Microsoft working on passive-voice activation for Cortana

While asking Cortana various questions on Windows Phone 8.1 is natural, the way in which one enables the personal voice assistant isn't quite as fluid. Users need to press down on the Search key or press the Live Tile and microphone before Cortana enables, and while it's not difficult there are better solutions available.

One of the solutions is passive-voice activation, which is something users of the Xbox One with Kinect should know. Passive-voice activation allows users to engage the AI or system simply by using their voice. Likewise, the Motorola Moto X with Android can also activated by explicitly stating 'Ok, Google Now.' The phone will then come awake, letting users make appointments, starting their music, or sending a text message without ever touching the phone.

Although Cortana is still in its beta stage, many are curious as to when passive-voice activation will arrive, as it seems like the next reasonable step. Marcus Ash, Group Program Manager for Cortana on Windows Phone, has once again taken to Twitter to tease the feature:

"Something to think about as we design voice activation for #Cortana. Don't want Jesse sending texts on your behalf."

Ash is referring to the new Xbox One commercial featuring 'Breaking Bad' Star Aaron Paul. In the ad, Paul says 'Xbox On' and for anyone with an Xbox One and Kinect system, their system may accidentally turn on as a result. Clearly, this is a downside to passive-voice activation and something that Ash and his team are taking into consideration when further designing Cortana.

Of course, the big surprise there is that Ash just admitted that Microsoft is actively developing voice activation. That news is not necessarily surprising, but it is reassuring.

The one caveat to this feature is when it arrives, it will most likely only work on newer Snapdragon 800 and higher chipsets. The ability to leave microphones on for passive activation is not a trivial thing to enable on any phone (unless you're the NSA, we suppose). That means in designing and enabling this feature, Microsoft may have to somewhat fragment the Cortana experience, frustrating some users who have older hardware. But like all things in technology, the spoils go to those with the latest gadgets.

Ash is no stranger to Twitter lately, often revealing interesting bits of information about Cortana, including alternative names and encouraging Windows Phone users that their OS was top priority for development. Ash also recently discussed the origins of the voice assistant, including mass adoption by international users in a video featured on Windows Phone Central.

What do you think of passive voice activation? Would it make using Cortana easier for you? Would it entice you to buy a new phone with supported hardware for the feature?

Source: @MarcusAsh; via Neowin


Reader comments

Microsoft working on passive-voice activation for Cortana



Why not Apple's way of making Siri listen when the phone is plugged into a power outlet? Chances are if the phone is far enough that I want to command it with only my voice, it's over there being charged. So I can send a text while it is out of my reach, which is most likely during charging.

Not necessarily right. At home, my phone always sitting on the desk while I doing other things. Watching TV, browsing on my computer, etc. I only charge my phone when necessary. If it only listen when it plugged, then people like me won't have that much the chance to use cortana from distance. Not that I frequently use cortana though :p

Umm... Some technology called Wireless charging alerady helps this case. Why waste battery juice even when it sits idle(albeit being charged)?

What if each phone knew your own voice so that another person couldn't use their voice to mess with your own phone?

The phone already knows the owners name. Just say," this is [ insert name here] start XBox Music"

why not? i am free to be as rude as i want. of course... you are also free to tell me to shut up. hmm... guess we are both rude.

Now that would be incredible, but if done around another wouldn't it be easy for them to simply say what you've said to wake it as well? Though it would set it apart from the others in being able to set your own voice wake code.

No reason why that couldn't be done. In fact, it's been available for years, and no reason why Microsoft couldn't license an embedded version of VoiceIt for this purpose. http://www.voiceit-tech.com/ Disclosure: I know the guy behind the tech, but otherwise have no interest in the company.

What would be nice would be to bring the following to earphones more QUICKLY:
A long press on the earphone button should wake Cortana up and start listening - "hi Daniel".
I find that it is much more private to have a conversation via earphones, especially while walking I don't feel so silly talking to my phone, but also not having to take my phone out of my pocket, switch it on and start finding buttons. Also in this mode, Cortana should always read to you even if just to say "maybe try this", indicating a screen only response.
The passive listening sounds great, but I would rather get the above feature sooner rather than waiting an age for MS to perfect something which sounds quite complicated.

This feature is already there, I use it almost all the time I am out, like when I am on the road or in the bus, I long press the earphone buttons, and it brings up cortana, no need to take out my phone. As simple as that.

"Ok, Google Now!" --- seriously? that's what you have to say? could that be any stupider?

i'd much rather just push the button for a second.

Actually I think it's just 'Ok, Google'  and it's really kinda nice.  For example "Ok, Google when do the reds play next?"

The moto x is passive from the sleep state so they say "ok Google now" (only unique to the moto x) and when your on the home screen of pretty much any newer android phone all you have to say is "ok Google" and it will activate.. But the screen has to be on for anything but the moto x at the moment

No it is not "Okay Google Now."  Not on my phone. I have the LG G2 and when I say "Ok Google" , Google wakes up. 

My wife activates her SGSIII with "Ok Google"
And my Xbox never activates with the above mentioned commercial using the "Xbox On" Command.

Yeah, I've always thought the Voice Activation for Google is dumb having to say OK before you say Google Now...but, it may be the way they get around the issue raised above in the article: how do you handle accidental activation?  Just saying Cortana is a bit generic because what if you are talking to a friend about Halo, or explaining Cortana on WP to a friend and say "Cortana"?  Well, now you have your phone activated and it's recording everything you said after you accidentally launched it.  Then it's going to try do something with what it just heard.

OK, Google Now! is not something you would say in everyday conversation, so it's the perfect way to know that the user actually wants to engage the AI.

Actually, while this tech is not present in all phones just yet, it is getting more common to have phones (and especially with Lumias) that can detect where it is.  So if you are on the phone or the phone is in your pocket, then Cortana could be set to not activate when she hears her name, and only when she knows you are holding the phone and nothing else is going on.

But if I'm holding the phone I can probably push the button. I want it to work when it's on the counter away from my grip.

Thats a good idea, but I'd like it better to activate it with a voice command no matter where the mobile is. lets say you are laying on the couch your mobile is playing music over the stereo. You could then activate Cortana by voice and switch the track, mute it or whatever. Thats just one scenario, but there are plenty of it where I don't want to hold my phone in the hands when Cortana can read everything to me.

I think it would be cool if the user chose how to activate Cortana. You could type it in. Same as you type in your name. The system is really good at text to speech and I think this would be building on the whole customization of the OS for individuals.

kayb27 +1

Maybe we don't need to type it in...just record the phrase you want Cortana to recognize and she just listens for that phrase.

I'm okay with passive if it is secure, in the sense that somehow the system doesn't allow other apps to listen passively, or at least let you control it.

Seems like a no-brainer to allow users to enter a preferred word how to activate Cortana. Like "Hey" for "Hey Cortana" or "Go". I admit it seems rather simplistic, but sometimes the dev teams forget how to think simple.

However, leaving the decision to the Dev team would mean thaat everyone will have to use a phrase that appeals to the hypothesized majority. Allowing the users set the phrase not only makes the phone more personal, it also just makes sense for a larger group of users.

It would be great if we could manually set what could wake up Cortana in settings. I would set it "Hey Cortana"
Edit - lol, seems I am not the only one.

On my Nexus 5 I just say "Ok Google" and the phone starts listening. That is faster than holding the search button on my Lumia devices.

The downside to the voice activation on the Nexus 5 is that it only works from the home screen whereas you can hold the search button from anywhere in the UI and get Cortana.

I would like the option to have both the current way and the voice activated way.

Voice activation has it's downsides in rare situations for example, I taught an Android class a few months ago and my Nexus 5 was projected on the wall and during the history of Android section, Google Now kept activating the listening mode because it thought that I said "Ok Google" when in actuality I was saying "Google" with other phrases. I had to tell the phone to stop listening several times.

The students thought it was funny, I found it quite annoying. It did lead to another segment where I introduced Google Now and the need for special hardware to have an always listening mode from the Home Screen.

Any suggestions as to what we would say to activate Cortana? We can't just say "Cortana" as that would activate it any time the word Cortana is used, even if we don't need anything.

Cortana would reply: "Demitrius, I heard you talking about me earlier, you know that I can hear you right?"



To be honest, I wouldn't be saying "Cortana" unless adressing my phone; plain "Cortana" for activation would work just fine for most non-techies, I would think... "Google", or any phrase including "Google", on the other hand! Google is almost synonymous with doing a search, and is used all the time.

How goes is the voice recognition with Google Now? Cos with the issues MS have with the Xbox picking up Jessie, surely there's a risk other people are going to go "Ok Google" and make your phone do something without you wanting it to. Real chance for mischief if you're not careful.

My friend has a Droid Maxx and I know that with the google now feature, The phone only responds to his voice. I tried to say it after he demonstrated in front of me but it would not wake up or listen. I like that it knows only your voice and I would like MS to bring this type of feature to the Lumia group.

i think they should do it like this

- use the proximity sensor to detect if the phone needs to be listening or not (eg in pocket - off, on hand - on) 

- then if its connected to bluetooth then keep listening, and allow the user to determine a list of bluetooth device to be listening at all times (except when you are on the phone of course)

- then the most important fix for when the phone is on the table and get activated is give us an option to say, stop listening for the next (duration) or just stop listening to turn it off until we turn it back on by going to the search screen


that should take care of 90% of the issues, would be nice if it can recognize our voice/tone though it would reduce error rate even further

Well, a phrase like, "Cortona, listen" or "Cortona, are you there?" would be distinct enough to keep random activation from occurring. The phone would then respond, "How can I help you?" Or "I'm listening, ." This would work nicely.

On my Galaxy Nexus and Nexus 7 I can activate a search from any app by swiping up from the bottom and than saying "OK Google".

Posted via the WPC App for Android!

Google is so stupid that they have to put their name on everything - and making users have to say "Ok, Google" with personal assistant is even more stupid!

Microsoft names: OneDrive, Xbox, Windows, Windows Phone, Office, Bing, Cortana, Outlook, etc., and rarely you see the name "Microsoft" along.

Apple names: iOS, OS X, iWork, iTunes, App Store, iCloud, Safari, etc., and rarely you see the name "Apple" along.

GOOGLE names: Google +, Google Hangouts, Google, Google Drive, Gmail, Android, Google Now, Google Maps, Google Chrome, Google Books, Google Earth, Google Docs, Google translator, etc. I don't know how they didn't name YouTube "Google Tube".

Gmail and Android is probably their only saving grace, as for YouTube it just came that way when they bought it, would probably enrage alot of people if they decided to go with Google Video... (Google Tube sounds... I don't know, sketchy for some reason)

This is a ridiculous comment. Microsoft is famous for having some of the worst products names ever. http://bit.ly/1lBsQlH

Also, Microsoft is most certianly in the offical names of many of their products: Microsoft Office, Microsoft Word, Microsoft Excel, Microsoft PowerPoint, Microsoft Outlook, Microsoft Windows, etc, those are official names.

Having your brand isn't much different either: Xbox Music, Xbox Video, Xbox Live, Bing News, Bing Translator, Bing Sports, Bing Travel, Bing Food & Drink.

Microsoft also makes you say "Bing" when you want to search something with voice with the Kinect on Xbox, instead of just "search for..." It's stupid, why can I just say, "Xbox, find" or Xbox, search"

I've used Google Now with voice activation and it's actally pretty cool and works very well.

ya you're quite right.  Microsoft Azure, Microsoft Surface, Microsoft Dynamics, Microsoft Internet Explorer, and Microsoft Xbox are certainly valid tradenames.

the same can be said for apple:  apple mcintosh, apple tv, apple ipod, apple itunes, apple icloud, apple ipad, etc.

Excel is not "Microsoft Excel" it is just "Excel."  I just launched it on my Office 2013 version and "Microsoft" is nowhere to be found on the splash screen.  When launch System on Windows 8.1, the software is clearly just called "Windows" not "Microsoft Windows."

Indeed, with the exception of Dynamics, I cannot find a single example of MS generally including their corporate name in a product.  Sometimes they include their corporate name alongside the product name, but not usually.   In contrast, Google Drive's official name seems to be "Google Drive."  It is never referred to as just "Drive."  There really is a difference.

there is also google keep, google cloud print, google calendar, google finance, google music, and google play. 

we all live in a googly world.

well Xbox being the name of the product and is one word, and its a good thing it only listens for Xbox on to turn it on else I'd be creeped out too

Next time your in an area where many people are bound to have this feature, tell out "ok Google now" can be quite fun watching everyone's phone turn on

Doesn't work, at least not in a Moto X. It's trained to recognize and react to your voice specifically. For example, I am unable to trick my girlfriend's Moto X.

I can contest this is true on the Moto X. You aren't required to but you can train it to your voice and it works very well.

I am sad Google bought Motorola and sold it to lenovo.

I REALLY wanted a Motorola WP device, more so than I wanted a Nokia.   But then again I'm American.

No, u have to say "OK Google". That is how I activate Google Search on my Galaxy Nexus.

It would be nice to say "OK Cortana" on a Windows phone.

Posted via the WPC App for Android!

All you have to say is "Ok Google" and you can set wake up commands. If WP had a decent dev team then you guys would probably be in the 21st century.

I m gonna love it.. Bring it on.. I need to send messages via passive voice, and also need to support passive voice for 3rd party app like WhatsApp, wechat and many other, then it will be awesome..

I know Cortana use Bing, but this type voice control if integrated with WhatsApp or other app then it will be awesome , just say WhatsApp, send xxxxx to xxxx

I think that will come eventually, I believe the developers have to put the code into their apps to enable these features. I was thinking that and it would be convenient to have the radio apps (Slacker, MixRadio, etc.) be integrated so I can say what channel I want to listen to with which ever app I want opened, like "Open New Rock on Slacker Radio" and have the app open and jump right in.

I'm in the U.S,but I'm not going to run the craptastic Preview on my phone.Will wait for the official launch.

Missing out on having a phone that's unusable.No wifi,mms or text.Sorry!but I don't need a five hundred dollar paper weight.

I don't have problems buying apps with the region change. I'm guessing that it shouldn't be a problem for you either.

Having the microphone on all the day sounds like a big battery drain to me, but I guess there are ways around that.

The new chipsets mentioned in the post support a low power "listening" mode that doesn't put a massive drain on the battery.

Microphones in cellphones are passive anyway, so the battery is just going to the clock cycles the CPU needs to listen.

Sad that dedicated hardware might be required, means my beloved 1020 will probably never have this awesomeness :( but like Daniel says, to those with the newest gadgets goes the spoils...

look on the bright side, by the time it arrives it's probably time for us to upgrade anyways (my 1020 is only 6 month old lol)

For older device, at least have the option to enable it when charging. Then if you are driving and you have your phone hooked up and charging or in the morning and the phone is charging, we could activate it... Just a thought.

The Snapdragon 800 has a QDSP6V5 low-power DSP core that is required for always-on listening. You're not going to see this feature enabled on lower tiered chipsets.

Dafuq!! How the Moto x can do it then? If it's powered by 1.7 ghz instead of 2.2 ghz? And with iOS 8, users can activate Siri with voice on any iPhone from 4s... This is cheating!!

The Motox X incorporated the low power cores into their chipset. It's a modified Snapdragon S4 Pro specifically designed for voice activation.

are you sure they can activate Siri by voice on 4s? I remember a while back of hearing MotoX was getting some specific hardware to enable voice activation, so I guess this is what that report was talking about

Right now, Siri cannot be activated by voice on any iOS device. iOS 8 is bringing this feature, but your device will need to be plugged in to make it work. I guess it is a significant drain on battery.

I'm sure if they have the time they will find a way. I think it is one of those things that are harder to do without too many compromises so they will leave at last just in case they have the time and resources to do it. So there might be a chance.

They should enable it for BT only, IMO. That's where it's most useful, and you wouldn't have to worry as much about the mic having always on capability. I would rarely use it on my phone when I'm holding it anyway, just like I don't really use the voice commands much when I'm holding it, but when I have BT connected in the car or at home it's really convenient.

I find disapointing that the current Cortana unlocks my phone in my poket when I use it via BT. The previous voice commands on WP8 did not, and were actually useful to use. If I have to take the phone in my hands, at least half the usefulness is gone for voice commands...

How do you mean? The only difference I've noticed is that it no longer says "I turned on driving mode for you" when driving mode turns on (I kinda miss that).

As for starting apps etc it seems to work about the same. Except that Cortana is much more accurate.

Though to be honest what I mostly use it for is text dictation and starting music, so I am not well versed in the differences in voice commands between TellMe and Cortana.

It fully works on my BT already. Press and hold the answer button until you heat a beep, Cortana opens and takes commands ❤

Right, I know that. This article is about how to get it to activate WITHOUT pressing a button.

That is a great idea! I would just say "Hey Cortana" and have her reply "Yes Demitrius?"  (then I'd ask a question) even when the screen is locked (via a settings toggle to enable Voice Activation to work under the lock screen).

That way, I could just look at my phone and ask a question.

My biggest gripe is if I'm using my wired headset and I hold the talk button Cortana comes up but my phone is in my pocket and apparently then gets cancelled by my ly leg hutting the back button or something. Why doesn't the proximity sensor work when its in my pocket? I can see the screen lit up in this scenario its frustrating.

Yeah, that's my problem too. For me it's verry Microsoft-like to remove a useful feature, as the previous WP8 voice commands were useful with BT headset. Cortana needs to show off on the screen, defeating half the purpose.

It's already very useful for the lazies (like me) on the Xbox One. I do it everyday when I just want to watch Netflix. This is my most requested feat for Cortana in the future.

Yep that's what I'd really like. Especially for driving I thought car mode when it's charging would be a good setup... Mind you I'd like a UK version of Cortana first...

Battery drain ftw.

Constantly listening for the activation phrase... I would rather have an option in driving mode, activated when "on charge" in car cradle for example.

I encourage passive voice activation, but I can wait for it. Never really thought of pushing the button as a hassle.

How come Microsoft didn't think of the from the beginning? Anyway, very reassuring, that's for sure.

It's not a matter of thinking of it, of course they did, it's a matter or prioritizing features in design. You can't do everything in a beta. Not even Google had this at first and Siri still doesn't.

Cortana can go on all Windows Phone 8 devices. Cortana with passive voice? Maybe 5% right now. To me, that doesn't sound like a 'get this out on first release' type of feature.

I have an Icon now...I can't imagine getting another phone for quite some time unless it hovers overhead, records everything and has defensive weapons onboard.

Yes, it would definitely be a reason to get a new phone. I'm hoping the Lumia 930 supports that feature because that's the one I want.

I think it would be awesome, if we could set the passive activation word by ourselves. It would be amazing if I could activate it by saying "Heads up, Jarvis"

First they need to fix the dam Cortana, (whatever you want) resulting in friggin Wi-Fi On\Off. Gotten so bad I've turned Cortana off.

I think, passive voice activation will need extra sensors. So that means current devices are not able to get it. :(
We have to buy a new one.

For older devices atleast do like this,
Once we open cortana app(via tile or search button) microphone should stay on till we close the app or till display times out, now cortana stops listening once it completes the query,for next query we need to press microphone button again, its bit cumbersome.

I want passive voice activation. I've caught myself wishing I could just say "Cortana, send a text to..." or "Cortana, open..." already. Mostly, I just think of where the future goes with this -- I want to be able to say, on my way home from school or work or wherever, "Cortana, set the thermostat to 74 degrees in the living room," or "Cortana, open the garage door," or whatever else comes up. How close we are to the "smart homes" that were touted in the past is just great!

I would love to be able to say "Cortana my dear," to wake her up. Although my wife might be a little jealous. Lol!

Does anyone know where I can submit feedback to WP? I've been using WP8.1 on my Lumia 1020 for a couple months and after the most recent update my phone's email freezes when browsing, and internet explorer scrolling freezes consistently. Also the other day my Phone was receiving texts but my replies (though it showed as sent) were not sent until after I restarted the phone.
These problems were not so until this most recent update.


If they can somehow do it with minimal battery drainage, that'd be a major boost for WP all around the world. The easier (lazier) you make it, more people love it. :D

Don't worry. After saying "Hey Cortana" the phone will put your voice through vocal recognition software, scan your iris and fingerprint and ask you to input a 16-digit passwordall in the space of 17 seconds. Then it will redirect one of Microsoft's custom-built weapons satellites to geostationary orbit above your position that will vaporize any unauthorized users.

I want...

"Cortana, shut up!" to turn off my alarm in the morning.
"Cortana, pickup" incoming phone call.
"Cortana, hangup phone"

In Windows Speech Recognition there was a feature calles "Macros" that I really enjoyed which allowed you to create custom commands for WSR. It would be great if they could bring this to Cortana, but it would probably take up to much memory or someting

They could have it respond to "Bing", because no one says that in normal conversation.

/I'll show myself out.