shapesofsounds
More latest updates »
are called visemes. The problem is that there are just a dozen visemes for the 40 to 50 phonemes in English; "pan" and "banned", for example, look remarkably similar to a lip reader. That makes it rather taxing to reconstruct words from visemes alone. Instead, Dr Hassanat has been trying for the past few years to detect the visual signature of entire
Most commented
words all at once, using the appearance of the tongue and teeth as well the lips.
1
Russia and Ukraine Understanding Putin’s plans
His method has had some success. In a paper published late last year, Dr Hassanat trained his system by filming 10 women and 15 men of different ethnicities as they read passages of text. The computer first compared these recordings to a text it knew, then tried to guess what they were saying in a second video. When the computer was allowed to use the same person’s training speech, it was fairly accurate—around 75% for all subjects and up to 97%
2
Greece and the euro’s future: Go ahead, Angela, make my day
http://www.economist.com/news/scienceandtechnology/21640472betterautomatedacquisitionspeechmaybemoreaboutseeinghearingwatchwhatyou
1/3
2/5/2015
Speech recognition: Watch what you say | The Economist
for one speaker. But when the person’s own training video was excluded from the analysis —analogously to similarly untrained digital assistants—the program's accuracy plunged to 33% on average and as poor as 15% in some cases (moustaches and beards, it seems, are particularly confusing to the system).
3
Greece's debtreduction offensive: Threats and blandishments
4 5
War in Ukraine: No exit Greece and the euro crisis: An odd view of democracy
Another idea is not to focus on the mouth. In 2013, Yasuhiro Oikawa, an engineer at Waseda University in Japan, used a highspeed camera capable of shooting 10,000 frames a second of a speaker’s throat. The approach measures tiny, fleeting vibrations in the skin caused by the very act of speaking. The precise frequencies present in the vibrations can then, in principle, be used to reconstruct the word being spoken. So far, however, Dr Oikawa’s team has managed to map the visual vibrations of just a single Japanese word. The best results seem to come when the approach is used at closer quarters. VocalZoom is
Products and events Have you listened to The Economist Radio on Facebook? The Economist Radio is an ondemand social listening platform that allows you to listen, share and recommend The Economist audio content
an Israeli startup whose idea is to point a lowpower laser beam at a speaker’s cheek to measure vibrations, and use those to infer the frequencies of speech. The system combines those results with ordinary speech audio from a microphone, subtracting unwanted ambient noise or other talkers and leaving just the cheekwobble frequencies. Earlier this month, the firm took its technology to CES, a big trade show and a notoriously earsplitting environment, and impressed the tech press. But it is not yet ready for the mass market. The prototype system is currently larger than the smartphones it is intended to be
Test your EQ Take our weekly news quiz to stay on top of the headlines In Other Words Try our new audio app and website, providing reporting and analysis from our correspondents around the world every weekday
built into, and tempting manufacturers into adding components to everslimmer, eversleeker handsets will not be easy. The company may have more luck getting its technology into cars, another industry increasingly reliant on voice control; VocalZoom claims to be in early talks with a big carmaker. Perhaps the company will even get its kit into spacefaring vehicles.
View all comments (11)
Want more from The Economist? Visit The Economist estore and you’ll find a range of carefully selected products for business and pleasure, Economist books and diaries, and much more
Add your comment
More from The Economist
The Economist explains: Why the oil price is falling
Human mating strategies: Cads and dads
Game theory in American football: Defending the…
Russia and Ukraine: Understanding Putin’s plans
The Economist explains: Why currency volatility has got worse
America’s elite: An hereditary meritocracy
Living in Detroit: Surprisingly expensive
Canada’s economy: Beyond petroleum
Materials science: Wings of steel
Want more? Subscribe to The Economist and get the week's most relevant news and analysis.
http://www.economist.com/news/scienceandtechnology/21640472betterautomatedacquisitionspeechmaybemoreaboutseeinghearingwatchwhatyou
2/3
2/5/2015
Speech recognition: Watch what you say | The Economist
Contact us Help My account Subscribe Print edition Digital editions
Sections
Blogs
Research and insights
United States Britain Europe China Asia Americas Middle East & Africa International Business & finance Economics Markets & data Science & technology Special reports Culture Multimedia library
Buttonwood's notebook Democracy in America Erasmus Free exchange Game theory Graphic detail Gulliver Prospero The Economist explains
Topics Economics AZ Style guide The World in 2015 Which MBA? The Economist GMAT Tutor Reprints and permissions
Debate and discussion
Events
The Economist debates What the world thinks Letters to the editor The Economist Quiz
Jobs.Economist.com Timekeeper saved articles
Contact us
Help
About us
Advertise with us
The Economist Group » The Economist Intelligence Unit The Economist Intelligence Unit Store The Economist Corporate Network Ideas People Media Intelligent Life Roll Call CQ EuroFinance The Economist Store
View complete site index »
Editorial Staff
Staff Books
Copyright © The Economist Newspaper Limited 2015. All rights reserved.
Careers
Accessibility
Site index Privacy policy
Cookies info
Terms of use
http://www.economist.com/news/scienceandtechnology/21640472betterautomatedacquisitionspeechmaybemoreaboutseeinghearingwatchwhatyou
3/3