AllThingsD » voice recognition

Take a Note: Typing With No Hands

Walter S. Mossberg — Wed, 11 Apr 2012 01:08:24 +0000

I am writing this paragraph on an iPhone. But I am not typing it on the phone’s virtual keyboard. I am dictating it using a little-known feature that allows you to employ your voice instead of your fingers, wherever text entry is possible on the device.

[ See post to watch video ]

And now, for this paragraph, I have switched to an Android phone. Once again, I am composing these words using only my voice, and not typing them on the virtual keyboard.

Those two paragraphs, dictated as emails and then cut and pasted into this column on a computer, required far fewer corrections than you might think, given the bad reputation for accuracy that voice input on digital devices has acquired. I only had to add a comma I’d forgotten to specify in the first paragraph and capitalize the word “Android” in the second paragraph.

For me, a daily user of virtual keyboards, the process was quicker and more accurate than typing would likely have been, even for the relatively short blocks of text typically composed on phones.

So, on the suspicion that dictation on smartphones might prove useful for others as well, I’ve been testing it heavily over the past week. I used a top phone with Google’s Android software, the Samsung Galaxy Nexus, and an Apple iPhone 4S. In general, I found that, while dictation could occasionally fail badly, it worked surprisingly well in a wide variety of environments and applications.

On both leading smartphone platforms, I found that relatively short dictation—such as emails, texts, tweets, Facebook posts and notes—was at least as accurate, and often more, as typing on a glass screen. It was better in quiet environments, but did OK even in most noisy places like grocery stores, coffee shops and carwashes. It was also faster, since, as long as you don’t have to correct numerous errors, speaking is usually faster than typing on glass.

For this review, I am not mainly referring to Siri, the widely publicized, voice-controlled feature on the new iPhones, which can do things like tell you the weather, or stock prices. Nor am I discussing the “voice actions” on Android, which can perform Web searches and other tasks. Both can also help with some text dictation. I concentrated on a much simpler feature of both platforms: a small microphone key that’s included right in the phones’ on-screen keyboards.

Apple’s dictation system did better at capitalizing proper names.

Android phones have had this microphone key for a couple of years, and Apple added it to the latest iPhone, the 4S, last fall, and to the new iPad, when it came out last month. But I’m guessing that many users of these phones either haven’t used this special key, or haven’t even noticed it.

While the microphone keys work a bit differently on the two platforms, they are basically similar. When the keyboard appears, ready for you to type, you can instead hit the microphone key and simply dictate what you want to say. The phones then send your spoken words to a remote server, which rapidly translates them into text and sends them back to the phone’s screen. If corrections are needed, you make them by typing, though both platforms make this easier by indicating the likeliest errors, and suggesting alternatives.

A couple of caveats are in order. I didn’t compare dictation to typing on a phone with physical keys, whose devotees are often speedy and accurate. Instead, I thought the apt comparison was with a virtual keyboard, which is becoming the norm on phones, but is still a source of frustration for many users.

But Android was more reliable.

I also didn’t try dictating a long document, like this column, because phones are rarely used for lengthy composing.

I found that both platforms’ dictation systems worked well enough for me to recommend them. In case after case, both phones got it right, or close enough to require little correcting.

But there are differences. Android has an advantage in that, in the newest version of its operating system, it displays the dictated text almost in real time, lagging just slightly behind your spoken words. On the iPhone, the system only reveals its rendering of your dictation after you’ve tapped on a “Done” button.

Android’s dictation system also supports many more languages than Apple’s—40 languages and dialects, including Spanish, Chinese, Arabic and Hebrew. On the iPhone, only English, French and German are currently supported, though Apple says Chinese, Korean, Italian, and Spanish will be added later this year.

However, I found the iPhone 4S worked better than the Galaxy Nexus in noisier environments. For instance, in a crowded shopping-mall food court, while neither phone was perfect, the iPhone understood me to say: “I am dictating this email from the very noisy Court at Montgomery Mall on the iPhone”—missing only the word “food” and capitalizing “Court.” The Android phone mangled a very similar sentence as: “I am dictating this email on droid phone from the bearing noise for it montgomery mall.”

Google notes that, unlike Apple, it supports many phones, and that the results might have differed on another model, with better noise cancellation. Apple says the iPhone 4S does have noise cancellation. And, in any case, the two phones’ results were more comparable in quieter settings.

Apple’s system also did better at capitalizing proper names, like Stradivarius, or Red Sox, or even Google (which my Android phone, ironically, always rendered in lowercase). But Google says it will be updating its dictation feature in weeks to better handle proper names.

On the other hand, I found that, when Android did err, it had a more extensive and easier to use manner for correcting those mistakes than the iPhone did. Android was also more reliable; sometimes the iPhone returned no text at all.

Still, I found these differences less important than the fact that, for me, the results on both platforms were impressive. On both, if you say words like “period” or “comma,” you generally get the punctuation mark (though both try to make the distinction when you actually want a word like “period.”)

And, in test after test, both did a good job. Errors were generally fewer than if I had typed the words quickly.

Both have a downside: Because they do the transcription on their servers, and they are anxious to improve, they do retain some information about what you’re saying. Both companies say they respect your privacy, but, if you worry about transmitting your messages or notes to Apple or Google, don’t use dictation.

Otherwise, especially for those who find typing on glass clumsy, the microphone key on Android and the new iPhone is something you might want to add to your arsenal of ways to use your phone.

Email Walt at mossberg@wsj.com.

Year of the Talking Phone and a Cloud That Got Hot

Walter S. Mossberg — Thu, 22 Dec 2011 02:04:22 +0000

While other industries struggled, consumer technology seemed to march ahead as always in 2011, with important new products and services continuing to roll out. Sure, some tech companies, like BlackBerry maker Research In Motion, suffered reverses. And some products, like Hewlett-Packard’s TouchPad, flopped. But many shone.

[ See post to watch video ]

So here is a look at a few of the biggest tech products of the past year, with some analysis of what they signified and what issues they raise for 2012. As with all my columns, this one is focused only on products and services provided to consumers. Also, as usual, this column isn’t meant to offer investment advice or to evaluate the management skills or financial condition of companies.

The iDevices

Siri, right, the voice-controlled artificial-intelligence system, made the iPhone 4S stand out even though it looked like its predecessor.

Even in a year when its iconic leader, Steve Jobs, resigned as CEO and then passed away, Apple kept going from success to success. In March, it introduced the iPad 2, a thinner, lighter, faster version of its groundbreaking tablet and sold tens of millions of them. In October, it brought out the iPhone 4S, which proved popular even though it looked identical to the prior model. One reason: The phone introduced a voice-controlled artificial-intelligence system called Siri that answers questions and performs tasks without requiring typing or searching. Siri, while still rudimentary, could herald a revolution in practical artificial intelligence for consumers.

The lesson here is that Apple is driving the industry toward simpler, more reliable digital experiences tied into ecosystems of content and cloud services. It is expected to bring out radically new iPhones and iPads in 2012. But can it fend off challenges from popular, rapidly improving rivals using Google’s Android operating system? And, in the absence of Mr. Jobs, can it keep churning out game-changing hits?

With its ultralow price and Amazon connection, the Kindle Fire may be the first tablet to gain significant traction against the iPad.

The Kindle Fire

Despite some initial software flaws and its chunky, plain hardware, the diminutive Fire appeared to be the first color tablet to gain significant traction against the iPad. The biggest reasons are its ultralow $199 price and its tie-in to Amazon’s huge content library. But the Fire may have started a trend that could be a problem for Google: It demotes the Android operating system to an under-the-covers piece of plumbing, ignoring Google’s user interface and apps marketplace.

In 2012, Amazon is expected to bring out a larger, possibly sleeker Fire, and, if it continues to prove popular, it could attract larger numbers of apps designed for the Fire and sold only through Amazon. But despite its success with simple e-readers, Amazon has little experience as a maker of general-purpose computing devices, and it will have to be nimble and creative to keep up with Apple and more-traditional Android rivals.

LTE

Though several cellular technologies claim the moniker “4G” to indicate fast data speeds and greater capacity, only one, LTE (Long Term Evolution), delivers true broadband speeds consistently. This past year, it finally spread significantly in the U.S., both in terms of geography and in the number of devices supporting it. The LTE leader by far is Verizon Wireless and it has the potential to make the wireless Web, and wireless streaming of video, the equal of their wired counterparts. AT&T is racing to catch up and Sprint, which uses a different 4G system, says it will join the LTE parade.

But at this stage, LTE still consumes too much battery power. And LTE networks, if they become the norm, could get overwhelmed. To fend off this prospect, the biggest carriers in 2011 began charging more for greater data usage, a move that could curb the spread of innovative services that rely on large data downloads, such as video streaming and sharing of music and high-resolution photos.

More companies took advantage of cloud computing, with Google introducing the Chromebook, which relies almost entirely on the cloud.

The Cloud

Many players began offering consumers the opportunity to both store their data on, and run apps from, remote servers on the Internet, a system called cloud computing. Google even introduced a new kind of laptop, the Chromebook, that has almost no internal storage and relies almost entirely on the cloud. An example of a cloud service: music “lockers” that store all your songs on multiple devices. Cloud services are sure to expand in 2012, but questions remain on their reliability, security and privacy. And while most now cost little or nothing, these offerings could become another monthly fee burden for consumers.

Android became easier to use with the release of the Ice Cream Sandwich version, used in the Samsung Galaxy Nexus.

The Android Army

In 2011, Android overtook Apple’s iPhone and iPad operating system, called iOS, in users. Though no single Android device is as popular as the iPhone or iPad, Android is now the collective leader, with hundreds of devices using it. Samsung, in particular, had success with its Android-based Galaxy devices. And a new version, called Ice Cream Sandwich, continued Android’s steady improvement by making it easier to use. However, Google may be losing control of Android, as hardware makers and cellular carriers redefine it to suit their own needs, and fail to offer consumers updates in a timely fashion. Except for the Kindle Fire, the operating system hasn’t caught on in tablets.

Windows

Microsoft has been way behind in the new areas of super-smartphones and tablets. In 2011, the software giant began to try to reverse that situation. It introduced the first competitive version of its sleek, sophisticated Windows Phone software, called Mango, though so far without much uptake by consumers. And it previewed a bold new version of main Windows, called Windows 8, with a multitouch interface that, unlike Apple’s approach, is a single operating system meant for both PCs and tablets. It will start shipping in 2012.

Following in the Apple MacBook Air’s footsteps, a crop of thin and speedy ultrabooks, such as the Toshiba Portege Z835, pictured, became the new standard for laptops, with Windows PC makers coming up with their own versions of the machines.

Still, Windows Phone must somehow attract many more users. And Windows 8 is a gamble, because it includes two interfaces: the new tabletlike face and the old, familiar Windows look, which could confuse consumers.

Ultrabooks

In 2011, Apple’s MacBook Air, previously a niche product, became the new standard for laptops—thin, light, speedy, with long battery life and solid-state memory for storage instead of a hard disk. Now, Windows PC makers are following suit with similar machines called Ultrabooks.

Ultrabooks may recharge the Windows laptop scene in 2012. However, they will have to become less costly—they now hover at around $1,000—and their solid-state drives don’t offer the capacity of hard disks at an affordable price.

The Lenovo IdeaPad U300

Television

The reinvention of television picked up steam in 2011, albeit in a small way. Despite some miscues, Netflix streaming of TV shows to many devices grew in popularity. Set-top boxes that bring Internet video to TVs, like the Roku box and Apple TV, got better and more popular, though Google’s competing effort was a dud. Microsoft’s Xbox is set to compete strongly, using its Kinect add-on to find and play media apps with gestures and voice commands.

The big test may come in 2012, when Apple is believed to plan to ship a whole new type of Internet-connected TV, which the company hasn’t confirmed. A big obstacle: Cable and media companies will have a huge say in this potential revolution, and the current system serves them well.

So, 2011 was an exciting year in consumer technology. I can’t wait for 2012.

Email Walt at mossberg@wsj.com.

2012: Siri Is a Stunner, Amazon Is Amazin' and Security Gets Spendy

Arik Hesseldahl — Fri, 09 Dec 2011 04:59:40 +0000

On Thursday night, I attended a dinner at New York’s Waldorf Astoria Hotel, hosted by Mark Anderson, the CEO of Strategic News Service, a newsletter that many senior tech execs subscribe to. At this annual event, which I missed last year, Anderson makes predictions concerning what he thinks will be the dominant forces shaping the technology world in the coming year. And his predictions are always interesting.

Ahead of the dinner, Anderson stopped by my office to let me have a peek at his 10 predictions, and we talked them over a bit. All 10 are below, along with some comments from Anderson that emerged from our conversation.

Before diving into the predictions, Anderson tells me there is a grand theme that unifies them all: “Integrating everything.”

What does that mean? “It means a whole lot of stuff that needs to be integrated. We don’t need anything new at all. There’s so much work that needs to be done with the existing tool sets. Steve Jobs didn’t really invent anything at all. But he was great at integrating things into a product. There’s a lot more of that work to do. We have to do it in the phone world and the TV world and the health care world. We have lots of devices and lots of chips and lots of operating systems and lots of content. The bigger question is, how do human beings use it all efficiently?”

As an example, he cites the collaboration between Nuance, the speech software company, and IBM, bringing the Watson computer of “Jeopardy” fame into the area of health care. “For the first time, the idea of evidence-based medicine won’t just be in a magazine article,” Anderson says. “A doctor will be able to pick up his phone and describe four symptoms, and find out what the likely diagnosis is, what the indications are. It’s fantastic.”

So here are those 10 predictions, with additional comments from Anderson:

1. TV becomes the new center of gravity in the tech universe. All the other devices find their niches in the TV galaxy. Microsoft’s attempt to integrate Kinect into TV is a strong if qualified success. Smart phone-TV integration software becomes a new category. Pad-TV integration becomes common.

“Apple will hustle to launch the next version of Apple TV, and it will be a roaring success and be seen as Tim Cook’s first great product success. But what it really will be is Steve’s last product.”

2. 2012 will see tectonic shifts in phone markets. “Nokia will fail to come back, which is pretty clear to everyone except the people in Finland.” Samsung, Anderson says, will retain its spot as the new global leader in mobile phones by volume, and will keep this crown despite the debut of Microsoft’s Windows Phone 7.

Meanwhile, Anderson says, Google will lose control over the Android operating system, mainly because unlicensed versions of Android will multiply in type and in installed base, especially in Asian countries. “It’s already a balkanized environment. Now Google loses control of the technology entirely. China is already running an unlicensed version of Android, and I think there will be more of that.”

Finally, the smartphone will finally emerge as the dominant category of wireless phone. “Why would you have anything else? And why would sellers of content and services want you to?” he says. “It doesn’t matter if you’re in a rich country or a poor country. This stuff is cheap.”

3. Clouds are for consumers, and for start-ups. Even as a large number of big companies move pilot projects onto external clouds, it will become clear that the real trend is for enterprise to stay away from clouds in all key areas, for reasons of both security and reliability.

“The cloud guys hate this because they want to sell to enterprises,” Anderson says. “But the security issues are becoming really intense. If you’re a CIO, it’s a terrible environment, and you’re a target, for sure, especially if you’re a company with a lot of intellectual property. I’m not implying that things like SAAS (software as a service) aren’t a big trend. But no one is going to put their valuable IP on the cloud.”

4. Security splits the tech world in two, finally getting attention from CEOs. Companies with real IP start to realize they have to “go big or go home” with their security response, and their spending on protecting their “crown jewels” rises dramatically.

5. Siri stuns the world. Siri, on Apple’s iPhone 4S, has sounded the arrival of Internet personal assistants, and the world will spend this year marveling at what Siri and its rivals can and cannot do — and what they can learn to do.

“I think we’ll see a bunch of these things,” Anderson says. “Siri will get much better. It will learn how you learn. We’ve never seen people have long-term relationships with machines before, but it will be a long-term relationship, and she will remember everything, but make good use of it. She will know you learn better by seeing than hearing, or that it takes three times to tell you something. All those things that you have to program today should be learnable. None of that has been done yet. That creates a real friendship. And I think we’re going to start seeing personal assistants not just for everyday life, but for professions like medicine or car repair. Instead of just having Siri be everything, there will be many Siris for different contexts.”

6. We enter the amazing world of Dave and HAL, as voice recognition comes of age. From hospital to car, mobile to home, Kinect to Siri, exercise to play, work to entertainment, remote control to direct action, from Microsoft to Apple, from Tellme to Nuance — the time has come for computers and humans to talk to each other. With lots of funny stories, big bloopers and amazing breakthroughs, humanity at the end of 2012 will be talking to machines in a normal voice, and it will not seem unusual, nor be the cause of unending frustration.

“The voice-recognition part is almost trivial,” Anderson says. “The important part is context-sensitive understanding. It used to be that all the researchers at Carnegie Mellon used to think that all you needed was more computing horsepower to do better at voice. It turned out that was wrong. It was right for a little while, but the real problem is context. And so, if you can build up that database where you can search it contextually for what to expect, that is where you get all the mileage.”

7. E-readers prosper, but pads continue to dominate what Anderson calls the “carry-along” market. Pads and tablets will come down in price and get closer to prices of e-readers. Meanwhile, Anderson says, Amazon’s Fire will move upmarket and evolve into a full-fledged tablet.

“If you look at the specs on the Fire, it’s a tablet, but it’s hobbled,” Anderson says. “So I think that this is part of the whole strategy: Come in and sell at a low price, and then later unveil a more complete tablet. Apple will stay ahead, though. A lot of people are asking me if Amazon will catch Apple, and the answer is no. The way it’s configured right now, there’s no way the Fire will catch up with the iPad.”

8. The consumption world explodes. Get ready for new devices, new content, new bundles, new connection techniques, new distribution channels, new aggregators, new tablets, new phones, new players, new self-published authors, new garage bands, new consumption models riding on social networks. There is nothing but high energy in the content consumer market. People are now ready to spend subscription money, and the publisher response will be huge. “It’s going to be a huge melee of stuff,” Anderson says. “We’ll invent more stuff to consume, and it will be very hard to figure out who the players are from week to week, and how they’re doing. They may not even know themselves.”

9. Governments and corporations focus on intellectual property as though it were their most prized asset. It is. This new global understanding leads to a reevaluation regarding giving critical IP away for nothing versus protecting it. The age of what Anderson calls “IP naïveté” is over, and the question of proper IP valuation is here.

What is IP naïveté? “When Jeff Immelt stood on the steps of the White House the day after he was named jobs czar, and handed the plans for GE’s most important jet-engine project to Hu Jintao in order to get the permission to be allowed to bid on maybe selling engines to China — that’s IP naïveté,” Anderson says. “Thinking that’s not going to come back and show up for sale in Houston from some Chinese company in about six months is IP naïveté.”

During 2012, he says, companies and countries will start valuing their intellectual property not for its replacement value, but for figures that are magnitudes larger. State-sponsored IP theft will shift from being considered a nuisance and more along the lines of an act of aggression.

10. Amazon gets it all. Between outdoing Wal-Mart online, to beating the booksellers and delivering groceries, and making new inroads in video streaming, Amazon will prove that one company can indeed have it all. Strong Kindle and Fire sales will only be icing on the cake.

Here's How Microsoft Is Adding Voice Control and Gestures to the Xbox (Video)

Tricia Duryee — Mon, 05 Dec 2011 05:00:29 +0000

Microsoft is planning a massive software update on Tuesday for the Xbox, beginning the game console’s transformation into an entertainment hub for the whole family.

The free update will allow users to control the console using their voice and gestures, or even their Windows Phone (if they have one).

In addition, Microsoft will begin to add more than 40 content providers to the console to increase the catalog of live and streamed TV, movies and music.

Microsoft has announced nearly all of these details previously, including some of its content partners, so today’s announcement serves as a reminder now that the final product is ready to go.

Last week, I met up with Michael Suraci, Xbox’s director of marketing, to get a preview of the updates.

According to Suraci, Kinect, the motion sensor that launched last year, is a central part of the update. When it was introduced, it seemed that all it was good for was dance games, but clearly Microsoft had much bigger plans for the camera and the microphone.

Now users can speak naturally to the Xbox, which tears down a number of barriers to family members in the household that weren’t comfortable with the clunky controller. If Microsoft pulls it off, it could teach people that televisions are meant to be talked to, just as Apple has taught people that screens are meant to be touched.

An unknown subset of the nearly 60 million Xbox owners worldwide that have purchased Kinect will be able to use all the new features in the update.

But everyone will have access to many of the updates.

One major improvement is in navigation. For example, the old interface required the user to decide which category they wanted to go into. For example, games, video or music. Then, they had to choose the application, like Netflix, ESPN or Zune.

In the new user interface, the person can search across all of the categories and apps.

As Suraci demonstrates in the video, a user can say: “Xbox: Bing, ‘Fast and the Furious.’”

The results show all of the content that matches that criteria across games, music, video and other categories. The style of the user interface will be recognizable to anyone using a Windows Phone. The format will also be carried over to the upcoming Windows 8 update.

During Suraci’s demonstration, the software got confused a couple of times, but still, searching by voice will be much faster than typing in a string of words, letter-by-letter, using the controller to scroll through the alphabet.

Going forward, the Xbox could replace the need for a second set-top box in the household, but as Peter Kafka has mentioned before, it’s not a service for customers looking to cut the cord. In order to stream live TV, or watch movies, you’ll either have to pay for a subscription — like Verizon FiOS or Comcast’s Xfinity — or pay a la carte.

On Tuesday’s launch, the amount of content that will be available in the U.S. will be somewhat disappointing. But later in December and in early 2012, you will start to see integrations with Verizon FiOS, YouTube, HBO GO and Xfinity On Demand, TMZ, UFC, Wal-Mart’s Vudu service and others.

[ See post to watch video ]

Wal-Mart's New Apps Will Integrate Coupons and Voice Recognition

Tricia Duryee — Wed, 09 Nov 2011 05:00:33 +0000

Wal-Mart is launching its first iPad app and is refreshing its iPhone app just in time for the holidays.

Both applications will enable consumers to shop online or see what’s available locally. Orders can be shipped or picked up.

The apps are being released ahead of the end-of-the-year shopping frenzy, which many retailers are hoping will be one of the busiest mobile commerce events ever.

“As customers use more smartphones and tablets, the Wal-Mart customer is doing the same thing,” said Gibu Thomas, the company’s SVP of mobile and digital.

That’s true, and it’s scary for big retailers. Increasingly, consumers are scanning barcodes to get reviews and compare prices before deciding whether to make a purchase or go somewhere else.

Thomas said the apps are not a defensive move against this trend, but rather a way to give consumers what they want. ”We want to know how we can help our customers shop better with us, which will make them shop more with us,” he argued.

Wal-Mart’s first iPad app launched about a week ago.

It lets people browse inventory in their local store, while also seeing what else is available online. Instead of duplicating the online experience, they’ve created categories. For instance, in the home section, shoppers will browse a catalog-like experience, where they’ll see pots and pans, stereo speakers and outdoor fireplaces without any additional information.

To see prices or more details, a user will have to choose an item.

The major new feature of the iPhone application, which originally launched more than a year ago, will be a shopping list with integrated voice dictation using Nuance’s speech recognition, and discounts through a partnership with Coupons.com.

The app will be available as soon as it receives Apple’s approval.

This app gets pretty close to what many companies have been describing as mobile commerce — minus near field communication that would allow users to tap and pay. And, at Wal-Mart’s scale, this is truly something for the mass market.

Thomas said 90 percent of consumers who come to a Wal-Mart store on a weekly basis come with a shopping list.

The iPhone shopping list feature will allow people to enter items manually using predictive text. Type “cheddar,” and you’ll see a number of cheese brands and the price of each item.

Consumers will also be able to enter items using their voice. When Thomas demonstrated it to me, the app was fairly accurate. It recognized orange juice and sour cream as one item each, but Campbell’s soup came up as two.

Because the real price is listed, budget-conscious shoppers can decide how much to pay for groceries before even getting to the store. Coupons from Coupons.com will also be integrated into the shopping experience. You might type in “yogurt,” but decide on Yoplait for the discount.

For some stores, Wal-Mart has given consumers the ability to find out what aisle the items are in from the phone. That capability will roll out to more stores as it comes out of beta. Eventually, the information could be overlaid on a map to show the most efficient route for getting all of your items.

Thomas also said the company anticipates being able to store shoppers’ receipts electronically, so that items purchased will be uploaded to an individual’s device, making it easier to create shopping lists in the future.

“We think of your mobile phone as your loyalty card. We don’t have cards but we think of it in the sense that you can use your phone to surface real-time discounts,” he said.

The applications were built by @WalMartLabs in Silicon Valley, which serves as the technology hub for the Bentonville, Ark.-based retailer.

Sell Me an iPhone, Siri

John Paczkowski — Wed, 02 Nov 2011 10:30:05 +0000

No surprises here: one of the biggest selling points of Apple’s new iPhone 4S is Siri, the speech-recognition personal assistant that’s built into the device.

While Siri still has a way to go if it is to popularize voice as the next major user interface, its natural-language processing and automation abilities are already good enough to be a real competitive advantage for Apple in the mobile market.

It’s pushing competitor costs up, as rivals scramble to come up with equivalent voice command offerings — not a cheap or easy feat, considering the level of Siri’s applied artificial intelligence and speech comprehension.

And better than that, it’s creating a consumer bias towards the 4S, Apple’s newest iPhone and presumably the one with the highest margins.

Sterne Agee analyst Shaw Wu says his latest checks with industry and supply chain sources show broad sales strength across Apple’s entire iPhone portfolio, but most of all for the 4S. Evidently lots of folks who could be spending $99 on the iPhone 4 are opting to fork over another $100 for the 4S — and a lot of them are doing it for Siri.

“Despite global macroeconomic headwinds, Apple continues to defy conventional wisdom with a higher-end product mix,” Wu says. “Talking to industry sources, what’s driving the 4S is better than expected reception of its new Siri software.”

The reason for that? Siri’s voice recognition actually works — and pretty consistently, unlike competing solutions, which are often unreliable. And as of today, it’s still in beta. So it will certainly get better and more powerful over time. And it will continue to drive sales.

To wit, Wu’s forecast for the December quarter, which we’re only about a third of the way through: 26 million iPhones — a new record.

Yelp's Mobile Visits Could Soar With the Help of Apple's Latest iPhone

Tricia Duryee — Wed, 12 Oct 2011 19:59:18 +0000

Already, millions of visitors are looking up Yelp reviews on the phone, but that’s only likely to grow with the recent integration of Apple’s upcoming iPhone 4S.

In the month of September, Yelp said it registered 8.5 million visitors on its mobile Web site and iPhone, Android and Blackberry applications.

It said nearly 60 percent of that traffic was coming through its apps, but the remainder was from visits to its mobile Web site. To help those users, Yelp is updating its site today with a better design.

However, Yelp’s mobile visits are likely to spike in the coming months due to a new partnership with Apple.

Earlier this month, Apple announced it was integrating sites, like Yelp and Wikipedia, into its voice-activated service, called Siri.

In his recent review of the phone, Walt Mossberg had this to say about it:

Siri can find information in Wikipedia, Yelp and Wolfram Alpha. It successfully answered when I asked it, “Who’s the president of Iran?” (though it misunderstood me the first time) and “Who stars in ‘Boardwalk Empire?’” When I asked for a “French restaurant in Bethesda, Maryland,” it instantly returned a list from Yelp, ranked by user reviews.

Yelp will still have a long way to go, however, before a majority of its visits come from mobile.

Compared to the 8.5 million mobile visits in September, 54.5 million visits were made to its traditional PC site.

Apple's Siri: Game-Changer, Not Gimmick

John Paczkowski — Mon, 10 Oct 2011 10:10:22 +0000

Siri, the voice-operated, natural-language-based personal assistant Apple has built into its forthcoming iPhone 4S, might seem, as Gizmodo’s Mat Honan complained, “like the most amazing thing I’ll never use.”

But the application — which promises to reply to questions with answers, and to orders with actions — has the potential to be transformative, another of Apple’s industry-changing innovations. Analysts say Siri’s savvy mix of voice recognition, artificial intelligence and operating system integration may prove to be a far more potent combination than many expect. The software’s ability to interpret meaning and execute actions brings a potentially revolutionary new feature to the iPhone and, soon, to other Apple hardware as well.

“Early reactions have missed the extent of Siri’s capabilities, perhaps confusing it with mere speech recognition or simple keyword-based voice response systems,” says Cross Research analyst Shannon Cross. “We believe the use of natural language and potentially the ability to distinguish between voices could one day change the way we interact with electronic devices and provide a substantial technology advantage to Apple. Quite simply, we have not seen a demonstration of comparable AI in any other consumer system.”

Which means it will likely be difficult and time-consuming for rivals to match it. And if it works as flawlessly as it did during Apple’s demonstration last week (onstage and off; Siri nailed every question and command I threw at it in the hands-on room after Tuesday’s event) the company will have yet another tentpole point of differentiation from the competition.

And once Apple has it dialed in on the iPhone, it will almost certainly be extended to other hardware as well. Cross sees it headed to the iPad 3, the iPod, the Mac and, at some point, Apple TV or that mythical Apple television set.

“We think it would be very compelling to own a TV or a device that could quickly answer the request, ‘I want to watch the Yankees/Red Sox game,’ by changing the TV channel without requiring the user to look at a guide or use a remote control, or even specifying HD or standard definition feeds, since you would want the HD channel if available,” says Cross. “Or, you could instruct the device to record all new episodes of a show, without leaving the program you are currently watching. Finally, since you are online, a Siri-enabled TV could answer whether your iPhone or computer has received a new message, and let you respond accordingly.”

Nuance to Buy Virtual Keyboard-Maker Swype for $100 Million

Tricia Duryee — Fri, 07 Oct 2011 02:04:27 +0000

Nuance, a voice recognition technology company, has acquired Seattle-based Swype, which is known for its very popular virtual keyboard commonly found on Android phones, according to a person close to the situation.

A source pegged the deal’s size at more than $100 million, but less that $150 million.

The company was not reachable for comment. The deal was first reported by Uncrunched, but I was able to confirm it separately.

The acquisition is expected to be announced tomorrow by Nuance, a publicly held company that has a market value of nearly $7 billion. [UPDATE: Here it is. Total price: $102.5 million.]

Swype was co-founded by Cliff Kushler, who previously worked at Tegic Communications, which is known for developing T9, the predictive-text software that is now owned by Nuance.

As touchscreen phones became more popular, the predictive texting that eliminated triple-tapping became less necessary, as new forms of input on a piece of glass became mandatory. Nuance bought the technology from AOL.

The idea was first hatched by Kushler and co-founder Randy Marsden, who developed the onscreen keyboard included in Windows. In 2008, Mike McSherry, who was the co-founder of both Amp’d Mobile and Boost Mobile, became CEO.

The company raised a $3.5 million round earlier this year from existing investors, including the venture capital arms of Samsung, Nokia and DoCoMo.

Swype has found a huge following on Android devices, and at the time of its last round was estimating that it was being shipped on half of all Android phones. It also had its sights set on the tablet market. Nuance is an obvious acquirer of the technology because of the wide swath of patents it owns in the space, but Google or another handset maker would have been an option, too.

The importance of voice and text entry became even more prominent this week after Apple unveiled Siri, a voice-activated assistant.

Will Apple's Siri Make Talking to Your Phone Seem Normal?

Ina Fried — Wed, 05 Oct 2011 21:25:52 +0000

While Siri was probably the biggest technology advance that Apple introduced on Tuesday, it remains unclear just how ready smartphone owners are to spend time talking to their phones.

The fact is that voice recognition has been a part of cellphones since long before their was an iPhone. Voice dialing has been commonplace since the 1990s. The issue has always been how good is the recognition and how powerful are the capabilities.

What Apple is promising with Siri is a leap forward in both the amount of things one can do, as well as in the ability to do those things using just natural speech. While prior voice control handled things like dialing and playing a particular song, Siri promises to answer all manner of queries, regardless of how they are phrased.

Apple is not alone, though, in pursuing speech as a more dominant form of input.

Both Google and Microsoft have been beefing up the voice capabilities of their products as well. The just-released Mango version of Windows Phone allows a user to have text messages read to them and to dictate replies. That’s in addition to existing Windows Phone features that stem from the company’s Tellme acquisition. Microsoft is also working to expand its use of speech technology into all manner of other devices from cars to Windows laptops to the Xbox.

[ See post to watch video ]

Android, meanwhile, has an app that adds a variety of “voice actions” to Android, including the ability to dictate a memo, get directions, search the Web and more.

Beyond the competitive issues, though, there is also the question of just how cool it will be to talk to one’s phone. If Apple can convince people it is cool, there is a lot of opportunity there. That said, Siri has also debuted to plenty of mockery.

A Twitter user posing as HAL from “2001: A Space Odyssey” got in a couple good one-liners.

“Siri thinks she’s so cool because she has voice recognition but can she lip read? That’s the sign of a truly awesome computer,” @HAL9000 posted. Others have noted that the name Siri sounds a bit like some not pleasant words in certain languages, such as Japanese, where it sounds a bit like a word for buttocks. Of course, almost any name can be twisted into something not so nice, as this classic Nicholas Cage Saturday Night Live sketch reminds us.

I had a brief chance to play around with Siri yesterday and was impressed that it could handle queries seemingly no matter how they were phrased.

“Do I have a meeting at 2?” or “How hot is it?” all worked. The query “Where can I get a taco?” brings up a list of nearby Mexican restaurants and their Yelp ratings.

There are some limitations. Siri, for example, lets one hear an incoming text message and dictate a reply. However, there doesn’t appear to be a way to have that dictated message read back to make sure that Siri got everything correct.

For now, Siri is in beta and only works in a couple of languages, though Apple is promising additional features and language support to come.

Mostly Sunny With 100 Percent Chance of Apples

Shawn Carolan — Mon, 03 Oct 2011 21:43:00 +0000

The forecast is certain. Tomorrow, Apple will rain features from the Cloud, and it’s a very big deal. The iPhone 5 will be the first device that relies on the Internet and server farms to complete its functionality rather than a PC. The company that popularized the personal computer in 1977 is officially telling us we no longer need one. It’s the mark of a new age. The features will be awesome and the implications vast, of that I’m certain.

But I don’t know the details. Employees new and old keep their secrets close. That said, certain aspects of iOS5 have been made public for developers and speculation seems high that this is the juncture where Siri, a company on whose board I sat, will re-emerge as a core part of the operating system. Here’s a heads up on what’s coming now — and perhaps later — so you can prepare.

PC Free
Apple’s recently announced iCloud offers a host of new features but the most underappreciated is device configuration in the Cloud. It will have everything you need to configure and keep your iPhone up to date without a PC. Every Apple device you have will be linked with an Apple ID and iCloud will know the configuration of each one. No more long sync required before a phone upgrade, no more painful restore, simply enter the Apple ID and password and voila, good as new. This will make life easier for people with multiple iOS devices, but the implications go far beyond.

When configuration lives in the Cloud, modification to the configuration happens in the Cloud as well. That means you could install an app onto your phone while clicking a Facebook ad, reading a blog, or responding to an email. Every banner advertisement you see on the web will be an opportunity for app developers to entice you, and with your browser already cookied, a single click could make the new app magically appear on your devices.

This marks a major change for mobile app developers to promote their wares. Being on the Top 25 list won’t matter quite as much; there will be lots of ways to get the word out and drive downloads. The same PPI (pay per install) ad economy that multiplied AdMob’s revenues and led to the Google acquisition will be available to the rest of the Internet ad landscape. With the sophistication of ad exchanges today, app developers could bid on impressions of only Apple users and efficiently target the right people. Developers will need to have instrumented analytics all across this marketing funnel to maximize their opportunity and not get killed by others who have figured it out. Advice: App devs need an Internet acquisition animal in-house.

Assistant
Sixteen months ago Apple acquired a technology company named Siri. Siri was small, with three amazing founders — Dag Kittlaus, Adam Cheyer, and Tom Gruber — as part of a total of only 20 employees. Their size did not match their monumental ambition. Founded out of SRI where the technology originated and a Series A round from Menlo Ventures and Morgenthaler, the company was the first to make a “Virtual Personal Assistant” actually work. As an app running on the iPhone, users speak in natural language to book tables, order taxis, check flight times, and many other functions. In fact, their original venture pitch called the service “Hal” after the computer personality in the movie “2001: A Space Odyssey.” The app requires the Cloud because although voice is captured on the phone, the computation required to parse the words into intent and then invoke the chain of web services to accomplish the user’s goal is too much to run on the phone. Siri’s server farm does the heavy lifting.

The native integration of Siri into iOS could change the game in three ways. First, voice input will be a breakthrough for touch screen devices. Although users tolerate soft keyboards in exchange for larger screen size, typing anything of length is still painful and even short bursts are more convenient with voice. Siri found the overwhelming majority of queries were spoken rather than typed. High-quality voice recognition along with Siri’s semantic processing could allow a new level of instant gratification when capturing a reminder, queuing a playlist, or sending a text message, especially while driving.

The second game-changer could be voice access to apps. While Siri had to do all of its integration with other services (e.g., OpenTable) in the Cloud via web APIs, as part of iOS it would be possible to interact with any app on the phone seamlessly, with login credentials already there. Imagine being able to say “Checking account balance” and the banking app comes up to that page, or “Directions to Jim’s house” and the phone starts TeleNav to navigate you there.

The third implication is that Apple would be joining the search game and squaring off with Google. For the category of searches that people do on the go, the desired result is often a completed action rather than a page of blue links. Siri is a superior technology for getting the job done quickly. Both companies have a mobile operating system, a mobile device, an app store, and now an engine for navigating the web. It will be a fun one to watch. Advice: 2011 will mark the year a voice user-interface delivers real value and will rapidly become a must-have feature, prepare to respond.

Media on Demand
In addition to the PC Free features deriving from configuration in the Cloud, iCloud also stores media and data in the Cloud where it belongs. For personal media like pictures and videos, that means no more priceless baby shots at risk of deletion on the phone. All of the pics will get synced between devices and likely be shareable from the web.

Purchased media like music and videos will also live in the Cloud. For $25 a year, iTunes Match allows users to unlock pristine copies of all those songs they, um, ripped from CDs. Time will tell if it’s enough to stop the flow of people from iTunes to Cloud music services like MOG and Spotify. Movies will be in the Cloud too, allowing start times to be counted in seconds rather than the minutes required for download and sync. Apple already dropped the hard drive in their 2nd gen AppleTV to turn it into an Internet streamer like Roku, and I wouldn’t be surprised if a larger screened device resembling a TV appeared soon. To monetize all of this newfound cost, Apple has a model that seems inspired by Dropbox: Offer the first 5GB free and when users hit their limit they have little choice but to plunk down some extra bucks for an annual storage fee. Advice: Don’t waste your money on an Internet-enabled TV, it will be obsolete by the time you plug it in.

iNavigator
In 2007 after the iPhone launched, I emailed Steve Jobs to encourage him to take a meeting with TeleNav, a portfolio company who pioneered navigation on the mobile phone. Before TeleNav, the gigabytes of data required to render U.S. street maps was too large to put on phones so TNAV pushed the data, routing engine and traffic into the Cloud and streamed just the information required for GPS-enabled handsets to give turn-by-turn directions. They now power the majority of carrier navigation services in the United States. The fit with the iPhone was a natural. I fired off my carefully crafted note and got back a brief reply:

Shawn, Which provider does TeleNav get their map data from? Why would it be hard for Apple to also license this data and extend its own map application to do what TeleNav’s does?
Thanks, Steve

I replied with reasons of course, but never heard back. Apparently he decided Apple should do it themselves. Since that time, Apple has hired a number of engineers with navigation expertise. Though it has taken them several years, the service has been spotted recently in the wild and will likely show its face soon. Advice: Get an iPhone mounting kit for your car.

The move to the Cloud represents a tectonic shift in the IT landscape for both enterprises and consumers and the disruption is just beginning. It’s wonderful to see Apple embrace it wholeheartedly to drive value for their customers. For companies that are prepared, there will be great opportunities as a part of the ecosystem. The analysis could continue for pages, but I still haven’t found the iPhone5 pre-order page so you’ll have to excuse me for now.

Shawn Carolan is a Managing Director at Menlo Ventures, where he has been for nine years; he focuses on consumer Internet and mobile investing. He sits on the Boards of IMVU, PlayPhone, Roku, Talari, TeleNav (NASDAQ: TNAV) and YuMe.

IBM Launches Network of 'Spoken Web' Sites In India

Arik Hesseldahl — Sat, 05 Mar 2011 01:15:33 +0000

The conventional wisdom says that access to the Web is good for the poor because they get access to information that can help them improve their lot in life through education and finding jobs. But what good is a PC and a Web connection if you can’t read or write?

It’s a fundamental question that’s not often answered in western countries like the U.S. where most people can read and write, but a more common one in developing countries like India, where the literacy rate tends to be lower. In 2006 UNESCO pegged the average rate of literacy for people in India over the age of 15 at less than 63 percent. Meanwhile, only 7 percent of the population has access to the Web according to McKinsey & Co.

In certain areas of India there are construction jobs that go unfilled because the people who would normally qualify for them can’t find them, because they can’t read job ads. But they can speak and probably have a wireless phone. It’s a different kind of digital divide — the gap between information “haves” and “have-nots” — than what we usually talk about in the US, and researchers at IBM are trying to bridge it using a network of what it calls the Spoken Web.

IBM Researchers in India teamed up with the Karnataka Vocational Training and Skill Development Corporation, branch of India’s Department of Labor on a real-world trial of the Spoken Web. Using a combination of cloud computing, mobile phones and voice recognition technology, people looking for jobs will be able to call into so called “voice sites” — think Web sites that you navigate by voice over the phone — to search for jobs that match their skills.

“The whole idea is to apply some advanced technology to a problem that the government of India has been trying to solve: Help people get trained as well as get them matched up with job opportunities,” says Paul Bloom, IBM’s CTO of Telecom Research.

IBM created a new technology called HyperSpeech Transfer Protocol (HSTP) for navigating by voice. which is described in excruciating detail in this academic paper from 2007. (PDF). Initially the project will run in two districts, Mandya and Bijapur, and will be expanded further.

For now there’s no direct link between the voice sites in the project and conventional Web sites, but eventually there will be, Bloom told me. He also said there are lots of other possible applications for the Spoken Web technology including access to the Web for the disable, or people who are injured, or when driving a car.

The video below explains the project in more detail.

Ford Drives Digital Dashboards to Next Level

Walter S. Mossberg — Thu, 07 Oct 2010 01:03:22 +0000

Americans spend vast amounts of time in their cars, where many feel cut off from the digital world. Using a cellphone or digital music player, even in a legal manner, can be clumsy. And a car’s user interface for non-driving functions can seem ancient compared with how other devices are controlled outside the vehicle.

[ See post to watch video ]

So, auto makers have been trying to bring some of the feel of computers and consumer electronics to the dashboard, making it easier to use phones and music players in a safe way, through big screens and voice-command systems that allow the use of these devices without handling or seeing them. Ford placed a large bet on this trend in 2007 with a system called Sync, which I liked when I reviewed it then.

Now, Ford (F) is taking another big step, introducing a second generation of Sync that aims to redesign the entire user interface of the dashboard with color-coded touch screens, better voice recognition and five-way control pads on the steering wheel. This new system redefines the way you control in-car entertainment and climate settings; permits personalization of things like instrument-cluster gauges; and even lets you set up a Wi-Fi network in the car.

I’ve been testing the new Ford system, called MyFord Touch, for a couple of weeks on a 2011 Ford Edge Limited SUV, one of the first two models on which it is being offered. (The other is the Lincoln MKX, with a MyLincoln Touch system.) The Edge starts at $27,000, but the configuration I tested, on which the new interface is standard, lists for $36,000. On other versions of the Edge, MyFord Touch is available as part of a $1,000 option package.

MyFord Touch’s 8-inch touch screen, with function icons in the corners that switch the screen among four main functions: multisource audio entertainment, navigation, phone and climate control.

In general, I liked MyFord Touch, once I got used to it and configured its settings and its connection to phones and music players. The layout of most of the displays is clear and logical, and the voice-command system is still the best I’ve ever used in a car.

But Ford’s new user interface has so many options and functions that I believe it presents a challenging learning curve. Learning the new system can be distracting while driving, at least at first—even though Ford disables some functions while the car is in motion and even though voice commands are easy and plentiful, allowing you to keep your eyes on the road and hands on the wheel.

I urge caution, because this a very different dashboard than you may be used to. I only had the car for a short time, and put very few miles on it, so I can’t say how quickly the new features can become second nature and nondistracting. But anyone buying a car with MyFord Touch should always set up and configure it while parked, use voice commands whenever possible and avoid experimenting with new features and functions while driving. My advice is to learn these in the driveway, gradually.

Instead of the usual array of knobs, dials and passive screens, MyFord Touch is dominated by a giant 8-inch touch screen, with large function icons in the center and color-coded corners that you touch to switch the screen among four main functions: multisource audio entertainment, navigation, phone and climate control. There is also a “home” view, combining common functions that can be personalized.

The system also has several other elements. There are twin 4-inch screens on either side of the speedometer. The one on the left presents vehicle information, such as miles traveled, and allows you to customize some of the gauges so that, for instance, you can finally banish that tachometer you never use in favor of, say, a digital readout on gas-mileage efficiency. The one on the right replicates, in simpler form, the main functions of the center screen, so you can select and check things like audio and climate control without looking at, or touching, the main screen.

These smaller screens are controlled by five-way arrow clusters on the steering wheel, like controllers on iPods and other devices usable by touch alone. There also are some large, touch-sensitive buttons below the main center screen for things like setting volume and fan speed.

Finally, most, but not all, of these functions can be controlled by tens of thousands of available voice commands. And many of these commands can now be spoken without prefacing them with special terms. For instance, you can dial a contact by saying her name at any time, even if you’re not in the phone module on the screen, and even if you don’t first say “phone.”

This voice system worked very well for me, and is the crucial element of reducing distraction. But it wasn’t perfect. For instance, it had trouble with some names in my contact list with multiple entries, and with some streets in the navigation system.

Ford believes the combination of the touch screen, the instrument cluster screens and controls, and the voice commands provides a redundancy and ease of use that should allow both a familiar digital experience and safe driving. Of course, some believe doing anything but driving, no matter how those tasks are performed, is dangerous.

I tested MyFord Touch with an iPhone and an Android phone, which I connected wirelessly; and a standard iPod, which I connected via one of the two USB ports built into the car. I also tested a USB flash drive containing music and a couple of photos, since you can add a personal photo to one of the available views on the big screen. In addition, I tried a USB cellular modem lent me by Ford that creates a Wi-Fi network in the car, presumably only for the use of passengers with laptops and other devices.

All of these devices worked pretty well, but not without issues. The car easily recognized and used both phones for calling, and the iPod generally worked fine. But Bluetooth streaming of music from the phones, which is still an evolving industry feature, periodically failed and never displayed song or artist names. On the physically connected iPod, some album covers didn’t display.

At the moment, only one USB modem, an AT&T (T) model, works with MyFord Touch, and setting it up proved complicated. My general view is that, while operating the touch screen’s main functions is easy, the various setup and option menus are too complex.

The new Ford system can read text messages to you and let you send a limited number of canned responses—the idea being to make texting in a car somewhat safer. But I couldn’t try this as neither of my test phones supported this function.

Ford also has announced that the new system will support some third-party apps, like the Pandora music service, and will eventually have a Web browser for the big screen that would only work when the car is parked. But neither of these features is available yet.

For those who believe doing anything but driving in a car is dangerous, no amount of touch screens, voice commands or redundancy will do. But, for people who would like to enjoy some of their digital lifestyle in a car, MyFord Touch is worth checking out—as long as you take it slowly.

Write to Walt Mossberg at mossberg@wsj.com.

Here's Hoping Google "Speech-to-Speech" Translation Not as Hilariously Inaccurate as Google Voice Translation

John Paczkowski — Mon, 08 Feb 2010 15:43:02 +0000

Google has built up quite a business scanning the written word for contextual advertising opportunities. Now it hopes to do the same for the spoken word as well. The company is reportedly developing a real-time translation technology for our phones.

“We think speech-to-speech translation should be possible and work reasonably well in a few years’ time,” Franz Och, Google’s (GOOG) head of translation services, told the New York Times. “Clearly, for it to work smoothly, you need a combination of high-accuracy machine translation and high-accuracy voice recognition, and that’s what we’re working on.”

A worthy goal, and admirable too, since such a technology, were it to work well, would really herald an end to language barriers. That said, given the hilarious inaccuracy of Google Voice’s visual voicemail transcriptions, I’d say a truly usable speech-to-speech translation service is quite a bit further off than Och claims.

No Hands, All Ears for Sound in Cars

Katherine Boehret — Wed, 13 Feb 2008 00:01:00 +0000

Bluetooth headsets, which wirelessly connect an earpiece with a cellphone to allow hands-free conversations, are particularly useful in cars — especially since, in many states, drivers can be ticketed for using a cellphone without one of these headsets.

To make Bluetooth even easier to use in cars, most new luxury cars and some standard cars are sold with optional built-in Bluetooth speakerphone technology which doesn’t even require an earpiece. Some cars also come with built-in iPod integration, displaying song titles on the dashboard and controlling the iPod using buttons on the steering wheel, again to minimize distractions. One new built-in product, SYNC, the voice-activated system created by Ford Motor Co. and Microsoft Corp., lets users do both things hands-free: play music or make phone calls using simple voice commands.

The $120 Parrot PMK5800 uses voice recognition for a more hands-free experience.

This week, I tested two devices that bring Bluetooth technology to older cars in hopes of integrating hands-free phone calls and music with a car’s stereo system. I tried the $130 Venturi Mini from NextGen Venturi Ltd. (www.myventuri.com) and $120 Parrot PMK5800 from Parrot Inc. (www.parrot.com) on three cars made in 2000 and 2005.

Neither device offers a surefire solution; each is limited by your particular car and cellphone. But the Parrot sails ahead of the Venturi Mini by using voice activation for phone calls, something the Venturi Mini can’t do. Venturi’s version of “hands-free calling” requires initiating a call on the cellphone itself or by looking down at a tiny, grayscale screen and painstakingly scrolling through names of contacts. Furthermore, the Parrot worked after only a few steps, while the Venturi took much longer to set up and get going.

Both of these are one-piece black gadgets that plug into your car’s cigarette lighter and use FM transmitters to play on unused FM radio stations. Each has a tiny screen, though the Parrot screen is used solely to display the current station so as to match it with the radio. The idea is that after initially “pairing” a Bluetooth cellphone with one of these devices, the phone and device will automatically find each other whenever both are in the car and on, making calls easier and music a bit more hands-free.

The Parrot and Venturi Mini will only play music via Bluetooth using cellphones that have a technology called A2DP, which enables music streaming. More and more new cellphones have this technology, such as the Nokia 6555 that I used, but many — including Apple’s iPhone — don’t. Most people will play music by attaching an iPod or other portable music player to these devices using cables that come with them.

The $130 Venturi Mini doesn’t work with voice commands.

Neither the Venturi Mini nor the Parrot PMK5800 enable voice commands with music: songs streamed via Bluetooth are controlled using buttons on the devices, and music coming from a wired-attached player can only be operated using buttons on that player.

The Parrot isn’t as compact or as stylish as the Venturi Mini, but its best feature is unseen: built-in voice recognition software that guides you while using this device. If your cellphone has built-in voice-recognition software, which many do, you can plug in the Parrot and get started. Otherwise, contacts must be loaded on to the Parrot and assigned a voice tag. Large green and red buttons initiate or end phone calls, but speaking commands also works. A knob turns to different stations or can be pressed and turned for audible descriptions of menus. Three glowing Play/Pause, Skip Ahead and Skip Back buttons are easy to find without looking down so as to navigate through music.

While I was listening to music coming from one paired cellphone, my sister called me on another paired cellphone that I had forgotten I had in my purse. The music automatically paused, and the sound of a ringing phone came from the car speakers until I answered it by pressing the Parrot’s green button (speaking the word “phone” also works).

If voice tags are assigned to contacts in your phone, the name of the person calling can be announced over the speakers, like caller ID. Music automatically re-starts after a call ends.

I made calls on the Parrot by pressing its green button and speaking directions like “Call Allison Mobile” to call the correct number from my phone’s contact list. The voice recognition sounded a bit robotic, but almost always found the right number.

Most voice calls sounded rather clear to the people with whom I spoke, but in the car, calls suffered when stations were interrupted with static as I drove around the Washington, D.C., area. One major issue with relying on FM transmitters in major cities is the small number of unused radio stations. Static also affected streamed music, making it sound scratchy at times. Music playing from a cord-connected iPod had no trouble.

The same static problems arose with Bluetooth calls and music on the Venturi Mini. This device’s rectangular face has a scroll button in its center, which seems like it would be a useful addition. But this can’t be pressed down to select anything on the screen, which is maddening. Instead, selecting anything from the Venturi screen must be done using a separate button, as if it wasn’t even designed to be a hands-free device.

To set up the Mini, a paired phone’s contacts must be copied from the phone onto the device. Once these contacts are added, calls can be initiated through the device by finding the correct name on the screen using the scroll wheel and pressing more buttons to select that name and place the call. None of this involves voice recognition, and it’s all supposed to be done while you’re driving.

For all its faults, the Venturi does have a few features that the Parrot doesn’t, including the ability to display Bluetooth data — like the name or number of an incoming caller and a song title and artist — on your radio display if your car has this ability. But most older cars don’t allow this, and I couldn’t quickly figure out how it worked even while driving in a 2005 car. You can also charge devices through the Mini using a built-in USB port.

The position of a car’s cigarette lighter matters to the Venturi Mini and the Parrot. This plug is often positioned near the gear shift, and in my manual car, it would’ve been difficult to operate these gadgets while in fifth gear (I didn’t try). The location of this plug also determines how loud or soft your voice sound to callers. So as to not sound so far away during calls I tried to lean closer to the devices, but this isn’t safe while driving.

The Venturi Mini looks like a hip device, but without voice recognition software and a smart interface, it’s frustrating and dangerous to use. Parrot’s PMK5800 plugs in and works and its voice-recognition software makes it a true hands-free device that will improve the way you use Bluetooth in your car. Just look out for static, especially in big cities.

Edited by Walter S. Mossberg

Email mossbergsolution@wsj.com

Guess 'Puddin' Head Media' Was Already Trademarked, Huh?

John Paczkowski — Mon, 24 Sep 2007 13:06:09 +0000

ThePudding uses breakthrough technology that makes your conversations fun and interesting.”
–Pudding Media press release

As business models go, this is surely one of the most ill-conceived to come across the decks in some time. This morning Pudding Media announced the beta of “The Pudding”–an advertising-supported Internet phone service that serves up contextually relevant ads to subscribers based on their conversations. How? Pudding’s voice recognition software eavesdrops on calls and selects ads based on what it hears. A conversation about the NSA wiretapping scandal might, for example, serve up a promotion for AT&T. A chat about Pudding Media itself might display an ad for a commemorative edition of George Orwell’s “1984.”

“We saw that when people are speaking on the phone, typically they were doing something else,” Pudding CEO Ariel Maislos (who apparently spent several years doing intelligence work for the Israeli military) told the New York Times. “They had a lot of other action, either doodling or surfing or something else like that. So we said, ‘Let’s use that’ and actually present them with things that are relevant to the conversation while it’s happening.”

Interesting idea, but as Silicon Alley Insider notes, one that’s riddled with problems:

“Why will Pudding Media flop? Four reasons:

Free or nearly free phone calls are already available from dozens of companies, including Skype. The difference between two cents a minute and zero cents a minute (for SkypeOut users) is negligible.
Except for people trying the service to see how relevant the targeted ads are (of which there will be many, especially after the NYT article), users will be absolutely freaked out by the idea that someone/something is listening to their phone calls.
Advertisers will be absolutely freaked out by the potential consumer backlash.
PC-based calling is a niche market to start with.”

Guess 'Puddin' Head Media' Was Already Trademarked, Huh?

John Paczkowski — Mon, 24 Sep 2007 13:06:09 +0000

ThePudding uses breakthrough technology that makes your conversations fun and interesting.”
–Pudding Media press release

Interesting idea, but as Silicon Alley Insider notes, one that’s riddled with problems:

“Why will Pudding Media flop? Four reasons:

Free or nearly free phone calls are already available from dozens of companies, including Skype. The difference between two cents a minute and zero cents a minute (for SkypeOut users) is negligible.
Except for people trying the service to see how relevant the targeted ads are (of which there will be many, especially after the NYT article), users will be absolutely freaked out by the idea that someone/something is listening to their phone calls.
Advertisers will be absolutely freaked out by the potential consumer backlash.
PC-based calling is a niche market to start with.”

Hate Voice Mail? New Services Turn Recordings Into Text

Sarmad Ali — Thu, 24 May 2007 00:01:00 +0000

Punching in a password and listening to voice mails on hand-held devices is inconvenient — or rude if you’re in a meeting, at a party or in a restaurant.

A slew of new services using voice-recognition technology aim to eliminate the hassle of checking voice messages on wireless devices. These services transcribe recorded messages into text, which then is sent as email to email-equipped devices such as personal digital assistants and BlackBerrys, or as a short text message to phones that don’t have the email function. If in doubt about transcripts’ accuracy, you can always click on the attached audio files or dial in to hear the original voice mail.

To see how efficient these services are at transcription, I tested two voice-recognition applications: one from start-up SimulScribe, based in New York, and another from Atlanta-based SpinVox, a subsidiary of SpinVox Ltd. of the U.K.

Overall, both services work pretty well and are easy to use. I was able to read transcribed texts in a fraction of the time I would have spent dialing in to hear them. I was also able to sift through my messages and go directly to the ones I wanted to check, as opposed to having to listen to every single one sequentially. And the transcripts end the hassle of having to jot down names, numbers or addresses.

Another advantage is that the transcribed messages are sent immediately after the voice mails. When I was on a train and passing in and out of reception areas, I got my written messages faster than if I had tried voice mail only. Users can forward the messages to others or reply by calling back, sending an email or text messaging.

Signing up with SimulScribe.com took just minutes. After setting up an online account naming a cellphone carrier, I got a confirmation email with instructions on how to activate the service on my phone. SimulScribe, launched this past September, costs $9.95 a month for 40 transcribed messages plus 25 cents for each additional message. The service is compatible with all wireless carriers.

Currently, SimulScribe transcribes English voice mails only. A voice mail left on my phone in Spanish wasn’t transcribed at all. The company says it’s testing a Spanish system to add to the service in a couple of weeks.

Activating SpinVox’s Spin-my-Vmail on the phone was easy, too. After signing up on spinvox.com, the company sends subscribers a guide to using the service on the phone. I was able to activate “call forwarding busy/no answer” to divert messages to my SpinVox voice mail just by changing a setting in my T-Mobile’s cellphone settings menu.

I received an email in my mailbox whenever someone left me a voice mail. Each had the number of the caller, the transcript, date and time the message was received, plus a message code I could type into the keypad after accessing my voice mail to hear the original audio.

U.S. customers can get a free one-year trial by sending an email to gamma@spinvox.com, during which time the company expects to launch the service in the U.S. first through Cincinnati Bell and later through other carriers. Bloggers can also test Spin-my-Blog, which lets users speak a posting to their blogs from any phone. Sign-up for that is also at spinvox.com.

SpinVox, first launched in the U.K. in May 2005, transcribes in English, French, Spanish and German. Unlike SimulScribe, messages left to me in Spanish were successfully transcribed.

One friend who left me a voice mail on both services said she was pleased to be reminded by SimulScribe that her message would be transcribed so she should speak slowly and clearly.

Users of SimulScribe get unlimited inbox storage so they don’t have to delete old mail. Both services work better with hand-held email devices such as Treos or BlackBerrys than with the cellphones that don’t have the email capability. The number of characters that can be transcribed into SMS text is limited.

With SimulScribe, long messages delivered by SMS are parsed over multiple text messages. The same happens when customers use SpinVox on CDMA cellphones. Customers using SpinVox on GSM phones like those from Cingular or T-Mobile fit three-minute calls on one text.

Sometimes, the transcriptions contain misspellings, missing words or unnecessary punctuation marks. A friend left me a voice mail on my cellphone with SpinVox’s Spin-my-Vmail service. She ended it by asking me if I was sick of Thai food, but the transcribed note, amusingly, turned it into: “Hi Sarmad, it’s Kain(?). I’m calling at 4:09(?). I just wanted to see what the plans were for tonight. Are you interest in dinner, are you up for Lasagna(?).”

Words — mostly names — spelled phonetically, some numbers and undecipherable words are usually followed by a question mark.

Aside from sporadic imprecision, I liked SimulScribe more, mostly because it eliminates the need to dial in any passwords to get a voice version of transcribed messages. But it’s hard to beat the free trial of SpinVox and its multiple-language transcription capability. Both services are a nice addition to hand-held devices if you can overlook a few nuisances.

Email me at sarmad.ali@wsj.com. Walt Mossberg is on vacation.

Securing a Wireless Network

Walter S. Mossberg — Thu, 15 Mar 2007 00:01:40 +0000

There’s no other major item most of us own that is as confusing, unpredictable and unreliable as our personal computers. Everybody has questions about them, and we aim to help.

Here are a few questions about computers I’ve received recently from people like you, and my answers. I have edited and restated the questions a bit, for readability. This week my mailbox contained questions securing a wireless network, adapters that play the music stored on a PC through a stereo and voice-recognition software.

Last week you advised readers that, in order to stop people from piggybacking on a wireless network using a Linksys router, they should simply set up a password and keep it private. But don’t they need to enable an “encryption key?”

Yes. Most nontechnical folks would consider an encryption key as a kind of password, and that’s the word I used because I always try to write my columns in plain, conversational English. In this case, however, my use of the term may have caused confusion, because there are, in effect, two kinds of passwords on Linksys routers and most other brands of routers. One just prevents strangers from changing the router’s settings. The other — the one to which I was referring — is required to actually access the wireless network. That’s the one that’s technically called a “key.”

To enable the encryption key, use the router’s setup software to turn on security. On newer models, the strongest security system is called WPA, and on older models, it’s called WEP. Once it’s enabled, only people who know the key can get onto your network. There are further steps you can take, like hiding your network’s name (called an SSID) from others, or even restricting access to the network to specific computers with specific identification codes (called MAC addresses, a term that has nothing to do with Apple’s brand of computers.)

For more details, go to Linksys.com, select “Learning Center” at the top of the page, and click on “Network Security” from the menu that appears. And then click on the link called “How to Secure Your Network.”

Is there a stereo that I can buy that can wirelessly connect to my computer and play the same music I play with Apple’s iTunes on that computer?

Yes. If you literally mean a stereo, rather than an adapter for a stereo, the best I know of is called Sonos ZonePlayer 100. It involves a module you connect to your computer that links wirelessly to stereo units with built-in amps in remote rooms of your house. It has a beautiful remote control with a color screen and many other great features.

But it’s expensive — about $1,200 for the starter package, without speakers — and can’t play copy-protected music you buy from iTunes, only music you import from your CDs or other unprotected files. Information is at sonos.com. For much less money, you can also buy adapters that can play the music on a PC, or an iPod, through an existing stereo.

I have been keeping a journal or diary. It is hand written but I would like to make the journal more readable by others. Is there voice-recognition software that you would recommend, so I can dictate the entries?

It has been years since I reviewed voice-recognition software, but Dragon Naturally Speaking, a Windows program, works well. For more information, see nuance.com. And Microsoft’s new Vista operating system has a decent built-in voice recognition system. If you are using a Macintosh, one widely known speech-recognition program is called iListen, but I haven’t tested it. Information is at macspeech.com.

* * *

Because of the volume of e-mail I receive, I can’t routinely answer individual questions by e-mail, or consult on individual problems or purchasing decisions. I read all questions I receive and select three each week to answer in the column.

Write to Walter S. Mossberg at mossberg@wsj.com