It's Sunday afternoon here, there is no better time to spend by having some fun with some of the [in]abilities of operating systems. There is definitely quite some rocky road still for speech recognition and in extend natural language & artificial intelligence. Hopefully we will still be alive when that happens.
Post a Comment
Member since:
2005-11-23
Member since:
2006-01-02
While the guy was probably trying to take the micky out of the Voice Recognition. I think it handled really well.
He purposely made stupid noises, so of course the Voice Recognition is going to pick them up. Towards the end his only THEN decided to 'turn microphone off' etc...
I'm quite impressed to be honest, nothing as good and as integrated as that exists in the Linux world. That's one thing in Vista that is certainly impressive.
Member since:2005-11-23
Member since:2007-02-22
Member since:
2005-10-01
Member since:
2007-02-22
That'll teach some of us to mutter under our breaths at our computer... it's listening now! :>
But then again, this wasn't a total failure -- even when he made a mistake, he was able to use the VUI to fix his errors.
Who knows? Maybe future OSs will listen to how often their users curse and ramble and dumb down -- er, that is, selectively limit the interface appropriately. (Clippit pops up: "I heard you talking about 'finding' something. Are you trying to find a file?")
Edited 2007-03-19 00:07
Member since:
2006-01-07
Member since:2007-02-22
Is this beginning of the end for thousands of languages that happen to be not English?
As most languages out there are phonetic, Vista's speech recognition will be fairly easy to localize. In fact, this could be a boon for non-standard languages, because even if documentation can't be produced quickly or cheaply in various minor languages, it's possible that an oral interface for those languages would work much more naturally.
Member since:2005-07-06
Is this beginning of the end for thousands of languages that happens to be not English? Are we going to build another Babel tower? This time with windows i.e. Windows?
Vista's speech recognition supports English (U.S.), English (U.K.), German (Germany), French (France), Spanish (Spain), Japanese, Chinese (Traditional), and Chinese (Simplified).
Member since:
2005-06-29
http://www.youtube.com/watch?v=S9T3RXFDtfo&NR
good stuff. Just over 1:30 to do it. Not the exact same code, but still good.
Edited 2007-03-19 00:51
Member since:2006-01-02
Dragon's been doing the speech recognition game for longer, so I'm not surprised that they have pretty refined voice commands. Also the guy seemed to have a lot more experience with Dragon than the first guy did.
Anyway, the lesson is that neither speech recognizer is really meant for writing code. I tried writing code with my TabletPC stylus once. I had to quit after taking a minute to write a single line.
Member since:
2005-08-10
Member since:
2006-05-14
Member since:
2005-07-06
Member since:
2006-05-24
Really? This one stupid video is "A Look at Vista's Speech Recognition"? Man what an egregiously lame post. (For starters this was dugg to death a *month* ago - "osnews, you read it here...er, 100th!"). I like osnews because often you will have an anchor link but then will point to a couple takes on a subject and that's what a real "look at Vista's speech recoginition" would do. Show a couple different takes rather than just this one guy who seems vaguely disappointed when it does work. We have no indication that this is anything but the first time he's ever put on a headset. The "response video" doesn't really make much sense because you have to train both and it is clear the dragon dude is very familiar with it and the moron in the first video is just messing around.
Also, it is hardly suprising that a consumer os's speech engine wouldn't be too well suited to the specialized syntax of script and so people much smarter than either of these two are building an engine optimized for code:
http://voicecode.iit.nrc.ca/VoiceCode/public/ywiki.cgi
IMHO Vista's speech recognition is pretty good. As good as Dragon? Probably not but certainly the best that has ever shipped as part of an OS thus far. The grid and numbering system is a tiny boon to handicapped people and great to have it built in. No less than Apple fanboi deluxe David Pogue agrees.
Pogue
http://tinyurl.com/2gpwja
btw- Leopard is also going to also tranlate to Braille devices which is awesome.
Spending all of 2 hours with it Eugenia has no credibilty on Vista in the first place (the same way someone spending a couple hours with linux and declaring it useless isn't to be taken seriously) but why not at least post a video of someone getting different results and have the post be an actual look at the app's charms and pitfalls instead of a silly gag video?
Another goofball trying to program with it
http://www.nxtgenug.net/article.aspx?ArticleID=150
Navigation (tiny bit of dictation) demo in French
http://tinyurl.com/25obht
Member since:2005-07-06
One of the better videos online demoing Vista's speech recognition (though the video's audio isn't the best) is a demo given at SpeechTek.
http://video.google.com/videoplay?docid=5549626370637807497&q=vista...
Member since:2006-10-08
"btw- Leopard is also going to also tranlate to Braille devices which is awesome. "
That's fine, but as long as it is usual to put barriers into the WWW, blind people mostly keep outside of it. More and more "Flash" stuff, no proper image references (img alt, longdesc), frames without proper encalsulation and alternative...
I did some studies in rehabilitational pedagogics, we found the poor quality of content presentation increasing. "Web designers" just don't care.
I'm looking forward ad MacOS X how it will serve the needs of blind people.
(This is off topic, so don't really mind. Sorry, cound not resist.)
Member since:2006-01-03
I agree that YouTube video was "made" to look as poorly useable as possible, but you say its the best of any OS, have you even tried Speech Recognition in OS X?
Without any training I have found it to be quite robust and with applescript it is a snap to make custom speech commands for programs that don't even support speech.
You are right in that OS level speech commands is a totally different beast than dictation. People need to distinguish between products like Natural Speaking and iListen and the OS feature.
BTW, did I mention this post wasn't typed. Speech worked fine for me. IMing without typing is a god send also.
Member since:
2005-08-09
The earlier commments says most of it. But this guy seems to be intentionally making it look worse than necessary by using the same voice non-commands repeatedly: "press caps lock" should be merely "caps lock". Worked the first time for me.
"Doctor, doctor, it hurts when I do this."
Member since:
2005-11-15
Member since:2005-07-06
Actually, the TTS and Recognition engines in Vista (and in other products like MS Speech Server) are based off of the work of MSR and MSR Asia.
http://research.microsoft.com/srg/
http://research.microsoft.com/speech/
Member since:
2005-07-12
IBM bundled both "voice dictation" and "voice navigation" with its OS/2 Warp 4 product back in 1996. Some Warp 4 packages even included a headset microphone.
It's been ten years. How much better is the voice stuff package with Vista? Does it require less training time than the Warp stuff from IBM did?
It would be interesting (at least to me) to see the level of progress that has been made, if any.
