Nice find. Unfortunately, the copy is a vivid illustration of the difference between a copyeditor and a spellchecker. That's the sloppiest job I've ever seen on an Apple page.
A lot of people have been waiting for this tech, though. More accessibility == Good.
Nice find. Unfortunately, the copy is a vivid illustration of the difference between a copyeditor and a spellchecker. That's the sloppiest job I've ever seen on an Apple page.
A lot of people have been waiting for this tech, though. More accessibility == Good.
This may explain the crazy rumor that Apple was going to 'bundle' all its applications into the operating system and allow you to access them by voice control.
Hopefully they'll revamp the whole MacinTalk framework...it hasn't changed much if at all since 1993-4
Yeah. I hate to say it but Micro$ofts speech recognition is a lot more accurate. I was really impressed when someone showed it to me the other day. I am used to getting 90-95% on my Mac with a small vocabulary. The M$ stuff got 100% with the same vocabulary. Apple really needs to update their recognition engine. It is embarassing.
What kind of speech input were you using under Windows? Was it actually *from* MS, or a third party? Was it dictation or discrete speech?
I believe XP has dictation built in. I only know because I was helping a friend track down a strange problem that his friend was having with XP where every so often "..." would appear when he was typing and then random short words like "and", "it", "she" would appear. I thought it was a crazy virus at first but after consulting Google it appeared that Voice Recognition had been turned on by accident.
You do realize that MacInTalk is speech *OUTPUT* right? It's got nothing to do with the speech recognition engine.
Sorry, I guess I don't know my Speech API terminology. Whatever you call the speech recognition component, it doesn't seem to have been updated since the mid 1990s and it shows.
Quote:
What kind of speech input were you using under Windows? Was it actually *from* MS, or a third party? Was it dictation or discrete speech?
The guy who was showing it to me was doing some programming using the interface. He referred to it as "Microsoft SAPI" and when I asked him what that stood for he said it was Microsoft's speech API. We are really only interested in discrete speech. We were testing on just a four word vocabulary, "Red", "Green", "Blue", "Yellow". I speak pretty standard American English, so I wasn't exactly tricky. Microsoft got 100% on me, and over 95% for 15 out of 16 random people we tested (most were essentially perfect; the other one was about 90%), with no training.
No, really, the smaller the dictionary of words to look for, the easier the problem is.
That sounds about like my experienced %s with the Mac speech recog. I've got almost no problems with it, my advisor has it work for him *most* of the time. (He's a mushmouth though.)
And, just like your experience with SAPI, no training needed there either.
I've noticed that people either have almost no problems with the Mac speech recog, or it drives them nuts. I think it's based solely on how closely your speech matches the 'expected norm'. The SAPI 'expected norm' may be calibrated closer to your voice.
Asked, whether Apple would introduce a new version of MacOS X at this year's WWDC, Apple VP Ron Okamoto said:
"In typical Apple fashion, we will not comment; however, as I said, it is our opportunity to talk about our technology direction, so you can bet we will be doing that."
Asked, whether Apple would introduce a new version of MacOS X at this year's WWDC, Apple VP Ron Okamoto said:
"In typical Apple fashion, we will not comment; however, as I said, it is our opportunity to talk about our technology direction, so you can bet we will be doing that."
It does seem odd that Apple would start a campaign about features to be added to the NeXT version of Mac OS X now if it is 9 -12 months away from fruition? Any comments or ideas?
Comments
A lot of people have been waiting for this tech, though. More accessibility == Good.
http://www.businessweek.com/technolo...6454_tc056.htm
as long as they don't use the "Trinoids" voice.
Originally posted by Amorph
Nice find. Unfortunately, the copy is a vivid illustration of the difference between a copyeditor and a spellchecker. That's the sloppiest job I've ever seen on an Apple page.
A lot of people have been waiting for this tech, though. More accessibility == Good.
Like what? I've only seen a few things so far.
Originally posted by curiousuburb
sounds cool
as long as they don't use the "Trinoids" voice.
Originally posted by kim kap sol
Hopefully they'll revamp the whole MacinTalk framework...it hasn't changed much if at all since 1993-4
Yeah. I hate to say it but Micro$ofts speech recognition is a lot more accurate. I was really impressed when someone showed it to me the other day. I am used to getting 90-95% on my Mac with a small vocabulary. The M$ stuff got 100% with the same vocabulary. Apple really needs to update their recognition engine. It is embarassing.
What kind of speech input were you using under Windows? Was it actually *from* MS, or a third party? Was it dictation or discrete speech?
Originally posted by Kickaha
What kind of speech input were you using under Windows? Was it actually *from* MS, or a third party? Was it dictation or discrete speech?
I believe XP has dictation built in. I only know because I was helping a friend track down a strange problem that his friend was having with XP where every so often "..." would appear when he was typing and then random short words like "and", "it", "she" would appear. I thought it was a crazy virus at first but after consulting Google it appeared that Voice Recognition had been turned on by accident.
Originally posted by Kickaha
You do realize that MacInTalk is speech *OUTPUT* right? It's got nothing to do with the speech recognition engine.
Sorry, I guess I don't know my Speech API terminology. Whatever you call the speech recognition component, it doesn't seem to have been updated since the mid 1990s and it shows.
What kind of speech input were you using under Windows? Was it actually *from* MS, or a third party? Was it dictation or discrete speech?
The guy who was showing it to me was doing some programming using the interface. He referred to it as "Microsoft SAPI" and when I asked him what that stood for he said it was Microsoft's speech API. We are really only interested in discrete speech. We were testing on just a four word vocabulary, "Red", "Green", "Blue", "Yellow". I speak pretty standard American English, so I wasn't exactly tricky. Microsoft got 100% on me, and over 95% for 15 out of 16 random people we tested (most were essentially perfect; the other one was about 90%), with no training.
Um... wow.
That's kind of, er... easy.
No, really, the smaller the dictionary of words to look for, the easier the problem is.
That sounds about like my experienced %s with the Mac speech recog. I've got almost no problems with it, my advisor has it work for him *most* of the time. (He's a mushmouth though.)
And, just like your experience with SAPI, no training needed there either.
I've noticed that people either have almost no problems with the Mac speech recog, or it drives them nuts. I think it's based solely on how closely your speech matches the 'expected norm'. The SAPI 'expected norm' may be calibrated closer to your voice.
Originally posted by rok
"Tea... Earl Grey... Hot."
HHGTTU?
http://www.ecommercetimes.com/perl/story/33133.html
Asked, whether Apple would introduce a new version of MacOS X at this year's WWDC, Apple VP Ron Okamoto said:
"In typical Apple fashion, we will not comment; however, as I said, it is our opportunity to talk about our technology direction, so you can bet we will be doing that."
Originally posted by Kickaha
I've noticed that people either have almost no problems with the Mac speech recog, or it drives them nuts.
Tell me about it!
I say "What time is it?" and it tries to switch off my computer!
Originally posted by durandal
Well, we'll get to know more about the spoken interface in June, I guess...
http://www.ecommercetimes.com/perl/story/33133.html
Asked, whether Apple would introduce a new version of MacOS X at this year's WWDC, Apple VP Ron Okamoto said:
"In typical Apple fashion, we will not comment; however, as I said, it is our opportunity to talk about our technology direction, so you can bet we will be doing that."
It does seem odd that Apple would start a campaign about features to be added to the NeXT version of Mac OS X now if it is 9 -12 months away from fruition? Any comments or ideas?
Johnny
Originally posted by Aquatic
Speech is good. Speech recognition isn't. Hopefully they'll focus there since Vicki is basically understandable.
Vicky is not good either, at least not by todays standards...hope Apple license this: http://naturalvoices.att.com/demos/
Now, THAT'S good speech!
Originally posted by Power Apple
Vicky is not good either, at least not by todays standards...hope Apple license this: http://naturalvoices.att.com/demos/
Now, THAT'S good speech!
Mmmm....crystal...
I know what I'm doing tonight
(Oh, what I typed!)