Ford's SYNC learns 100x more voice commands, integrates Nuance technology
By Darren Murph 
posted July 15th 2010 10:10PM
posted July 15th 2010 10:10PM
Ford and Nuance Advance Voice Recognition of SYNC: Now Faster, Friendlier, More Personal
DEARBORN, Mich., July 15 /PRNewswire-FirstCall/ --
* With the introduction of MyFord Touch™ driver connect technology, Ford makes it easier to control in-car systems with fewer steps and more natural language; customers can now speak more than 10,000 first-level commands, up from only 100 in first-generation SYNC®
* Working with voice control leader Nuance, SYNC will recognize more direct voice commands such as "Call John Smith," "Find ice cream" and "Add a phone," allowing users to do more with fewer steps
* Innovative features boost recognition accuracy and provide "Samantha," the voice of SYNC, with smoother, more natural speech patterns
* Consumer acceptance of voice control is increasing; the Harris Interactive® 2010 AutoTECHCAST survey found an 8 point year-over-year improvement, and industry analysts predict continued segment growth
Ford made in-car voice activation a reality for millions of drivers with SYNC, first introduced in 2007. Now, Ford engineers – working with voice technology pioneers Nuance Communications (Nasdaq: NUAN) – plan to once again raise the bar with the next generation of SYNC, a system that can understand 100 times more commands than the original, thus delivering a more conversational experience between car and driver.
The voice upgrades will be available on the next generation of SYNC powering the new driver connect technology, MyFord Touch, launching this year on the new 2011 Ford Edge. The system will make it easier for drivers to use voice control and get what they want more quickly using more natural phrases.
"Ford is committed to making voice recognition the primary user interface inside of the car because it allows drivers to keep their eyes on the road and hands on the wheel," said Jim Buczkowski, director of Ford electronics and electrical systems engineering. "The improvements we've made will make it easier for drivers to use and interact with it, even those customers that have never used voice recognition before."
Improved vocabulary
At the heart of SYNC is the speech engine, and Ford is working with speech technology leader Nuance to create and integrate a vast library of possible driver requests. This library will enable the SYNC speech engine to listen for and respond to more voice commands directly, recognize different words that mean the same thing (aliases), and integrate a vast number of point-of-interest (POI) names and business types into its navigation system.
"With this latest generation of SYNC, users can control the system without having to learn nearly as many commands or navigate as many menus," said Brigitte Richardson, Ford global voice control technology and speech systems lead engineer. "As we've gained processing power and learned more about how drivers use the system, we've been able to refine the interface. Customers can do more and say more from the top-level menu, helping them accomplish their tasks more quickly and efficiently."
Examples of some improvements to SYNC powering MyFord Touch-equipped vehicles include:
More direct, first-level commands
* "Call John Smith" dials the phone number associated with John in a connected phone's phonebook directly – the user isn't required to say "Phone" first
* Direct commands related to destinations, like "Find a shoe store" or "Find a hotel," place users in the navigation system menu where they will be walked through the POI search process
* The command, "Add a phone," will enter the phone pairing menu and walk users through the connection process – users don't have to enter a phone submenu to initiate the pairing process
Quicker, easier entry and search
* Navigation entries can be spoken as a single one-shot command; for example, "One American Road, Dearborn," instead of requiring individual city, street and building number entries
* Brand names are recognized by the navigation POI menu, allowing drivers to look for chain restaurants, shoe stores, department stores and more, as well as regional and local favorites
* Direct tuning of radio stations by simply saying "AM 1270" or "FM 101.1," or using SIRIUS station names or numbers such as "21" or "Alt-Nation"
Use of aliases
* Within the climate menu, users can voice-request the same function using several different phrases, such as "Warmer," "Increase temperature" or "Temperature up" – helping reduce the need for drivers to learn specific commands
* When requesting a specific song from an MP3 player, users can now say "Play song [title]" in addition to saying "Play track [title]"
Personalized access
* If an occupant's USB-connected device, such as an MP3 player, has been named, users can simply say the device name, such as "John Smith's iPod," rather than the less personal "USB" command
More friendly and adaptable
Ford voice engineers refined SYNC beginning with the two features customers interact with first: the voice recognition system and Samantha, the digital voice behind system commands.
To help SYNC react to driver commands more quickly and accurately, the team integrated Nuance's Unsupervised Speaker Adaptation (USA) technology. USA learns the voice of a driver within the first three voice commands, quickly creating a user profile and adapting to tone, inflection and even dialect for a 50 percent improvement in recognition performance. USA then continues to learn during that same trip, even picking out another user and creating a second profile if the voice is markedly different. Currently SYNC can actively adapt to voices in English, French-Canadian and Mexican-Spanish – with more languages on tap.
"The power of the SYNC voice control system is its ability to understand and respond to more natural language commands – and the advanced adaptability of the speech recognition technology enables the system to train itself with each successive use," said Michael Thompson, senior vice president and general manager, Nuance Mobile. "The adaptability of SYNC is pretty remarkable – a feature functionality Nuance and Ford worked hard to develop to ensure seamless customer interaction with the system every time it starts up. So even if the car owner has a cold or someone borrows the car, SYNC will adapt to the changed voice and process spoken commands without missing a beat."
Initial interactions also involve Samantha, the "voice" of SYNC. In an attempt to help Samantha sound less computerized, Ford boosted the size of her speech profile approximately fivefold. The additional speech units will help Samantha speak in a smoother, more human voice as she helps vehicle occupants accomplish their in-car tasks such as making phone calls, playing songs from a connected digital device and getting directions.
Voice poised to become primary in-car communication interface
With smart phones expected to replace desktop and laptop PCs as the primary web access point by 2015, some industry analysts believe voice control will replace touch devices like keyboards and screens as the primary method of search. Dr. Philip E. Hendrix, Ph.D., founder and director of immr and analyst with GigaOM Pro, says that a majority of smart phones will have optimized a Voice User Interface by the end of 2012.
Research trends show strong consumer acceptance of voice recognition technology. The Harris Interactive 2010 AutoTECHCAST study found that 35 percent of drivers(1) say they would be likely to adopt voice-activated controls or features in their vehicle, up from just over one-quarter (27 percent) in 2009. In recent Ford-conducted market research of SYNC owners, more than 60 percent reported they use the voice controls while driving.
Datamonitor, an independent research firm, predicts that the global market for advanced speech recognition in the mobile world will triple from 2009 to 2014. Market growth of speech recognition in vehicles is expected to grow at a similar rate, from $64.3 million in 2009 to $208.2 million in 2014.
Voice commands may reduce distracted driving
Ford knows that customers are increasingly using mobile electronics while driving, and studies show hands-free, voice-activated systems such as Ford SYNC offer significant safety benefits versus hand-held devices.
According to a 100-car study conducted by Virginia Tech Transportation Institute, driver inattention that may involve looking away from the road for more than a few seconds is a factor in nearly 80 percent of accidents. The improvements to SYNC should help drivers accomplish tasks hands-free using natural speech patterns and fewer commands, enabling them to focus on the task of driving.
The fact that there is ANY software update for any part of a vehicle is nearly miraculous to me. I have two fairly late model vehicles with horribly outdated audio and GPS interfaces. They were outdated less than 6 months after I got the car and I hold no hope of them ever seeing an update. One (Volvo) is impossible to ugrade, even by going to aftermarket hardware, because it's optical interconnect system is so ingrained in every function of the car, it can't be touched for even a crazy amount of cash. I'm rooting for Ford,. Although this sort of feature is woefully late for automotive consumers, I applaud this first step.
@appsman
While it's being touted as an update, existing cars won't be able to "update" their existing systems to run this software. The "update" is to the new line of vehicles. It's as much of an update as a 2010 vehicle with 250 horse power engine to a 2011 getting a 300 horse power engine...it doesn't affect the older models.
@coop486 Thanks for pointing that out, Though it bums me out since it means that NO progress has been made on this front. Oh well.
As a non-native speaker of English, I used to evaluate my English pronunciation based on the success rate of recognition from my Saab's voice command. I wonder how good I'll do on a Ford...
http://hiexpat.com
Gees, majority of the reviews for this system have stated that this is the best implementation of voice commands in a vehicle to date, regardless of the car manufacturer. But this dunce of a journalist just had to make it sound like SYNC is very flawed. It was almost perfect and now this makes it even closer to that. Let's get some objectivity on this site.
this is good sign as I am looking more and more at getting either a F-150 or F-250. SYNC is a HUGE Selling point to me
In my opinion, SYNC is one of the best voice control systems out there. There are a few quirks with it as people have mentioned, but the great thing is that it's upgradeable, so they can continue to improve upon it, which they seem intent to do (as long as they don't abandon the original SYNC platform in favor of myFORD Touch in the 2011 vehicles).
My biggest annoyance with it is that if I have multiple tracks with the same name by different artists, and I say 'play track ...', it always plays the first one it finds, there is no way to say 'play track ... by ...'. Other than that it recognizes my commands 99% of the time, and functions as I think it should. I've used 3 different versions in my 2010 F150 (since version 1), and each version keeps getting better.
Great trolling, Darren.
Wow go Ford. Make a speedier version of your new 2011/2012 Focus hatchback and I'm your 1st time customer.
"Sync: Go go gadget rocket engine."
This seems to be a good system, and like all navi/automation systems, it will continue to improve over the years. As usual, in five years we'll all look back at be amused at what an antiquated system it is. The navi in my 2010 Nissan Altima is pretty nice, but it's not as sophisticated as the SYNC system. However, the voice of my lady (the navi system voice) sounds much more natural and soothing. And don't we all need a little soothing as we weave in and out of traffic trying to get to our destination three seconds faster?
fwiw, sync's voice control has always been powered by nuance's voice control software. myfordtouch just uses a newer, fancier version of the software.
The Ice Cream example demonstrates a UI flaw. The driver has to take their eyes off the road to look at the list. Will it read the searched list I wonder? Bad example Ford & Nuance!