Alexa learns to whisper and speak like humans do

Voice synthesis has come a long way since those days when computers sounded like robots, but there's still a long way to go and Amazon takes another step forward by teaching Alexa to speak with added emotions and human-like characteristics.

Google's Gboard keyboard for iOS gets voice transcription

Google's keyboard for iOS was missing a nifty feature that you have on Android for a long time, but they finally found a way to bring it to its competitor's platform: voice recognition.

Speech recognition technology is now on par with humans

The time when speech recognition systems were considered a joke, requiring numerous tries to understand things right, has come to an end. According to Microsoft, speech recognition technology is now on par with humans... or may be even better.

Find the secret "Ok Google" voice commands you didn't know about

Voice commands are slowly becoming more popular, but do you know everything you can ask Google assistant to do for you?

Hound promises to better understand what you tell your smartphone to do

Do you love interacting with your smartphone using your voice? Do you hate finding out just how stupid the current digital assistants are, proving to be useless as soon as you stray a bit from the formal inquiry formats they're expecting? Well, maybe it's time to try out this new Hound app.

Google Now knows a lot more than Siri and Cortana

Siri, Google Now, Cortana... it seems all major companies are working hard to make our smartphones answer all our questions. But... which one does the best job? Some folks decided to put it to the test, with over 3000 questions, and the results show a clear winner.

Google revolutionizes computer programming with Talk-to-do

Some say that the language everyone needs to learn these days is not a "human language" but a computer language. Programming is becoming essential in our day and age, where computers have the potential to do everything we want. But not everyone is able to speak "computer language", and even when they do, they often fail to translate their ideas into something computers can understand. But soon, all that can change... and anyone and everyone might be able to make their dreams come true.

Just the other day we saw how anyone can do complex tasks using easily understandable common language. Now, with this new Talk-to-do project by Google, thing will get even simpler.

Instead of having to learn a computer language, this project allows you to simple talk to your computer and tell it what you want to do! You can thing of it as a mashup of wolfram dynamic language, Google Now voice actions, Siri, IFTTT, and more - all mixed together to make whatever you want come true.

Want to make a program that alerts you whenever a Facebook friend goes out of town? Simply say something like: "Send me an email whenever a Facebook friend goes out of town". Yes, it's that simple.

In a closed session demo, Google engineers showed how you could easily code a pac-man style game using nothing but your voice (even the graphic elements were created by simply describing it to the computer - and the system is intelligent enough to ask you for added details whenever it needs to do so).

This will definitely revolutionize the way people interact with computers, and I can't wait to see what "regular people" will be able to code using Talk-to-do when Google opens it up to the public. Computer programming will never be the same again.

Google brings "Ok Google" hotword to the Desktop

If you're an Android user and have grown fond of the "Ok, Google" to initiate voice commands and queries, you'll be pleased to know that you can now do the same in your desktop or laptop, as long as you're using Google's own Chrome browser.

Thanks to the Google Voice Search hotword Chrome extension (beta), you can now enjoy the same handsfree operation for every thing Google does. I find it particularly liberating to use my Android to do quick unit conversions as well as USD to EUR and back, and lots more. I just wish that the Nexus 5 could listen even while "sleeping" as the Moto X does. Likewise, with this Chrome extension, you'll only be able to say Ok Google when you're a Google search page... which I also find to be quite limiting. It would make much more sense to be able to ask Google things even while we're navigating other sites - results could be simply spoken, or open up a new results tab page.

But I understand Google needs to take it slowly, particularly when there's so much going around regarding potential privacy invasions, and such. At least this way Google can assure users it's only listening when you're on the Google page, and it's sent to Google's servers only after you've said the Ok Google (or Ok Computer). Though I think it's only a matter of time till the voice listening mode becomes system wide.

... When it does, I can already imagine someone entering an office and loudly saying: "ok google, search for porn..." and have dozens of computers popping up not so suitable pages on their screens. Or maybe even a "Ok Google, shutdown computer"... before quickly running out.

Google Now gets to be more conversational

Google has been improving its Google Now "assistant" a lot, and now we get to have a few more cards and a more talkative voice recognition mode. Until now, if you asked Google Now to send a message, you'd be thrown into the message app without further ado. With the latest update, Google Now will ask you for additional information, like who you want to send the message to, etc.

This isn't the only thing that has been done, but it's definitely one that will amaze anyone not used to using voice commands/questions. (Yes, it's a bit more like Siri... but what's the problem with that?)

We also get a bunch of new cards, including ones that notify you when there's has been a change to a web site with the Website update card; a News topic card that keeps you up to date on whatever topic you may want; the The What to Watch card will notify you of movies and TV series; and you also get to see traffic incidents on your commute using real-time data from Waze, set repeat reminders, get real time scores for rugby teams, and also see when your packages are ready for in-store pickup.

What's not to like? :)

Siri Learns How to Pronounce Names in iOS7

If you're an avid Siri user, you've certainly come across some contacts' names that Siri insists on not recognizing and//or reading aloud the wrong way. It isn't easy dealing with names, particularly foreign name that even us humans might need some training to say correctly. In iOS you can already add a phonetic name entry into your contact list, but in the upcoming iOS7 Siri will be intelligent enough to ask for your help to learn how to say those strange names.

Considering most of the time we use Siri to send call someone or send message to someone, correctly detecting who we're talking about is one of the most important aspects. I was forced to change some contacts names in order for Siri to work properly - as it sometimes got "stuck" in deciding between two different contacts with the same first name (and when she asked which one I wanted, she would get back into telling me she had two contacts, and so on, and so on).

Hopefully that won't be the case with iOS7 - even though what I really would like to see was Siri finally expanding into lots more languages... such as Portuguese. Apple seems to forget that since it introduced Siri with the iPhone 4S nearly two years ago, lots of international customers haven't yet been able to take advantage of it, unless they're willing to speak english, german or french. Even Google, that got to the party later on has already introduced lots more voice recognition languages (including portuguese) and offers weekly improvements on Google Now - we're now getting the latest news and movie info on our nearest cinemas.

Apple really should remember there's a "rest of the world" whenever it launched lots of "new features"... that only work in the US.

Siri's About to Learn new languages?

Siri might be quite helpful, but there's still a small "big" step that Apple needs to address and where it has been failing for a very long time: there's a whole world out there... and there are hundreds of different languages Siri will need to understand to go global.

Google is doing a much better/faster work in that area, with its Google Voice Search already working with dozens of different languages, and although in my case I'm interested for Siri to hear me in Portuguese (seems that is also in the works, as it's about to reach Brasil), it seems that Apple is finally speeding up Siri's polyglot abilities.

At least that's what we can expect from the job posting where Apple says it needs native/fluent people in Arabic, Danish, Dutch, Finnish, Norwegian and Swedish.

I would feel a lot more comfortable if Portuguese was also in that list... but... I'll hope for the best.

Google Adds Voice Response to iOS App

Android users can already see it in action, but it will soon arrive to Google's Search App on the iOS as well: the spoken search results... Siri style.

Sure, iPhone users have had Siri for quite a while now, but this Google Search App will surely show them that there are things it can do better than Siri itself. For starters, it will be available in iPhones without Siri. Second, it will support more languages (at least in the Android version, it does... and I don't see why it won't do the same on the iOS). And last but not least... its voice recognition technology will certainly be appreciated by all. No disrespect to Nuance's amazing voice tech, but... seeing the voice being transcribed in real-time, and not just "after you've finished speaking", it's really great - particularly when you see it in action, correcting already recognized words due to the context of later words.

In a time where Apple is cutting all ties to Google... it's "interesting" to see Google offering more and better Apps to the iOS... including this one that goes head to head with Siri search results... and - dare I say so - beating it to the punch. Apple better have some very nice tricks up Siri's sleeves for the iOS6/iPhone 5 presentation... :)

Swype gets Voice Recognition

If you've been around for a while, you'll know that I'm a very big fan of Swype ever since it first showed up, nearly 5 years ago (and more recently, of TouchPal as well). If you don't know what Swype is, it's that amazing touchscreen keyboard where you can simply "draw" the word you wish, swiping your finger across the letters, and the words pop up as if by magic.

Now, Swype is going a step further, and adds voice recognition with Dragon Dication!

For now, it's limited to just a few languages, (the usual: english, french, german, italian, spanish) but Swype promises many more will be added in 2012 - which considering the number of languages Dragon Dictation supports, shouldn't be that hard to do.

Other than that, this new beta version also adds a improved language modeling algorithm, designed to increase the accuracy of its suggestions (and that learns over time), and also download any of the 50+ additional languages it supports directly from Swype's Settings window.

Google Beatbox

Who would have thought... you can use Google translation voice service to make Google beatbox to any rhythm you want! :)

If you want to try out your own beat, here are the secret phonemes that make Google beatbox:

• zk = suspended cymbal
• bschk = snare
• pv = brush
• bk = bass
• tk = flam1
• vk = roll tap
• kt = flam2
• kttp = flam tap
• krp = hi hat tap
• pv = short roll
• th = better hi hat
• thp, ds = instant rimshot

Lips for Gmail

How do feel about being able to write or answer to your emails using nothing but your voice - and not requiring any fancy Flash plugin or operating system wizardry? That's exactly what this Lips for Gmail by Nuno Morgadinho does, using nothing but pure HTML5 - thanks to the HTML5 Speech Attribute - and works rather nicely in Chrome.

As Nuno himself says, it's kind of sad that this sort of thing was already possible in 1990 in the NeXT computer (YouTube link) and to this day is has yet to become "mainstream".

Android Voice Actions and Chrome to Phone

Google showed off some new features for their Android platform.

The first one is Chrome to Phone, allowing you to send any piece of information from your Chrome browser directly to your Android device: a link, youtube video, phone number, or any simple text selection.




The second one is a lot more interesting and amazing: Voice Actions for Android. These action take voice control to the next level, allowing you to call people, find your way to anywhere, write emails and notes, find any music you want, and even set your alarm clock - all just by saying it!



It really looks like we're finally getting to the point we're actually carrying "digital assistants" in our pockets... :)

Voice Search

Search widget

YouTube Auto-Captions

Remember that amazing project Google was developing that would allow the automatic creation of captions on YouTube via voice recognition?

Well, this auto-caption service is now available to all english spoken videos.



I just hope we don't have to wailt long for other languages to be included.
(And just imagine all the "data" this will give them, in order to provide an instant voice translation service in the future! :)

Voice Recognized Captions in YouTube

Amazing! YouTube will get auto captions via voice recognition.



It's only available in English for now but... it's an excellent first step!

Accessibility

This Sunday, I'll be talking about accessibility, because not everyone can "surf" through the web (or the real world) as easily as most of us do, and a simple image, sound, ou improperly placed tag can become a major hurdle for someone with impaired vision, sight, or motor abilities.

Let's start with the:

Intel Reader

I know eBooks are all the rage right now, but considering how some "stupid" rights management issues are - making it impossible for you to listen to an ebook reader, for instance - maybe it's not such a bad idea to stick with the plain old paper books for a while longer.This Intel Reader scans and does OCR on it, and reads it back aloud.



It's a pity it costs $1,500 USD! As I find it hard to believe you wouldn't be able to do the same with the hardware you get in a $350 netbook.


Visually impaired Google search

Google has a visually impaired search page allowing you to quickly and easily browse through search results using the keyboard and better integrated with screen readers.
It's an experimental feature for accessibility enhancements but it might be worth it - try it out.

Freedom Leg - crutches B-gone


If you ever had to hop around in crutches, see how much better you'd be with one of there Freedom Leg devices. With the billions spent in medical research each year, why didn't they come up with this sooner?

Speaking without Talking

If speaking is no problem for most of us (some might even have a problem not speaking!) there are thousands of people out there where every single word is a pain.

Thanks to Audeo they'll soon be able to speak just by thinking about it. The tiny electrical impulses from your brain will be picked up by an electronic device and spoken alound by a digital voice.

I can already imagine a cell phone manufacturer wanting to use this technology to power up their hands-free device. Imagine to be able to talk on your cellphone in the middle of a restaurante (or even inside a movie theater) without anyone listening to you! :)


Painting with you Eyes

Thanks to the EyeWriter, which you can build yourself using cheaply available components like PS3's EyeCam even paralyzed people can interact with a computer using their eyes.


The Eyewriter from Evan Roth on Vimeo.

And to think so much money is often wasted in "professional medical devices" costing thousands more and doing infinitely less!

vLingo - Voice Recognition for the iPhone

Although there are over 10,000 Apps for the iPhone available at the App Store, most of them aren't worth a thing. Thankfully that isn't the case with vlingo.

If you recall Google Search for the iPhone recently introducing voice recognition to search for stuff without typing,  this vlingo takes the concept even further...

For starters, it completely replaces Google App by allowing you to search for anything on the web (using either Google or Yahoo) by simply saying:
web search "whatever I want to search"

Of course, it does more.
Just say:
find italian restaurants
to see what's available near you, on the map.

If you're into social websites, like Facebook and Twitter, you'll be pleased to know that you can now update your status without typing a word. Just say it:
Twitter This voice recognition posting is really awesome.

And last - but not least - it also handles voice dialing.
Just say:
call John at Home
or
call John at the Office

A must have app that should have come with the iPhone right from the start!

I hope they can improve this great App even more, by adding emailing capabilities (emailing while driving would be awesome!) and maybe some text-to-speech verification to allows us to review what we just said to ensure it was properly recognized (so we wouldn't even have to look to check it got it right.)