backtop

Print E-mail del.icio.us 33 comment(s) - last by cfaalm.. on Nov 3 at 9:16 AM
Speech recognition will be as big as touch capability in the future reckons Microsoft

Microsoft has been touting touch enabled computers in a big way with the build up and launch of Windows 7. Now that Windows 7 is here and touch capability is ingrained into the user experience of many computers and smartphones, Microsoft is looking to the next big thing in interface design.

Microsoft's Zig Serafin said, "Voice is the new touch. It's the natural evolution from keyboards and touch screens. Today, speech is rapidly becoming an expected part of our everyday experience across a variety of devices."

Microsoft has reportedly been developing speech technology alongside its touch technology for over ten years now. The software giant beefed up its speech recognition development in 2007 with the purchase of Tellme Networks. Technology developed by Microsoft in speech recognition has now been integrated into Bing for Mobile, Windows Mobile 6.5, and Windows 7.

EWeek reports that an offshoot of speech technology has also been built into Exchange Server 2010 allowing a text preview of a voice mail message to be sent to a user. HP has integrated some of Microsoft's voice control technology into its TouchSmart PCs reports eWeek with an app that is called TouchSmart RecipeBox that allows chefs to navigate recipes with their voice so they don’t have to touch the screen with dirty hands.

Serafin said, "The climate in our R&D environment is optimally charged to accelerate advances, leverage the power of software plus services, and revolutionize the ways customers interact with a wide range of Microsoft products."

Voice technology certainly holds promise for both the PC and mobile phone markets. With the big push to get people in the U.S. to stop texting and driving voice technology would allow safer texting while behind the wheel. Accurate voice recognition technology on the computer could also mean faster typing for users and easier computer use for those with disabilities.


Comments     Threshold


oh no...
By UltraWide on 11/2/2009 9:59:18 AM , Rating: 3
I hope I never have to talk to my computer.




RE: oh no...
By Statzere on 11/2/09, Rating: 0
RE: oh no...
By Mitch101 on 11/2/2009 10:49:53 AM , Rating: 4
I remember doing stuff like this maybe 10-12 years ago with a creative sound card. You could control a lot of windows functions with your voice.

I also have a program that I cant recall the name but would allow you to configure keyboard shortcuts for gaming to voice commands that worked great. In unreal I would say Nuke and everyone at the lan party would hear me and immediately think I had the nuke. Instead It would feign and when they walked by I would pop up and gun them down. Was much better than trying to cycle through keystrokes.

I have yet to use the voice options in Windows 7. I hear its actually very good.

http://www.microsoft.com/windows/windows-vista/fea...


RE: oh no...
By The0ne on 11/2/2009 12:43:49 PM , Rating: 2
Reviews I've read also said MS voice is good. I've yet to test it myself though. Could be interesting if it is indeed accurate :)


RE: oh no...
By BladeVenom on 11/2/2009 4:02:00 PM , Rating: 2
RE: oh no...
By splint on 11/2/2009 4:07:04 PM , Rating: 4
I just upgraded to Win7 and played around with Voice. After going through the ungodly long voice recognition training session here is what the result was:

(dictating in Microsoft Word 2007)

Start of the station I feel several and they correct several saw wholly fit god nine one so far is wrong with this speech recognition everything I say, that is to encourage it's like a retarded child on revenue and trying to dictate what I say 1 to 5! Holy spirit everything I say is completely wrong off. Let's try this again
I'm trying to take a on this 5:00 AM computer. That's not what I was trying to say. I am trying to dictate on this god damn computer! That previous sentence took about 10 minutes.


RE: oh no...
By GeorgeH on 11/2/2009 6:31:04 PM , Rating: 5
Just did a quick run-through myself (using the admittedly poor built-in microphone on my laptop.) Below are the various interpretations of the phrase "Die, you filthy Nazi scum."

Diet you from the NAZI scum
That you should be Nazis gum
Die you feel the Nazis go
Got you felt the Nazis come
Die you fill the Nazis come
Diane you feel the Nazis come
Die you through the NAZI scam
Guide you through the top he’s come

Saweet. I then went through a few more "training sessions", and got these results:

Got you feel the NAZI scout
Dying for the NAZI storm
Cut you feel the NAZI scout
Guy you fill the NAZI scum
Daiei you filthy NAZI scout
The IE you to the NAZI stone
Die you felt the at the start
Guy you filled the Matisse,

So we move from dieting to Diane in a porno followed by a Japanese supermarket and a little fine art. Even better, when I tried to queue up each word individually it wouldn't accept the word "Die". Why? Because that "Command cannot be performed."


RE: oh no...
By Spivonious on 11/2/2009 12:27:34 PM , Rating: 2
The article says that MS has been working on voice recognition for over 10 years.

Also, I don't think voice interaction with mobile devices will ever be big. It's so much easier to push a button.


RE: oh no...
By grath on 11/2/2009 7:50:29 PM , Rating: 2
quote:
Also, I don't think voice interaction with mobile devices will ever be big. It's so much easier to push a button.


I beg to differ. While there will always be functions more easily accomplished through a button press or two, more complex functions can definetly make good use of speech recognition.

Voice dialing is the bad example and probably what youre referring to, and it has been around a while. "Call Mom" is easily accomplished by one or two presses.

Googling something on the other hand requires (on iPhone): physical button press to wake, slide to unlock, press Safari, press Google text box, type search phrase, press Search.

So thats 5 actions in addition to typing, and some of the steps can have a delay of a few seconds between them.

Compared to: physical button press, say "google how to jailbreak"

Assuming the recognition is sufficiently accurate, how can that not be a popular feature?

Navigation and search in the Maps app would be an even better use of recognition, as the Maps app is not exactly the most convenient thing to operate sometimes, especially while driving.

And texting/emailing of course, thats the ultimate nobrainer there.


RE: oh no...
By StoveMeister on 11/2/2009 9:34:40 PM , Rating: 2
Man I can just imagine trying to whisper into the search engine "mumblemumble goatse.cs mumblemumble". Wonder what sort of results THAT would get :)


RE: oh no...
By cfaalm on 11/3/2009 9:16:55 AM , Rating: 2
while you have absolutely no idea what your computer looks like by that time?


Voice recognition is useful, but gimmicky
By Yawgm0th on 11/2/2009 10:14:36 AM , Rating: 2
I can type a word or command faster than my brain can think of it and much faster than I can say it. I've tried voice recognition in the past on Word processors, and honestly it's not that bad. By now, I have no doubt the technology has progressed to the point at which we can see more serious support for it. Still, I don't see it realistically getting used very much in most applications.




By Digimonkey on 11/2/2009 10:26:09 AM , Rating: 2
They can slowly integrate it into the OS. I can see it being useful for things that usually take a few mouse clicks to get to. Like it'd be nice to tell your computer to just change the display resolution or IP address to what you say.

Also you could say email to such and such address/name and it'll open up your e-mail client with the compose screen and e-mail address in the to field. It seems like it could save some time in the right scenarios.


By nafhan on 11/2/2009 11:12:03 AM , Rating: 2
It's not that it's gimmicky; it's just being oversold. I would prefer a combination of voice control, a good touch screen, and a physical keyboard on the same device. Which control method you would use would be based on what you're doing in the real world and and what you're trying to do on the device. I can see the ability to switch between control methods or even use all three at the same time being desirable and useful.


RE: Voice recognition is useful, but gimmicky
By rudy on 11/2/2009 3:41:12 PM , Rating: 2
First of all you are probably rare and if you concentrated on speaking you could probably speak much faster then you can type.

Speech and touch are important but to me speech is more important the combination of all of them gives us the ability to overcome repetitive strain injuries.

I personally would love my computer to be able to accurately type what I say without alot of training and that is the part left out what M$ really means is that speech will be automatic much like mouse countrol you can just sit down at any device and do it. Currently it requires alot of training.


By lyeoh on 11/3/2009 3:46:11 AM , Rating: 2
1) You'll just be shifting the strain from your hands to your vocal chords.

2) While you might be able to do a faster "word per minute" when speaking, the correction and re-editing bit might more painful.

To me it all depends on what you need to do. Pianists can play more notes than they can hum, but you can't easily sing with a piano.

"Multichord" keyboards might allow people to type faster, but most people probably can't think that fast anyway.

And how fast can you talk and still be intelligible to a computer?

:) http://www.youtube.com/watch?v=NeK5ZjtpO-M


Hmmmmmmmm, so voice texting
By Turas on 11/2/2009 11:06:48 AM , Rating: 5
Why not just call them instead.




RE: Hmmmmmmmm, so voice texting
By Yawgm0th on 11/2/2009 11:42:42 AM , Rating: 1
QFT. I use SMS and email so I don't have to talk to people.


Chicken coop
By BernardP on 11/2/2009 10:26:25 AM , Rating: 2
Open floor office = dozens of people speaking to their computer = noise level of a chicken coop.




RE: Chicken coop
By JediJeb on 11/2/2009 1:52:59 PM , Rating: 2
True, the average office will now sound like a telemarketers phone room.


Not now, maybe in 30 years time
By AnnihilatorX on 11/3/2009 4:24:27 AM , Rating: 2
Speech recognition will never take off until AI with significant intelligence comes. Face it given ambient sound and noises there is no way, despite using how much training, to attain 100% accuracy. Not many people like to sit through long session of training just to improve accuracy from 40% to 90%, that is without ambient noises.

On the other hand, this will never work on corporate environments, so again, application is limited.




By exploderator on 11/3/2009 8:06:47 AM , Rating: 2
... so again, if sadly, it's a failed idea and a waste of time. And it would be a sad waste of AI too, forcing it to listen to our pathetic human drivel.


Computer...
By Spivonious on 11/2/2009 9:51:40 AM , Rating: 3
Where is Captain Picard?




Helloooooo computer!
By Fenixgoon on 11/2/2009 9:59:20 AM , Rating: 3
Why didn't someone tell me that computers in the 20th century weren't voice controlled when I tried to make transparent aluminum? You could have saved me from a little embarrassment there, Captain Kirk!

</Scotty from ST IV>




Yaye!
By judasmachine on 11/2/2009 10:07:19 AM , Rating: 3
Now they will have to sound proof the cubicles.




Still just a gimmick!!
By HighWing on 11/2/2009 7:08:27 PM , Rating: 2
Lets face it, as great as touch is, it will never replace the mouse and keyboard on a desktop computer. And neither will speech! It might help them evolve into something different from what we have, but neither will ever replace them.

And if your not convinced, get a HP myTouch desktop PC and spend an hour or more browsing the web without the keyboard and mouse, by just using your hands. After awhile your hands will get tired and you won't want to do it anymore and go back to the keyboard and mouse.

Touch is only the current "Gimmick" and Speech is the next.

As many have already pointed out there are several speech software titles out there already. And as cool as it is to "tell" your computer to do something, it will never be faster then touch or using a mouse. In fact many things are often slower if you were using speech alone. Try it for yourself and see. Plus, it would never do well in the workplace!

Neither of these features are anything new, and at the moment neither seems anywhere near being a standard. I would be very surprised if the number of touch screen windows 7 pc's is anywhere close to 50%. And I would bet money that it will never get to 100% in the next few years.

It's just another gimmick that will have a niche, but in the end will never really be used as much as it is claimed!




RE: Still just a gimmick!!
By exploderator on 11/3/2009 7:51:32 AM , Rating: 1
Amen. That's exactly what I thought when I read "Speech is the Next Touch According to Microsoft". The next pointless marketing gimmick.

Touch is retarded, and voice is only good for focused dictation. Can you imagine trying to teach someone, or show them something, when your computer is trying to do what you are telling them?

"Did you hear about that virus that deletes all the files on the C drive?"
... computer prompts on screen "are you sure yes/no"
... your friend answers you "YES, I heard it was a bad one."
... you screaming... CANCEL STOP UNDO FUCK AGHHHH!

Or, "Hey, grab me a beer?"... Command cannot be executed. Fuck.

Dunno about you folks, but I live in a world with other people, and we people talk. I can use a mouse (almost silently) while I'm doing that, too. This just doesn't fit the real world, unless you are locked in the basement, or carefully dictating in a carefully chosen environment.

This entire initiative, as with touch, is purely marketing fucktards pushing a bling bling sci-fi fantasy. Totally disconnected from the mundane practical reality of actually making good peripherals that actually work well. But that's no surprise from MS, they have always been utter tards when it comes to ergonomics and advanced usability. Their focus is people who can barely use computers, and they do a pathetic job at that. I think the general approach is "automatic", but when that only works 50% of the time, it's a total fail for anyone who expects a computer to actually work reliably.

Frankly, I just want a computer to do exactly what I tell it, and I want that method to be as precise and unambiguous and efficient as possible. Neither touch or speech come anywhere near close. I spent $100 on a Logitech MX Revolution mouse, and I would gladly pay twice that if it was truly a precision built machine with metal and bearings, instead of consumer grade crap kicked up a very small notch.

There are many many possible improvements to be made, but sadly they are all mundane, and nobody is willing to really work on them outside of extreme niches like big $$$ robotics control interfaces. And meanwhile, though I'm glad that brain surgeon has a great amazing $100k 3D whizbang, billions of real human hours are fettered away with shitty OS & UI, and bottom quality cheap plastic crap peripherals. MS and Apple do touchy feely mumbly jumbly. And still we spend all day every working with mediocre tools.


It's been around >.>
By The0ne on 11/2/2009 10:41:50 AM , Rating: 2
Dragon Naturally Speaking. Great program and one of the better ones. I followed and still following these tech since my college days. Unless you have something wrong with your speech then I think you're covered :)




bah
By Randomblame on 11/2/2009 2:49:31 PM , Rating: 1
Kids hate talking these days, text messaging has retarded the speach centers of their brains. At this point they probably feel too socially inadequate to even speak to a computer. Now if they could come up with a way to txt a computer and command it it would be a hit - I see a resurgance in dos use




RE: bah
By exploderator on 11/3/2009 8:03:27 AM , Rating: 1
Fucking brilliant. I love it. And totally true and realistic, no matter how deeply pathetic and ironic that is.

I can see it now, "Critter"?. Teens txt controlling from their cell phones. Idea even has some legitimate uses, for remote file access to home office computer blah blah blah, but it's the teens that will bring it on, to send off pics and emails and mp3's (kinda like extending cloud systems to include home PC control). hahahahaha txt (doslike) control of facebook and myspace too. hahahaha


By TomCorelis (blog) on 11/2/2009 6:26:25 PM , Rating: 2
And will it not suck for those of us with speech impediments?




By lecanard on 11/2/2009 8:03:22 PM , Rating: 2
I was afraid they had forgotten about speech recognition. They made a good effort to implement voice controls in windows vista. However, they didn't update it in windows 7, and the voice commands aren't really designed for the different UI, so it doesn't work very well.




By DOOA on 11/2/2009 10:21:43 PM , Rating: 2
So long ago we controlled our computers with voice. It worked well if you had a quiet room or used a headset. But then something happened to interest in the technology. Perhaps it was Dragon Naturally Speaking making money so the OS freebies were no longer developed?
Having the computer read back text us is very useful for reports. The ][gs had very good sound, so perhaps the technology stagnation was due to a lack of hardware?
I, for one, welcome its return - especially in gaming. Reliable voice commands + keypresses vs keypresses alone? Bring it on, I know who the winner will be.




"A politician stumbles over himself... Then they pick it out. They edit it. He runs the clip, and then he makes a funny face, and the whole audience has a Pavlovian response." -- Joe Scarborough on John Stewart over Jim Cramer









botimage
Copyright 2009 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki