Whether or not Apple releases an iPhone 4S or an iPhone 5 device in a few weeks, one of the key talking points at Apple’s upcoming media event will center on voice recognition in iOS 5.
For some time, there have been rumors that iOS 5 will feature deep integration of voice recognition technology powered by Nuance, a company renowned for its best-in-class voice recognition engine. Further, note that Apple in 2010 acquired a company called Siri for their voice-activated search technology.
Putting the pieces together, 9to5Mac has an incredibly detailed piece laying out what we can expect to see in iOS 5 from a voice recognition perspective.
First, the requisite hardware. If you wanna run iOS 5’s voice recognition feature, which has been dubbed “Assistant’, you gotta be rocking Apple’s next-gen iPhone. Yep, even if you’re sporting an iPhone 4, you’re out of luck. Apple’s ‘assistant’ feature apparently requires the A5 processor and 1GB of RAM Apple’s next-gen iPhone will house.
Assistant taps into many aspects of the iPhone, according to people familiar with the feature and SDK findings. For example, one can say make appointment with Mark Gurman for 7:30 PM and Assistant will create the appointment in the user’s calendar. On noting events, Assistant also allows users to set reminders for the iOS 5 Reminders application. For example, a user could say “remind me to buy milk when I arrive at the market.” Another example would be integration with the iOS Maps application. A user could ask: “how do I get to Staples Center?” and Assistant will use the user’s current location via GPS and provide directions.
Another interesting Assistant feature is the ability to create and send an SMS or iMessage with just your voice. For example, you can say “send a text to Mark saying I’ll be running late to lunch!” – and it will send. This is a super compelling feature for people who cannot physically or safely take the time to type out a text message. Users can also choose to have Assistant read back unsent text messages to ensure the system interpreted the speech correctly. If the text is written correctly, the user simply says “yes;” if not the user says “no” and Assistant will ask the user to speak again. Apple is also working to allow users to ask for a specific song to be played. Voice Control only allows albums, artists, and playlists to be chosen with your voice.
Another interesting aspect of Assistant is that it will speak back to a user in order to glean more information or perhaps clarify a command open to interpretation. For example, if a user tells Assistant to set up a meeting with someone who has two email addresses, the feature will “speak back” and ask if a notification email should be sent to said person’s personal or work account.
9to5Mac claims that Assistant’s speech interpretation is incredibly accurate and does not require user’s to speak at a measured pace. Simply talking at a normal speed and in a normal cadence is enough to get the job done.
Another compelling feature relating to Assistant is “Find my Friends.” Yes, Find my Friends is real and Apple may have been waiting on Assistant integration to release it into the wild. Basically, if a friend of mine wants to instantly know where I am, they could say “where’s Mark?” and Assistant will locate me and share my location information.
Other features included in Assistant include Wofram Alpha integration and Nuance powered speech to text functionality which reportedly works blazingly fast.
All in all, even if Apple only releases one new iPhone model this year – and an iPhone 4S type device to boot – the upcoming Assistant feature seems to be all that and a bag of chips and should keep Apple fans beholden to the iPhone upgrade cycle yet again.