Apple Patent: Speech Recognition Improvements?

[Kwyjibo]

arn, i think this should b e your next poll, Would you use Speech Recogition software if Apple Produced it?

as for the whole rotation mic maybe they will evnetaully make a whoel group ofmicrophones and a headset like oldstyle metal braces. people will put it on and have micrphoens spinning around. I also remember a segment on The Screen Savers where they had ultra directional laser speakers that were able to be aimed. Perhaps a laser mic is the key to tet recognition.

 

[Niknar]

Apple Tablet!!

 

[tpjunkie]

That wasn't just any computer...it was a Mac Plus!

 

[Niknar]

Originally posted by Kwyjibo
as for the whole rotation mic

The mic doesn't move. The mic is attached to the screen. The screen is what rotates.

Apple Tablet!!

 

[guifa]

Originally posted by AmbitiousLemon
i am interested in the multilingual properties of mac osx and how speech recognition could tie in with all of this. will it work for all the osx languages installed? even the tonal languages?

I was wondering the same thing. But then I thought about it...English is technically tonal too (we use different tones to distinguish betwixt a question and a statement), so, perhaps they're looking at implementing those elements of the English language so we don't even have to type question marks! If they got that, then going to tonal-based languages shouldn't be too hard, especially since most, if not all, the other languages of the world have inteligent spelling systems.

At the same time, it might drive me crazy since a lot of times I'm chatting on-line in both Spanish and English, and thus, how would the computer recognize which language I was using? Would I have to use a keystroke to change languages?

 

[Kwyjibo]

Originally posted by Niknar
The mic doesn't move. The mic is attached to the screen. The screen is what rotates.

Apple Tablet!!

I"m just stating what I imagine when someone says mic that filters out backgorund noise and I'm confused on how a person would use display thats rotation. as for the apple tablet, two posts with that maybe your only seeing everything as tablet speculation.

 

[Trimix]

To make it understand Spanish you would use a keyword like HOMBRE, going to french something like uuhhuu pooopooo, then switching to German SCHULZ or ACHTUNG comes to mind 😛

No seriously, any verbal command could do - somehow with what I have read about i-chat and the whole teleconferencing thing, would it not be great to have the mac serve as a typist just taking down everything being said during a teleconference ?

A little while ago I read somewhere about ORANGE working on a little button for one's ear, which serves as a personal digital assistant, in the same time measuring temperature and checking one's overall well-being - it reads e-mails and faxes to the user, can redirect phone calls, wakes us up in the morning etc..

I guess what I want to say is that I expect from apple not to fall into the trap again of making a PDA like we are used to seeing them everywhere, but to introduce a totally new form-factor.
With .mac, they have the base for us users to store all our info we ever need, now we only need something to access it on the go -
I know I am dreaming - sorry

 

[Niknar]

Originally posted by Kwyjibo
I"m just stating what I imagine when someone says mic that filters out backgorund noise and I'm confused on how a person would use display thats rotation. as for the apple tablet, two posts with that maybe your only seeing everything as tablet speculation.

I did try to delete my first post but it wasn't letting me.

This tech would probably find its way into all Apple Monitors and computers.

But it sounds like it was made for a table computer that could be healed either portrait or landscape.

I could be wrong.

 

[dscottbuch]

Directional Microphone

I believe this is aimed at using some sort of phased microphone array. If you notice in the first patent is 'two or more' microphones. With this type of array and some fairly simple processing, the output signal of the array can be adaptively focused on the speaker, even as the speaker moves (as long as its not too fast). This can result in a tracking directional microphone which tracks the voice in front of the screen, especially in certain frequency ranges which are critical to voice recognition. This all comes from standard beam forming theory in radar. I'm actually surprised the patent office would grant this patent even though, if done cost-effectively, it will create a great system.

 

[pyrotoaster]

Let's think about OS X for sec.
Speech Recognition serves very little purpose in the system, besides Universal Access. Then there's Inkwell, which serves even less purpose. Apple is testing technology using OS X.

The real question is what are they building up to?

What about a mini-tablet with voice-input as well as native handwriting recognition? Inkwell seems to work well for those who have tried it (I haven't, but the recognition in my eMate works very well, and Inkwell is based on Newton tech), which means Apple could really put it to use.

Just imagine the overall innovation in a device that can read handwriting and dictate speech. If we take other rumors into account, it would be one amazing little gadget.

What's clear here is that Apple is getting ready to "Wow" us. I'm hoping for a Blue Dalmatian blender that switches modes based on voice commands.

"Mix!" 😛

 

[noverflow]

Pic of the device!!!!

http://aiw1.uspto.gov:80/.DImg?Docid=us20030033148ki&PageNum=1&IDKey=919B2AAC9127&ImgFormat=tif

this is a drawing from the pat page.

you must scroll down the the VERY BOTTOM

looks cool

this is a pic of the device

all i did is click on images

 

[Freg3000]

Re: Pic of the device!!!!

Originally posted by noverflow
http://aiw1.uspto.gov:80/.DImg?Docid=us20030033148ki&PageNum=1&IDKey=919B2AAC9127&ImgFormat=tif

this is a drawing from the pat page.

you must scroll down the the VERY BOTTOM

looks cool

this is a pic of the device

all i did is click on images

What the heck is that thing?!?!!? It looks so strange, like a.....ummmm.......something living, with those two circles representing the eyes and the triangle as the nose. I suppose this is just what it looks like internally?

 

[voicegy]

It's better than it used to be

I remember how "fun" speech recognition was when it first came out and how Steve wowed us all with voice recognition password access to his computer on stage. Rarely did it ever work in real life. However, I've noitice that "Speech" on Jaguar has impoved markedly; I'm quite impressed with the forgiveness of the system, and easily, 9 times out of 10, it recognizes what my commands are.

My favorite thing to do is to play hands-free chess in front of a co-worker. Apple just keeps making it easier and easier for me to blow away those Pee-Cee guys and gals with stuff like this.

I have no idea what's coming around the bend with this latest patent, but any improvement upon what's already in place would be fantastic and most welcome. Perhaps Apple's own version of ViaVoice or some such product?

Bring it ON!😀

 

[Trimix]

Oh I hope this is not what it looks on the inside - I stopped buying gadget with two little men inside ages ago 😛

I am sure some studio technicians will tell us what it is but for me it looks like two people talking at the device from two different angles.

 

[Teqanjel]

Apple-branded speech recognition

SO glad to hear Apple is actively doing something on speech recognition!

Coincidentally, I posted some hypothesizing yesterday on MacRumors' Page 2 regarding that "new plastic enclosure". At the end, I said:

"So put this all together and what do you get?
- A product which opens up new possibilities, not just doing the same-old-things better;
- A product which encourages purchase by both Mac and PC users; and,
- A product which could introduce new Apple-branded software.

"I am going to go out on a limb here and suggest that this new enclosure is meant to house a portable computer which not only serves as a PDA and tablet, but which incorporates Apple-branded speech recognition software. Via AirPort Extreme or wired network, it can also serve as a remote console to the user's main desktop or laptop, regardless of whether that computer is Mac OS or Windows driven."

If I'm right, can I start up my own rumors site? 😉

Regarding speech recognition (SR)...I use it daily on my job (I use technology to help people with disabilities, and among other things, train others on Dragon NaturallySpeaking, IBM ViaVoice, and MacSpeech iListen). There are a lot of things Apple and others could do to improve SR and make it a more effective and less cumbersome tool for daily use. The use of multiple microphones is certainly one -- microphone arrays have been shown to improve performance. Then there's things like collecting other input to help the software interpret what it hears. For example, IBM is working on using a camera to watch the speaker's lips -- such additional clues increase the chances of the software correctly determining what the user says.

There is also the issue of speech recognition from multiple speakers (like in a meeting), and speaker-independent SR (little or no pre-training required) -- areas which I know IBM is working on with hopes of releasing a commercial product within a year. I can't believe Apple would let itself fall behind in this important area, so...bring on an Apple-branded tablet/PDA with SR, HWR, and any other technology which makes the hardware/software adapt to us rather than us adapt to the needs of the technology!

 

[Elenita]

One Handed Typing

Originally posted by newnomad
As for pda's (and all small input devices for digital gizmos for that matter) I think its about time we learn a whole new 1 hand operated way of typing. Man stands above machine and is smart enough to adapt.
If it works for a trumpetplayer......

There is actually a way of typing one-handed using a standard QWERTY keyboard that I discovered recently. Most of the people who use this method tend to have limited ability in one hand, making "normal" typing impossible, but I see no reason why people wanting to use PDA's couldn't use it either.

Check it out!

 

[imaswitcheryeah]

PIC OF iMAC MICROPHONE

Sorry if this is a little off topic, but its pretty close😉

Im a service technician and I saw this pic in the iMac service manual. I dont own an iMac or know someone that does, so I was wondering if this actually worked.

(The wire on the right is the LED.)

If anyone with an iMac sees this, lemme know if it works. Thanks.

 

[deepkid]

Re: Re: Pic of the device!!!!

Originally posted by Freg3000
What the heck is that thing?!?!!? It looks so strange, like a.....ummmm.......something living, with those two circles representing the eyes and the triangle as the nose. I suppose this is just what it looks like internally?

Hahah.

 

[Quark]

Re: Pic of the device!!!!- WRONG!

Originally posted by noverflow
http://aiw1.uspto.gov:80/.DImg?Docid=us20030033148ki&PageNum=1&IDKey=919B2AAC9127&ImgFormat=tif

this is a drawing from the pat page.

you must scroll down the the VERY BOTTOM

looks cool

this is a pic of the device

all i did is click on images

Hey, goofy! 🙂

That's a pic of the person sitting in front of the microphones. It is a person. Not a device.

That's why it looks so strange for a device. On the right they are showing the head that is speaking and that the mics will pickup the speech. The shoulders and the legs are also shown, as well as the eyes and nose. These are very basic representations and they didn't need to be elaborate.

Quark

 

[jimjiminyjim]

Speech Recognition is what I am waiting for. I was going to shell out some cash to MacSpeech for iListen, but they wouldn't give me a test version, and the reports gave it too low an accuracy rating for me to justify the expense without my own trial run.

While my iMac 233 is becoming dated (though it still runs everything I have wanted to run) It can suffice until a 970 with integrated voice recognition comes out.

Does that seem like wishful thinking...? 😉

 

[grapeice]

A similar project

As soon as I read this, I thought of a research project I did in college not even a year ago. The project was to take an array of microphones and position them in some way so that when a person moved around, the microphones would change their "beam" to continuously point to that source. The source would have a small device that was tracked by a receiver on the other end. The computer would then change the weighting of what direction each microphone would listen to, thus creating a "beam" or area where the sounds could be picked up the most, and all other areas outside of that beam would be cancelled out, to get rid of the background noise.

Electrical engineers are trying to test similar devices in cars, where instead of using a cell phone, you have microphones positioned around the main driver in the car such that they can speak when they need to and all the background noise of the car driving down the road, the radio, etc, could be cancelled out. Everything could be completely hands-free, command-activated. I actually know someone right now working on this very application, although no word as to which automotive manufacturer is funding this research.

The real question is if this is something that Apple is trying to accomplish, what would they be doing with it? They mentioned 2 or more microphones... could this be an array that could be used in a similar fashion? I don't understand what would be the use of a rotating display, other than if you think of it like an enclosed workstation, where if you need to move your head to another location, the display moves with you so that it is in front of you. Of course, I don't see how that could be a consumer application.

Or perhaps the microphones in the display know whether you are talking to it and need to add something, or talking away from the computer, and not add the things you are saying because you are saying something else. Either way, I'm glad to see that Apple is continuing to innovate. I wish I could work there and see all the cool ideas they are putting together...

 

[Dunepilot]

Originally posted by dstorey

What would be reallt intresting is if the display had a cam as well, and could automatically move and rotate to follow the user or turn to face a second speaker when they start talking. very useful for meetings, but very confusing for the poor mac if an argument ensued. The all seeing, all listening, dancing iMac anyone, how much cool factor would that have......

Add iApps to allow you to watch tv on the iMac (removes the need for a seperate portable in the study/kids room/ wherever the mac is) and allow to record a la tivo. and a iHome app that controls standard devices for home automation systems (possibly over bluetooth/WiFi depending on range). This would allow you to use playlists for the home, such as have settings for the lights, heating etc when your at work/holiday/night time/romantic night in(controlling the dimmer switch 😉) .......

If the display could even turn itself you could have bluetooth movement sensors for the home where if someone moves into the room where the mac is, the sensor transmits to the mac and it turns and starts to record with the cam and saves t your .mac account.

I thought all of this was leading up to

"...and then when you've dimmed the light switches using your Romantic playlist, your Mac camera can capture the action happening on your living room carpet, whilst pumping out some Herbie Hancock MP3s to complete the experience..."

and save it to your .mac account, of course😉

 

[Winston Smith]

Sounds like its not just a pda but mainly for desktop systems. The whole thing talks about a rotating display and this doesn't suggest a pda and i cant work out why or how a pda would have a rotating axis, unless it just means the user manually turning the device, but this doesn't use multiple axis. It sounds more to me like an iMac where the screen can rotate on the neck for whatever reason. And by the sounds of that description, a studio display that has mic's and a display stand so it can rotate and have free movement like the iMac.

My earlier feeling exactly, this is a technology for the whole Apple 'experience' not just a particular new piece of kit. This is all about simplifying the GUI for whatever machine is used- personally I think this could be the next reinvention of the wheel.
Apple generally builds computers that work (could be a marketing line) this is about ensuring that speech recognition 'works' then it becomes a marketable factor in OS X 10.2.5 or whatever.

 

[Winston Smith]

Originally posted by Dunepilot
I thought all of this was leading up to

"...and then when you've dimmed the light switches using your Romantic playlist, your Mac camera can capture the action happening on your living room carpet, whilst pumping out some Herbie Hancock MP3s to complete the experience..."

and save it to your .mac account, of course😉

Much more fun this though......😀

 

[dstorey]

Originally posted by Dunepilot
I thought all of this was leading up to

"...and then when you've dimmed the light switches using your Romantic playlist, your Mac camera can capture the action happening on your living room carpet, whilst pumping out some Herbie Hancock MP3s to complete the experience..."

and save it to your .mac account, of course😉

well that could be an option when it becomes fully intergrated with iLife 😉 As well as .mac access, a 1-click system will be incorperated that sends your movie/music to playboy enterprises and in three working days you get a proffesionaly produced top shelf dvd delivered to your door...or you could produce it in iMovie for free...just for the budding prawn enthusiasts. In all seriousness that is a disturbing suggestion... herbie hancock indeed..