Apple Spoken Interface

[MacRumors]

The following Spoken Interface page (http://www.apple.com/accessibility/spokeninterface/) has appeared on Apple's website.

The page describes an upcoming feature for Mac OS X:

Introducing the spoken interface for Mac OS X. The Universal Access capabilities of Mac OS X will soon be enhanced with a spoken interface that provides a new way, through speech, audible cues, and keyboard navigation, to access the Macintosh.

The technology appears to be related to a previous job opening (http://www.macrumors.com/pages/2003/09/20030922122609.shtml) at Apple for programmers for a "Screen Reader" application.

According to this article (http://www.macrumors.com/pages/2003/11/20031112121726.shtml), Apple was pressured to release their own solution when the only developer of such software for the Mac discontinued their version. A lack of availability of such software may have threatened Mac OS X's adoption in the education sector.

[aethier]

Hey, that sounds quite cool, think it will be a 10.3.x or in 10.4?

aethier

[adamberti]

Hey, that sounds quite cool, think it will be a 10.3.x or in 10.4?

aethier

I would think 10.3.x, just because if it was 10.4, they would hold off on the features until it was announced.

But then I read this:
"Apple's spoken interface technology is currently in development and scheduled for delivery in the next major release of Mac OS X."

So I guess 10.4. I hope we get improved voice sythesis, Apple hasnt really done much on that fron these days - except for that new Victoria in 10.3.

-Adam

[TEG]

I like the current system in OSX for Voice Activation. I hope this new one is easier than the current system.

TEG

[johnnyjibbs]

I think it's interesting to see that the Apple page says (halfway down first page):

"In fact it will be included in the next major release of Mac OS X"

So this answers the three above posters' question. We are looking at a shot of 10.4.

[AirUncleP]

No child left behind.

It's nice to see a company develop products to help ALL of us.

I'll say it again.... WAY TO GO APPLE!!

[andyduncan]

yay. This is a huge deal. Jaws for windows (the industry standard) is over $1000. It's also ridiculously difficult to learn and generally craptacular. It will be interesting to see how well Apple does the integration with safari; Jaws' integration with internet exploder is it's crown jewel.

[andyduncan]

oh and I also forgot to mention. With jaws you have to buy it or run the timed demo, which means many developers never bother to test against it. Having the screen reader built-into the system by default, freely available to developers means the applications they create are much more likely to be accessible. This is also a huge deal.

[elmimmo]

I expect to see this turn in yet another English-only feature... Speaking of which, Apple could at least re-bundle the Spanish (Mexican, not Spaniard, but less is nothing) voices that we had in pre-X systems (not compatible on X) so that at least we can have the oh-so-new feature like text-to-speech (I cannot dream of course of getting anything similar to Speech Recognition, that is indeed demanding too much for such a relatively newly introduced English feature).

[Diatribe]

Why do I get the feeling that if this continues and more stuff like this finds its way into 10.4 it is going to be an upgrade worthwhile to get.
Especially if they get the improved database in and more useful things, since 10.3 was more of a performance upgrade. We will see I guess, but it does sound promising.

[sw1tcher]

So this mean no new hardware (PM G5, PB G5, iMac G5) updates until OS X 10.4 is out, huh? :p

[drewbert]

I hope we get improved voice sythesis, Apple hasnt really done much on that fron these days - except for that new Victoria in 10.3.
That's be nice, but look at the image at the bottom. Same danged voices. There's much higher quality speech synthesis out there these days.

[silvergunuk]

I sent a link to macrumors a few weeks ago about rumours of a spoken interface. The article said that the next revision of osx won't be out till next year. heres the link http://macnet2.com/more.php?id=459_0_1_0

[Stella]

Sounds good.. especially since Mac no longer has any screen readers - *and* its "free".. with the OS, no additional software.

Mac OSX - the OS for everyone

[lucasraggers]

Great! I already use the zoom function very often.

[MacVault]

Any idea if Apple will include voice recognition so that one can control the Mac by speaking to it?

edit: stupd me. looks like this functionality is already built in. ;) Guess I've just never had the need for it.

[Frisco]

Seeing this won't be available for at least 1 year from now and no current developers making this for the Mac now--will Apple continue to lose major sales in education market?

[arn]

I sent a link to macrumors a few weeks ago about rumours of a spoken interface. The article said that the next revision of osx won't be out till next year. heres the link

umm.... this technology is speech synthesis/screen readers.... NOT speech recognition.

arn

[varmit]

They pull the feet out from under Apple, and Apple still lands on its feet. I'm sure they will allow developers to add speech into their programs, making the Mac even more user Friendly. How much is the new OS going to cost. Probably only $129 - $149, something like that, when to expect, I would say Winter '04 or Summer '05. If they come out with it this year, I will be surprised and proud of the work Apple is doing. Abandon the Mac, and it will bite you in the ass. They probably relieased this news to get Education sales up, or keep people from leaving. Which special ed teachers might be getting a Mac soon if they don't have one yet.

[Johnstkr]

This also appears to be the first fruits of the zeldman / bowman / apple collaboration (at least the first piece I've seen)

edit: The linked page not the application

[silvergunuk]

It's nice to see apple adding things to their os that makes using the mac more friendlier to people with disabilities such as sight and sound or people with reading difficulties such as myself. Probably has something to do with Steve Jobs himself being dyslexic.

[Snowy_River]

Well, I hope that Apple will have some new voices included with the next release. The addition of 'Vicky' was a nice one. A definite improvement over 'Victoria' in smoothness of her words. Now if they can do even more... :)

[simX]

umm.... this technology is speech synthesis/screen readers.... NOT speech recognition.

arn

I did find a link to this video (http://www.bu.edu/jlengel/kn65kfs.mov) at that speech recognition article, though... seems very interesting, but would be a miracle if they actually got it to work like that.

[MasonMcD]

I hope this is the beginning of a trend for Apple, giving hints of the roadmap.

Has Apple ever announced something like this for a major point release this early? Maybe they're trying to shore up some government accounts, or keep them from jumping ship.

[nagromme]

Yes, Panther gave us Vicki, the NEW Victoria.

But what happened to Bruce and Agnes? They took a turn for the worse! :) No more MacinTalk Pro?

[requies]

you would think they could proofread their web pages before just throwing them up there. i saw two errors just skimming.

/pedant

[Photorun]

This isn't as big as some people think, OS 9 had speech to command built in, but I never knew one person who used it.

Don't confuse this, as Arn pointed out, with ViaVoice (http://www.scansoft.com/viavoice/osx/)or iListen (http://www.macspeech.com/products/iListen.html), this wont allow Xers to talk their emails out, only control the computer, it wasn't utilized before, it probably wont be now.

[kenkooler]

Apple could at least re-bundle the Spanish (Mexican, not Spaniard, but less is nothing) voices that we had in pre-X systems.

I agree. I still wonder why Apple hasn't done it yet, after re-creating all the speech synthesys in X it shouldn't be a lot of work to port these voices to MacOS X.

[mdriftmeyer]

It's nice to see apple adding things to their os that makes using the mac more friendlier to people with disabilities such as sight and sound or people with reading difficulties such as myself. Probably has something to do with Steve Jobs himself being dyslexic.

Odd. I never heard him complain or demand we needed to add this into NeXTSTEP/Openstep or OS X when I was at both NeXT and Apple. He seemed to do just fine for over 20 years without it.

To me it is smart business and a marketing tool to leverage customers, via the OS, to switch to Apple.

If it is done correctly and saves schools money it just makes sense.

For the end consumer, it is even better and I know many people who could benefit from it who otherwise don't even think about computers.

[luggnutt]

For me, it's FRED or nothing.

[gwuMACaddict]

bubbles... c'mon... BUBBLES!!!

[scottstover]

This is straight from Apple, "At a recent conference for Accessibility in Los Angeles, Apple announced a new spoken interface that will be integrated into the next major release of Mac OS X. Apple also announced it will be making a technology preview of this software that runs on Mac OS X “Panther” available for evaluation by selected customers."

AND . . .

"Because the spoken interface is tightly coupled with the operating system, it will only run on Mac OS X version 10.3.3. Similarly, it is compatible only with applications designed for Mac OS X."

[Snowy_River]

Hey, how about 'Bad News'?

"The light you see at the end of the tunnel is the headlamp of a fast approaching train"

He he he... :D

[lind0834]

I think the MacNET article is based on rumors of this.. somebody said speech. They went to the extreme.

[Jerry Spoon]

I did find a link to this video (http://www.bu.edu/jlengel/kn65kfs.mov) at that speech recognition article, though... seems very interesting, but would be a miracle if they actually got it to work like that.

This is Star Trek type stuff here....can't wait! :)

[frozenstar]

I did find a link to this video (http://www.bu.edu/jlengel/kn65kfs.mov) at that speech recognition article, though... seems very interesting, but would be a miracle if they actually got it to work like that.

Natural language speech recognition is still a long way off. Forget having the processing power necessary to do it... The computational linguists just don't fully understand the mechanics of natural language processing.

[GUSTO]

PLEASE! let 10.4 be a 64bit OS :D

[obeygiant]

Dude,

Everytime I say to my iMac...

"Open the door please, Hal."

And it doesn't do anything!

And Its just like the movie, man!

YEAH!

[MacFan782040]

I heard OS 11 is supposed to support PC software :cool:

[ginoledesma]

By far the best sounding voices I've heard are those from AT&T's Natural Voices (http://www.naturalvoices.att.com/). The speakers sound very decent, and it beats anything Apple has to offer as of the moment. Check out their interactive demo (http://www.naturalvoices.att.com/demos/) and compare the bundled voices with what it can churn out. The demo allows you to save the short clips -- I use them all the time with Mail actions. :D

[Analog Kid]

This isn't as big as some people think, OS 9 had speech to command built in, but I never knew one person who used it.

Don't confuse this, as Arn pointed out, with ViaVoice (http://www.scansoft.com/viavoice/osx/)or iListen (http://www.macspeech.com/products/iListen.html), this wont allow Xers to talk their emails out, only control the computer, it wasn't utilized before, it probably wont be now.

It certainly won't be unless they make those voices easier to listen to. There's maybe 3 that are intelligible, the rest are just "look at what we can do with LPC"...

I used to use it for some things-- dialogs for example. If I wasn't at my machine I could hear what went wrong. The only thing I use it for now is iChat-- it's good to know which buddy just came or went.

[sonyrules]

It good to see Apple helping out disabled people. And to a good degree also. I dont think your going to see Microsoft or any otheer company show as much ambition, and then give it away for free.

[adamfilip]

Apple needed to post this on there website. so that schools and users of such technology will realize that they dont have to drop there macs.. and get PC's ,they are trying to keep the users they have and gain new ones.. too.

its a pre-emptive strike.. sort of..

[Analog Kid]

Odd. I never heard him complain or demand we needed to add this into NeXTSTEP/Openstep or OS X when I was at both NeXT and Apple. He seemed to do just fine for over 20 years without it.

To me it is smart business and a marketing tool to leverage customers, via the OS, to switch to Apple.

If it is done correctly and saves schools money it just makes sense.

For the end consumer, it is even better and I know many people who could benefit from it who otherwise don't even think about computers.

Personally, I think it's a wise move technically as well. By making it an "accessibility feature" you can develop a technology and it will be useful to some before it will be useful to all.

I tend not to use the speech I/O functions because they're way slower tome than reading an typing. Someone who has more difficulty on one of those tasks will make immediate use of even the earliest speech capabilities.

Then someone who types a little slower than I do, or sits further from their machine than I do will make use of it.

Then I'll use it in certain applications (web browsing comes to mind).

Then we'll all just throw our keyboards and monitors in a box and interact with our iPods... =)

[Analog Kid]

Apple needed to post this on there website. so that schools and users of such technology will realize that they dont have to drop there macs.. and get PC's ,they are trying to keep the users they have and gain new ones.. too.

its a pre-emptive strike.. sort of..

Who owns apple.com?

[Analog Kid]

Natural language speech recognition is still a long way off. Forget having the processing power necessary to do it... The computational linguists just don't fully understand the mechanics of natural language processing.

Yeah, the video goes beyond that. It's not just understanding the language, but it's making common sense judgments...
Woman: "Is this the last minute panic before the lecture?"
Guy: "No that doesn't happen until... *pause*"
Computer: "4:15"

C'mon... It's hard for a human to pick up that kind of cue...

[andyduncan]

This isn't as big as some people think, OS 9 had speech to command built in, but I never knew one person who used it.

Don't confuse this, as Arn pointed out, with ViaVoice (http://www.scansoft.com/viavoice/osx/)or iListen (http://www.macspeech.com/products/iListen.html), this wont allow Xers to talk their emails out, only control the computer, it wasn't utilized before, it probably wont be now.

you obviously missed the point of the article (you did read the article didn't you?). You also missed the point of Arn's post.

This isn't voice control, it's a screen reader. This allows people with no/low vision to use a graphical user interface. It reads aloud all elements of the screen and provides a consistent set of commands for controlling the interface. This is way beyond the "Speak selected text" feature we all played with when PlainTalk came out. You're right in that you probably won't ever use it. You obviously don't have vision problems (yet).

Screen readers are an absolute neccessity for people with vision deficiencies; they simply cannot operate a gui without one. OS X has no screen readers for it. That means if you are blind and you have a mac, you have to use the screen readers available for OS 9. Not only has that caused a lot of blind users to switch to PCs, but it affects Apple's viability in government IT situations. The ammendments to the Americans with disabilities act (section 508) requires all IT systems purchased by the federal government to be accessible to users with disabilities.

Having this technology integrated into the operating system is incredibly nice; the main problem with other readers (like Jaws) is that they are clunky, ineligant, have huge learning curves, and are nearly unusable. Oh, and they are ****ing expensive fringe products. Making the screen reader just another part of the OS is a huge deal. <insert joke about your lack of vision here>

[dashiel]

Natural language speech recognition is still a long way off. Forget having the processing power necessary to do it... The computational linguists just don't fully understand the mechanics of natural language processing.

while it's not as fantastical the above movie, there's a video of a new MIT voice recognition program they're working on that works quite well. it's more star trek with the language use requirements, you have to preface a command with "computer:" and wait for a confirmation tone (which is of course the same one you hear on star trek, nerds), and the language has to be pretty direct, but it was much better than viavoice or the current speakable commands.

[techie@surewest]

The following Spoken Interface page (http://www.apple.com/accessibility/spokeninterface/) has appeared on Apple's website.

The page describes an upcoming feature for Mac OS X:



The technology appears to be related to a previous job opening (http://www.macrumors.com/pages/2003/09/20030922122609.shtml) at Apple for programmers for a "Screen Reader" application.

According to this article (http://www.macrumors.com/pages/2003/11/20031112121726.shtml), Apple was pressured to release their own solution when the only developer of such software for the Mac discontinued their version. A lack of availability of such software may have threatened Mac OS X's adoption in the education sector.

This is sadly needed as currently blind folks CAN NOT use a macintosh with out a screen reader similar to windows JAWS. In my work I am forced to use Windows because of this lack. I hope they are serious about getting it.

Ken

[crenz]

I expect to see this turn in yet another English-only feature...

Same here, especially as Apple doesn't really seem to care much for the education markets outside the US :(. If you know people with disabilities and a mother tongue other than English, please ask them to fill in the questionnaire to show their interest: Questionnaire (http://www.apple.com/accessibility/spokeninterface/questionnaire.html)

[groovebuster]

I expect to see this turn in yet another English-only feature... Speaking of which, Apple could at least re-bundle the Spanish (Mexican, not Spaniard, but less is nothing) voices that we had in pre-X systems (not compatible on X) so that at least we can have the oh-so-new feature like text-to-speech (I cannot dream of course of getting anything similar to Speech Recognition, that is indeed demanding too much for such a relatively newly introduced English feature).

Nice of you being selfish! ;) But other languages would be nice too... At least the "big ones" like Spanish, French, German and maybe russian.

[groovebuster]

It's nice to see a company develop products to help ALL of us!

Please let me rephrase:

It's nice to see a company develop products to help ALL of us US citizens!

Sorry, but I couldn't resist! It gets on my nerves that so many features or services are "US only"...

[Belly-laughs]

Any idea if Apple will include voice recognition so that one can control the Mac by speaking to it?

edit: stupd me. looks like this functionality is already built in. ;) Guess I've just never had the need for it.


No need for it? Say the phrase: "Make Me World Leader!"

[TBR]

By far the best sounding voices I've heard are those from AT&T's Natural Voices (http://www.naturalvoices.att.com/). The speakers sound very decent, and it beats anything Apple has to offer as of the moment. Check out their interactive demo (http://www.naturalvoices.att.com/demos/) and compare the bundled voices with what it can churn out. The demo allows you to save the short clips -- I use them all the time with Mail actions. :D

So do I, best idea I ever had!

[Nermal]

No need for it? Say the phrase: "Make Me World Leader!"

I don't have a microphone. What does it do?

[aswitcher]

Screen readers are an absolute neccessity for people with vision deficiencies; they simply cannot operate a gui without one. OS X has no screen readers for it. That means if you are blind and you have a mac, you have to use the screen readers available for OS 9. Not only has that caused a lot of blind users to switch to PCs, but it affects Apple's viability in government IT situations. The ammendments to the Americans with disabilities act (section 508) requires all IT systems purchased by the federal government to be accessible to users with disabilities.



I think they should try and emulate famous actors like Sean Connery etc, to give their readers more personality and appeal :)

[SiliconAddict]

This is the next logical step in OS interaction. MS has been working on hand gestures but that will only take you so far.

[scottstover]

I don't have a microphone. What does it do?

Uhh . . . you're kidding, right?

[hayesk]

Honestly, if Apple is going to put effort into this, I would much rather they got all of the functionality working (and working well) than to have nicer sounding the voices. Sure, they sound old, but you can easily understand what they are saying, and that's the most important part.

Don't get me wrong, new voices would be good, but I would rather they do that later. And for all we know, they are working on new voices; 10.4 is still months away.

[7on]

lol, 17 negatives...

I wonder which 17 of you hate the disabled?

[Snowy_River]

By far the best sounding voices I've heard are those from AT&T's Natural Voices (http://www.naturalvoices.att.com/). The speakers sound very decent, and it beats anything Apple has to offer as of the moment. Check out their interactive demo (http://www.naturalvoices.att.com/demos/) and compare the bundled voices with what it can churn out. The demo allows you to save the short clips -- I use them all the time with Mail actions. :D

Well, I'm actually not that impressed with them. I just gave them a trial run, and the voice quality didn't seem much different than Vicki. In at least one instance, I'd even say that they were a far cry worse than Vicki. (Neither of the US voices properly pronounced my name 'Geoff'. The UK voices did, though.)

If I had to pick one of the english voices from AT&T's Natural Voices to port over to Mac OS X's voices, I'd choose Audrey from the UK english.

[Doctor Q]

Speaking of being selfish...

I'm not visually impaired, but I'd like to know if the new software will give me additional ways to control the GUI from the keyboard without using the mouse.

In other words, does this add new options even when you aren't having text read to you?

[wildo]

I work at the Royal National College for the Blind, the leading VI college in the UK. I've been crying out for this for many a year, due to the college being PC Based. This will certainly ruffle some feathers. Just today, my Tech Support Manager said he would budget for a Mac, for evaluation work. The tide is turning, lets ride the wave (cue Beach Boys, cue S Jobs in surfin' turtle neck gear!) :) :)

[Analog Kid]

Honestly, if Apple is going to put effort into this, I would much rather they got all of the functionality working (and working well) than to have nicer sounding the voices. Sure, they sound old, but you can easily understand what they are saying, and that's the most important part.

Don't get me wrong, new voices would be good, but I would rather they do that later. And for all we know, they are working on new voices; 10.4 is still months away.

It's not that they sound old to me, I just find them hard to understand, and hard to speed up without mangling the transitions between words. I find the electronic sound fatiguing. I can't imagine listening to that all day.

Just like I find the Aqua GUI less fatiguing because it's attractive and smooth, I'd expect better voices to make a spoken interface easier to use.

Granted, I don't use this feature for what it's meant to be used for, so my priorities may be off...

[SiliconAddict]

This is the next logical step in OS interaction. MS has been working on hand gestures but that will only take you so far.

:o My bad. From the description I thought this was voice control.

[Snowy_River]

It's not that they sound old to me, I just find them hard to understand, and hard to speed up without mangling the transitions between words. I find the electronic sound fatiguing. I can't imagine listening to that all day.

Just like I find the Aqua GUI less fatiguing because it's attractive and smooth, I'd expect better voices to make a spoken interface easier to use.

Granted, I don't use this feature for what it's meant to be used for, so my priorities may be off...

I couldn't agree more, about speeding them up. I find Vicki's voice quite pleasant and easy to listen to, so long as her voice speed is set to 'Normal'. But, as soon as you speed her voice up at all, there's some terrible clipping, and even some words get dropped. It makes it very difficult to listen to...

[rdowns]

This is the next logical step in OS interaction. MS has been working on hand gestures but that will only take you so far.

I have a hand gesture for MS. Unfortunately, Arn does not have the proper icon to attach to this message.

[Snowy_River]

I have a hand gesture for MS. Unfortunately, Arn does not have the proper icon to attach to this message.

I bet I could draw the hand gesture that you're talking about!! :D

[Snowy_River]

I did find a link to this video (http://www.bu.edu/jlengel/kn65kfs.mov) at that speech recognition article, though... seems very interesting, but would be a miracle if they actually got it to work like that.

Yes, well, if you're paying attention you'll note that the movie is set in 2011. So, let's keep pushing the boundaries, and we may get there in the next seven years.


Yeah, the video goes beyond that. It's not just understanding the language, but it's making common sense judgments...
Woman: "Is this the last minute panic before the lecture?"
Guy: "No that doesn't happen until... *pause*"
Computer: "4:15"

C'mon... It's hard for a human to pick up that kind of cue...

Actually (to be a teeny, tiny bit nit-picky ;)), it's
Woman: "Ah ha, is this one of your typical last minute panics for lecture material?"
Man: "No, no no no no, that's not until, um..." (Pause)
Computer: "4:15"

But the real point is that the computer is trying to fill in information that it sees is its human owner is trying to remember. While the leap to announcing when the lecture is does take a judgement, I don't see that the video is showing the man trying to cue his computer, so much as the computer trying to do its job. In this case, however, the computer manages to help the man stick his foot in his mouth by revealing that the lecture in question is that same afternoon, and therefore it is a last minute panic for lecture material. So, assuming that the computer is able to recognize natural language, and is therefore able to follow the conversation, I don't see this as too terribly extraordinary.

[elmimmo]

Natural language speech recognition is still a long way off. Forget having the processing power necessary to do it... The computational linguists just don't fully understand the mechanics of natural language processing.Well, I do not know if you think that way either because, as an English user, you have had for ages many more alternatives than what we Spanish have, and thus, you have a different quality threshold than what we can expect to find, or because you are far more demanding than I am.

Either way, I have been most impressed by Speechisimo (http://www.speechissimo.com/)'s quality. Having into account that it is the FIRST Spaniard Spanish voice I have stumbled upon, I could very well mistake the samples with a somewhat pretty dumb, albeit real life, human. That does not apply when you demand more expressive or punctated texts in the interactive demo, but it is definitely a thing a did not know was even close to technically possible, and at such a low price.

[elmimmo]

I cannot dream of course of getting anything similar to Speech Recognition, that is indeed demanding too much for such a relatively newly introduced English feature.I was being sarcastic, just in case someone did not get it. Speech commands have been on the Mac OS for ages, and still nowadays you could pretty much wipe out the Speech prefpane or whatever other trace of its function in the OS, and rarely any non-English user would miss it, since whatever that falls into that field is currently rather useless in a non-English environment.

[Tailwind]

Of course screen reader functionality will win friends and government contracts, but the broader benefit is that it helps lay the foundation for expanded voice interaction on the Mac.

Mix Speakable Items, Screen Reader, iSync, Mail and iCal - now we're talking!

User: "Computer, open a new email to Mom."
Mac: "Opened. It's been two weeks since you visited. Do you want to suggest a visit this Sunday?"

[jsw]

Of course screen reader functionality will win friends and government contracts, but the broader benefit is that it helps lay the foundation for expanded voice interaction on the Mac.

Mix Speakable Items, Screen Reader, iSync, Mail and iCal - now we're talking!

User: "Computer, open a new email to Mom."
Mac: "Opened. It's been two weeks since you visited. Do you want to suggest a visit this Sunday?"

Great. Then even my Mac could make me feel guilty... :)