|
|
When will Text to Speech and Voice Recognition get an upgrade?
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Ok both technologies are pretty much the same from when they came out in 1992ish. You'd think if it worked on 50MHz it would be MUCH better 15 years and 4 GHz later.
Yet here we are and the voices are just as silly and recognition no better.
I remember years ago there were rumors that the guy who voices HAL from 2001 was going to record lines to make it the new text to speech. Never happened.
I also recently read this story:
http://news.bbc.co.uk/2/hi/entertainment/4665254.stm
How awesome is that? They can now recored every instance of a persons voice and have it reproduced with text to speech?!!
In every futuristic movie people interact with their computers with voice and have the computer reply. i think in some cases it would be very useful (don't take away my mouse).
Shouldn't Apple get a head start!?
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Banned
Join Date: Jun 2003
Status:
Offline
|
|
Originally Posted by Landos Mustache
Ok both technologies are pretty much the same from when they came out in 1992ish. You'd think if it worked on 50MHz it would be MUCH better 15 years and 4 GHz later.
Yet here we are and the voices are just as silly and recognition no better.
I remember years ago there were rumors that the guy who voices HAL from 2001 was going to record lines to make it the new text to speech. Never happened.
I also recently read this story:
http://news.bbc.co.uk/2/hi/entertainment/4665254.stm
How awesome is that? They can now recored every instance of a persons voice and have it reproduced with text to speech?!!
In every futuristic movie people interact with their computers with voice and have the computer reply. i think in some cases it would be very useful (don't take away my mouse).
Shouldn't Apple get a head start!?
I'd be using the Mac OS voice recognition feature right now if it wasn't so horribly shitty. It's really sad that the state of voice recognition and text-to-speech hasn't changed on the Mac in 14+ years.
With voice recognition + mouse + keyboard, we could all be near twice more productive.
Being able to change a song in iTunes or find a contact in Address Book or a file in the Finder without removing the focus, such as when you're typing a document, would be really cool.
Or vice versa...dictate an e-mail while you're looking for a song in iTunes or browsing the web in Safari.
It's clearly the future...and to those that think things will become chaotic in a work environment with people talking loudly to their computer...it's also time for businesses to rethink the idea of emprisonning people in cubicles for 8+ hours a day. With video conferencing at everyone's finger tips, there's no reason why everyone must centralized in a building every day of the week. People can work from home and talk to their computer without bothering anyone.
(
Last edited by Horsepoo!!!; Jul 24, 2006 at 01:52 PM.
)
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Ya my luck with Text to speech is out of 10 commands it gets about 2 right on the first try.
On the plus side the tacky text to speech voices are great in techno songs.
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Mac Elite
Join Date: May 2001
Location: NYC
Status:
Offline
|
|
Man, can't be soon enough.
I've been hoping this will be a big feature in Leopard for a while now but expectations for Leopard are already running so strong with so many different things that I'm preparing myself for having to wait until Leopard+1.
|
|
|
|
|
|
|
|
|
Posting Junkie
Join Date: Feb 2000
Location: Washington, DC
Status:
Offline
|
|
Ha... this takes me back. Me screaming at my Performa "My voice is my password... verify me." Password verification was interesting... but just too geeky, inconsistent and slow for most people.
IMHO, it was creepy...
I think Landos Mustache is correct in that I could see it being useful in some situations... Also, having the guy from 2001 doing the voice would be cool!
|
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Jun 2000
Location: If I tellz ya, then I gotsta killz ya !
Status:
Offline
|
|
tts & VR will get an ungrade ONLY when:
A) people demand it.....for most people right now, it's simply NOT a big priority
B) the processing power is available to make it work.
C) the applications are coded well enough to make it work
It takes an incredible amount of processing power to ACCURATELY identify, reproduce, AND respond to human speech correctly. Yes, it can be done, but it would require a mainstream shift in cpu design, bandwidth utilization, and efficiency to become useful.....
something current cpu mfgr's just dont seem to focused on right now
|
Personally I find it hilarious that you have the hots for my gramma. Especially seeins how she is 3x your age, and makes your Brittney-Spears-wannabe 30-something wife look like a rag doll who went thru WWIII with a burning stick of dynamite up her a** :)
|
|
|
|
|
|
|
|
Clinically Insane
Join Date: Oct 2001
Location: San Diego, CA, USA
Status:
Offline
|
|
They can't make it any better than it was back then even with the several hundredfold increase in CPU speed since text-to-speech came out?
|
Chuck
___
"Instead of either 'multi-talented' or 'multitalented' use 'bisexual'."
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Jan 2003
Location: NYC
Status:
Offline
|
|
Yeah, someone posted the video of the original Macintosh introduction on here a few months back, and it played some text-to-speech sentence introducing itself and it sounded exactly the same as TTS does now.
But then, there was that episode of Family Guy where Stewie made his Mac speak, and Radiohead's "Fitter Happier," so I guess the voices do have some place in society
|
"I start fires!"
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by bowwowman
tts & VR will get an ungrade ONLY when:
A) people demand it.....for most people right now, it's simply NOT a big priority
B) the processing power is available to make it work.
C) the applications are coded well enough to make it work
I dunno, Apple has always been great at giving us new technologies that we never knew we needed but now can't live without.
Like I said a 50MHz Mac understood and spoke as clear as a Quad g5 does today, I see no reason they can't make it 3x better with just a little bit of effort. Heck my PHONE has better voice recognition that my mac!!!! Come to think of it... it has never got anything wrong!
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Apr 2002
Location: Illinois
Status:
Offline
|
|
Wasn't there some legal troubles with someone over this awhile back? IBM actually I thought.
|
|
|
|
|
|
|
|
|
Posting Junkie
Join Date: May 2001
Location: Brisbane, Australia
Status:
Offline
|
|
|
|
|
|
|
|
|
|
|
Clinically Insane
Join Date: Oct 2001
Location: San Diego, CA, USA
Status:
Offline
|
|
Most of those voices are not particularly better than Vicki. The British Charles voice is downright hilarious.
|
Chuck
___
"Instead of either 'multi-talented' or 'multitalented' use 'bisexual'."
|
|
|
|
|
|
|
|
Senior User
Join Date: Apr 2002
Location: California
Status:
Offline
|
|
Originally Posted by Chuckit
Most of those voices are not particularly better than Vicki. The British Charles voice is downright hilarious.
More important than that is how the voices perform at reading. The Mac OS X text-to-speech fails in each of these three cases:
The Mac takes a bow each time...
"The girl with the bow in her hair was told to bow deeply when greeting her superiors."
The Mac takes up guitar both times...
"My hobbies are bass fishing, and playing the bass."
The Mac only knows about tasks...
"My latest project is to learn how to better project my voice."
This is a truly pathetic state of affairs. The AT&T system (interestingly correct depending on the voice you pick) and an IBM system that can be tried online get these random tricky examples correct. There are others that probably get it right as well.
Not only is the synthetic sound less than state-of-the-art but the ability to read naturally and differentiate context and proper pronunciation is poor. It's difficult to do, sure, but they can create a heuristic database to address many such circumstances.
IBM's research demo is here: http://www.research.ibm.com/tts/coredemo.shtml
|
20" iMac C2D/2.4GHz 3GB RAM 10.6.8 (10H549)
|
|
|
|
|
|
|
|
Clinically Insane
Join Date: Oct 2001
Location: San Diego, CA, USA
Status:
Offline
|
|
I see. I thought the idea was that it sounded less like somebody being choked. Those are pretty good. I tested it with "minute" and it got every case right. Pretty impressive.
|
Chuck
___
"Instead of either 'multi-talented' or 'multitalented' use 'bisexual'."
|
|
|
|
|
|
|
|
Addicted to MacNN
Join Date: Nov 2005
Location: President Skroob's Office
Status:
Offline
|
|
Originally Posted by - - e r i k - -
Those sound absolutely awful. Almost worse in ways.
|
"She's gone from suck to blow!"
|
|
|
|
|
|
|
|
Clinically Insane
Join Date: Oct 2001
Location: San Diego, CA, USA
Status:
Offline
|
|
The IBM one Dru linked to is a little better.
|
Chuck
___
"Instead of either 'multi-talented' or 'multitalented' use 'bisexual'."
|
|
|
|
|
|
|
|
Mac Enthusiast
Join Date: Jan 2006
Status:
Offline
|
|
lol, i'd be swearing every time i lose or get pwned in a game of Star Trek Voyager: Elite Force, and i'd probably have the computer responding with "wtf... what do you want me to do?!" and i'd be like "damn! program me a wallhack or an aimbot, would you!?!?" j/k
I also wouldn't mind giving my computer the personality and voice of Star Trek's 7 of 9 (Jeri Ryan)... lol, forget about HAL!
|
Don't bully me, I got an Uzi... HOO-HAH!
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by Chuckit
The IBM one Dru linked to is a little better.
Ya a bit. I am dying to hear that Dr. Who one.
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Clinically Insane
Join Date: Oct 2000
Location: Los Angeles
Status:
Offline
|
|
If you guys recall, at a Macworld years back Jobs made an off-hand comment to the effect that Apple is interested in acquiring better synthesized voices from a third party company. It would be nice to see updated technology to come to OS X, but I'm sure most of us would rather see a better Finder first.
|
"The natural progress of things is for liberty to yield and government to gain ground." TJ
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Well in that case I hope they got around to it.
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Posting Junkie
Join Date: Feb 2000
Location: Washington, DC
Status:
Offline
|
|
Originally Posted by Big Mac
If you guys recall, at a Macworld years back Jobs made an off-hand comment to the effect that Apple is interested in acquiring better synthesized voices from a third party company. It would be nice to see updated technology to come to OS X, but I'm sure most of us would rather see a better Finder first.
I agree... the Finder update is 100X more important, but I would love to see Apple purchase Nuance ( http://www.nuance.com/naturallyspeaking/home/ ) and implement it at the OS level.
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by Big Mac
but I'm sure most of us would rather see a better Finder first.
Why would it have to be one or another? I am sure 10.5 will have a bigger selling point other than just "We fixed that ****ed up Finder".
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Grizzled Veteran
Join Date: Dec 2001
Location: Land of Enchantment
Status:
Offline
|
|
I'm fiine with the Finder, but would really really like to be able to input text by speaking it. Much more efficient than typing. Dragon make some pretty good stuff for PCs, is it to much to ask to have that kind of quality voice recognition on the ten times better OS X?
|
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Jun 2000
Location: If I tellz ya, then I gotsta killz ya !
Status:
Offline
|
|
Originally Posted by mitchell_pgh
ummm........that website hasn't been updated in 1.5yrs, and neither has the application
It's essentially the same p.o.s. app they've been selling since before OS X 10.0 came out..........been there, done that, a BIG waste of my time & $$ then, and it still is
So, like I already said, until people start DEMANDING that personal computers and software are (re)designed, from the ground up, to support/maximize/exploit VR and TTS capabilities, this subject is basically a d.e.a.d. horse..........
and we all know how far one can travel on one of those
(
Last edited by bowwowman; Jul 28, 2006 at 08:55 AM.
)
|
Personally I find it hilarious that you have the hots for my gramma. Especially seeins how she is 3x your age, and makes your Brittney-Spears-wannabe 30-something wife look like a rag doll who went thru WWIII with a burning stick of dynamite up her a** :)
|
|
|
|
|
|
|
|
Baninated
Join Date: Sep 2002
Location: http://www.rotharmy.com
Status:
Offline
|
|
i couldn't agree more
i fired up speech recognition today , thinking that after 12 years it MUST have improved ( last used it in 1994 on an 840av )
sadly , no
it's utterly useless as ever and apple should be really ashamed of even offering such garbage
is it that hard to get this thing working ??
and as for the voices - ugh
|
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Mar 2000
Location: Allston, MA, USA
Status:
Offline
|
|
The only thing about Text & Speech that amazes me is that Apple bothered to bring it over from OS 9. The only purpose it serves in OS X is to drive the Universal Access features that OS X provides. For that reason alone, it should be upgraded.
Oh, and Horsepoo, while your idea about dictation while surfing the web making you twice as productive sounds great, try dictating an email into a tape recorder sometime while surfing the web or finding a song in iTunes. I'm sure you'll find you do each task less than half as fast and half as well. I work with people who dictate, and when they try to do something such as find a document on their desk while dictating they inevitably end up with a few minutes of dictation that . . . uh . . .(shuffle) . . . souuuuuuuundssssss . . . uh . . . some . . thing . . .like . . . . . . . . . . this, and if what they are looking for is more than 3 sheets below the surface, they end up just stopping dictation to find it.
|
-- Jason
|
|
|
|
|
|
|
|
Addicted to MacNN
Join Date: Nov 2005
Location: President Skroob's Office
Status:
Offline
|
|
|
"She's gone from suck to blow!"
|
|
|
|
|
|
|
|
Fresh-Faced Recruit
Join Date: Aug 2004
Status:
Offline
|
|
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by Quadroclops1
Ya, and I have never understood what the hell the big deal is.
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Fresh-Faced Recruit
Join Date: Aug 2004
Status:
Offline
|
|
it's all hell, sweetheart.
thanks to dark helmet for the google vid!
|
Sailor: Did I ever tell ya that this here jacket represents a symbol of my individuality, and my belief in personal freedom?
Lula: About fifty thousand times.
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Sep 2001
Location: Santa Monica, CA
Status:
Offline
|
|
Originally Posted by bowwowman
ummm........that website hasn't been updated in 1.5yrs, and neither has the application
It's essentially the same p.o.s. app they've been selling since before OS X 10.0 came out..........been there, done that, a BIG waste of my time & $$ then, and it still is
What are you talking about? Dragon Naturally Speaking has only recently been updated to v9, and v8 was not so long ago. DNS v8 is pretty impressive - I use it on my PC at the office and it works pretty well out of the box, better and better as you go along.
Were you referring to the IBM ViaVoice product that Nuance distributes? Then I agree -- crap-a-doodle-doo...
MacSpeech's iListen has promise, but it is lightyears behind DNS. If Apple arranged for a port of DNS to OSX, I'd be ecstatic.
|
|
|
|
|
|
|
|
|
Baninated
Join Date: Sep 2002
Location: http://www.rotharmy.com
Status:
Offline
|
|
it IS weird how it works the same now as when cpu's were probably 1% of the power they are now, so much for advancing technology , eh ?
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Forum Regular
Join Date: May 2001
Location: Bill Gates' Basement
Status:
Offline
|
|
a while back I emailed apple and told them to buy rhetorical systems as they had by far the coolest tts voices out there. on rhetorical.com you could put in any text and it would read it back for you as an indian woman, a valley girl, a british butler guy, australian woman, etc. It was so amazingly good I always used it on my voicemail at work and everyone thought I had a butler. Anyway, they got bought out by nuance and now it's called realspeak. they finally put the demo page back up online but it appears to be inactive for the moment--
http://www.nuance.com/realspeak/demo/
|
|
|
|
|
|
|
|
|
Addicted to MacNN
Join Date: Nov 2005
Location: President Skroob's Office
Status:
Offline
|
|
"new voiceover is incredible - sounds like a real person reading."
OMG I WAS RIGHT!
|
"She's gone from suck to blow!"
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Jan 2001
Location: Helsinki, Finland
Status:
Offline
|
|
Yea, it's better.
I was still hoping for licensed versions of David Bowie, Christopher Walken and Fran Drescher.
|
|
|
|
|
|
|
|
|
Posting Junkie
Join Date: May 2001
Location: Brisbane, Australia
Status:
Offline
|
|
Originally Posted by Judge_Fire
Yea, it's better.
I was still hoping for licensed versions of David Bowie, Christopher Walken and Fran Drescher.
...or that guy who voiced HAL
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by - - e r i k - -
...or that guy who voiced HAL
Ya him. It would also been cool to have Dr. Who (As they already recorded his voice for that purpose) or for the old school geeks a Dalek! Unfortunately the Dalek dude just died.
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Mac Enthusiast
Join Date: Mar 2000
Status:
Offline
|
|
Originally Posted by lookmark
Man, can't be soon enough.
I've been hoping this will be a big feature in Leopard for a while now but expectations for Leopard are already running so strong with so many different things that I'm preparing myself for having to wait until Leopard+1.
Considering they just updated the text-to-speech part, there is a chance that one of the "top secret" features (if you saw the keynote) is better voice recognition.
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by cla
Considering they just updated the text-to-speech part, there is a chance that one of the "top secret" features (if you saw the keynote) is better voice recognition.
Vista also has a pretty good one (minus that one demo) so you can pretty much bet on it.
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Jan 2001
Location: Helsinki, Finland
Status:
Offline
|
|
Originally Posted by Landos Mustache
Vista also has a pretty good one (minus that one demo) so you can pretty much bet on it.
The secrecy could be about the things you can do with it?
Vista demos seem to have been about dictating texts, mostly.
|
|
|
|
|
|
|
|
|
Professional Poster
Join Date: Dec 2002
Location: Partying down with the Ewoks, after I nuked the Death Star!
Status:
Offline
|
|
Originally Posted by Judge_Fire
The secrecy could be about the things you can do with it?
Vista demos seem to have been about dictating texts, mostly.
No. Another demo I saw you can control everything on the computer with it. You can even draw in paint programs.
(
Last edited by Landos Mustache; Aug 10, 2006 at 10:54 AM.
)
|
"Hello, what have we here?
|
|
|
|
|
|
|
|
Senior User
Join Date: Mar 2004
Status:
Offline
|
|
You'll need that speech recognition feature in vista quite a bit, especially considering how often my mouse frozen up in the last beta. "Move the mouse... noooo left... your other left!"
Apple has significantly upgraded speech in leopard, and they tested it against Vista just to show you how much better it was. It sounds like a real person
It's good enough that I can actually listen to stuff dictated to me.. Can't wait till they put those voices on an ipod.. and read me off my tracks.
|
|
|
|
|
|
|
|
|
Mac Elite
Join Date: Jan 2001
Location: Helsinki, Finland
Status:
Offline
|
|
Originally Posted by Landos Mustache
No. Another demo I saw you can control everything on the computer with it. You can even draw in pain programs.
Microsoft Pain rocks.
Did you come across any advanced macros, like "mail this to bill" or "find pictures of steve"?
|
|
|
|
|
|
|
|
|
Posting Junkie
Join Date: Feb 2000
Location: Washington, DC
Status:
Offline
|
|
Originally Posted by inkhead
You'll need that speech recognition feature in vista quite a bit, especially considering how often my mouse frozen up in the last beta. "Move the mouse... noooo left... your other left!"
Apple has significantly upgraded speech in leopard, and they tested it against Vista just to show you how much better it was. It sounds like a real person
It's good enough that I can actually listen to stuff dictated to me.. Can't wait till they put those voices on an ipod.. and read me off my tracks.
Me: "New document... Dear Joe..."
Computer: *BLING!* "It looks like you are..."
Me: "close paperclip... close paperclip... close paperclip... close paperclip... "
Computer: *BLING!* *BLING!* *BLING!* *BSOD*
Me: Nice...
Computer: "Did you say erase? Yes or Confirm?"
|
|
|
|
|
|
|
|
|
Clinically Insane
Join Date: Oct 2001
Location: San Diego, CA, USA
Status:
Offline
|
|
Originally Posted by inkhead
Apple has significantly upgraded speech in leopard, and they tested it against Vista just to show you how much better it was. It sounds like a real person
You really think so? I thought it sounded creepy. The words sound disjointed. Vicki sounds a little unreal, but at least her intonation doesn't jump all around like they randomly recorded word fragments from various moments in a person's life.
|
Chuck
___
"Instead of either 'multi-talented' or 'multitalented' use 'bisexual'."
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Forum Rules
|
|
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is Off
|
|
|
|
|
|
|
|
|
|
|
|