Voice Recognition and Vista?

[alexander]

Well those two may not be the best combination you can imagine... yeah a superbloat and voice recognition software that is wicked tempermental in a demo.... check this out:

http://gizmodo.com/gadgets/windows/windows-vista-voice-recognition-crashes-and-burns-191020.php

 

lol... all i have to say

[Turtle]

lol... all i have to say

 

Sorry, I didn't quite hear you. Select all...delete. Please speak clearly into the microphone & forbid laughter. :eek2: Looks like this technology of voice recognition has finally caught up to the OCR technology.:lol:

[Queso]

OCR is so much fun on my mac.

 

I send all my poetry thru various voices, over abstract and ambient music.

 

It's hilarious, quite like that video.

[Jay-qu]

cant say I am that surprised, voice recognition on anything I have used has always been tempramental..

[alexander]

dude, my little sister can prolly write a peace of soft that is better then that, and she doesnt even know what programming is....

[michael1384]

Well there's a suprise. I always knew there was going to be something about that voice recognition.

[alexander]

lol, i like the part where the woman goes: microsoft is very upset that we show you this video, they claim the reason the software failed was noise in the room, but if you heard it was really quiet in there until the malfunction occured

[Mercedes Benzene]

Are there any effective voice-recognition programs out there that you guys have come across. There's one that our school computers have, called Crystel (or something like that), and it works better than that Vista crap does in that Video.

[C1ay]

http://ars.userfriendly.org/cartoons/?id=20060807 :steering:

[Mercedes Benzene]

http://ars.userfriendly.org/cartoons/?id=20060807

 

Hahahahahahhaha! That's super funny. :steering:

[alexander]

lol, thats awesome :thumbs_up

[CraigD]

Are there any effective voice-recognition programs out there that you guys have come across. There's one that our school computers have, called Crystel (or something like that), and it works better than that Vista crap does in that Video.

I used IBM ViaVoice and Dragon NaturallySpeaking on personal and biz Win16 (Windows 3.1) and unix machine. It worked fairly well. The voice recognition software included in Microsoft office since 1997 does a fairly good job of voice to text translation. I just dictated the previous sentence using it with only a single error. However, my anecdotal impression is that there’s little practical improvement between voice recognition over the past decade.

 

I’m inclined to believe Microsoft’s claim that the Vista voice recognition software was confused by background noise.

 

Nonetheless, despite early advances in voice recognition software, the state of the art in commercial software is less advance than one would expect. For example, when I spoke the preceding sentence, Office entered :”Nonetheless despite a rally advances in voice recognition software the state of the arms of new areas and since one would expect.” WTF indeed.

[alexander]

I’m inclined to believe Microsoft’s claim that the Vista voice recognition software was confused by background noise.

but if you listen closely it was really quiet in the room until the software started malfunctioning.... :D

[CraigD]

but if you listen closely it was really quiet in the room until the software started malfunctioning.... :D

When I listened, I noticed that the hapless product demonstrator was speaking into a headset boom mike – which should be pretty discriminating of outside noise, but that his voice is also being broadcast over a PA, either from his headset mike, or another one out of the picture. He’s standing in front of – and, one assumes, to the side of – some non-sound absorbing-looking stuff. It seems a perfect setup to produce severe echos.

 

Upon listening to the youtube video again, I noticed the demonstrator’s small, dismayed chuckle when the “double the killer delete select all” appears on the screen. I can so relate.

 

What’s appalling to me is not that the SR software so badly mishandled the audio input, but the lack of any context awareness. We humans perceive – and sometime utter – very wacked-out spoken words, but through the miracle of context awareness, know that a speaker simply can’t be saying “the killer” in the context of “aunt”, “double the”, or any default situation. Every commercial SR system I’ve seen fails to implement anything but the most rudimentary context sensitivity.

[alexander]

so you are left to write your own? :D

[C1ay]

What’s appalling to me is not that the SR software so badly mishandled the audio input, but the lack of any context awareness....

It would be interesting to run this little spell check verse through it :secret:

 

Eye halve a spelling checker,

It came with my pea sea.

It plainly marks four my revue

Miss steaks eye kin knot sea.

 

Eye strike a key and type a word

And weight four it two say

Weather eye am wrong or write

It shows me strait a weigh.

 

As soon as a mist ache is maid

It nose bee fore two long

And eye can put the error rite

Its rare lea ever wrong.

 

Eye have run this poem threw it

I am shore your pleased two no

Its letter perfect awl the weigh

My checker tolled me sew.

[alexander]

:beer: my spell checker said "I bet this will check out fine with Word or aspell" :secret: