Even thought it’s technically still a beta function, Siri is to many the hero feature of the iPhone 4S, through sheer force of futuristic dazzle if nothing else. It’s a voice control function, essentially, letting you speak commands to your phone, but presented with Apple’s trademark polish.
The iPhone 4S’s voice-recognition feature works by recording your voice and sending it to a server that interprets what you’ve said and returns plain text. If you haven’t got an internet connection, Siri won’t work. But it doesn’t require a strict vocabulary – if talk like Yoda even you try, it’ll generally figure out what you’re trying to say. That one leap makes interacting with Siri seem much more natural.
In addition to the Phone and Music apps, Siri is tied into Messages, Calendar, Reminders, Maps, Mail, Weather, Stocks, Clock, Contacts, Notes, and Safari. It’s also linked to Wolfram Alpha, the “computational knowledge engine” that can provide answers to numerous factual questions, and Yelp, the directory of local businesses. And when all else fails, Siri will usually suggest that it perform a web search for you.
There are two different reasons for Apple choosing to designate Siri as a beta product – one that’s still officially being tested. First, Siri currently only supports English (in US, UK and Australian dialects), French, and German – and that’s not enough languages for Apple and its worldwide customer base. Second, the company needs to expand the number of apps and information sources to which Siri is connected.
The truth is, once you start using Siri in earnest, you’ll discover where its boundaries are. It’s great at working with text messages, but not with email. It knows a lot about weather and restaurants but nothing about movie times. Apple says that understanding the words you say is the easy part, and that Siri’s true genius is in figuring out what you want when you say those words and getting you the answer. If that’s true, Siri needs to be tied into many more information sources and apps. (Including third-party apps, which are not capable of tying into Siri today.)
There are two scenarios in which Siri truly excels. The first of those is when you’re in a hands-free scenario, mostly likely when driving a car. (The iPhone 4S knows when you’re in a hands-free situation and becomes more chatty, reading text aloud that it might not if it knows you’re holding it in your hand.) When you get a text message, you can instruct Siri to read the message, and it will. You can then tell Siri to reply to the message, dictate the entire message, have Siri read it back to you to confirm that it makes sense, and then send it.
It’s a major step forward, though there are gaps. Siri can tell you that you have new email, and you can use it to send emails, but it won’t read your emails to you. (It’ll only read text messages aloud.) And while iOS 5 adds the nifty Notification Centre, which gives you granular control over how different apps notify you about what’s going on, there’s no option to read alerts out loud when you’re in hands-free mode. A missed opportunity.
If you aren’t driving, Siri can still be useful: In fact, the feature proves that some tasks can be done much faster through speech than through clicking, tapping, and swiping. It’s much easier to set an alarm or timer using Siri than it is to unlock your phone, find the Clock app, and tap within the app. Just say “Set a timer for three minutes” and your phone begins to count down until your tea is ready. “Set an alarm for 5am” does what you’d expect, instantly. “Remind me to record my favourite show” and “Note that I need to take my suit to the cleaners” work, too. These are short bursts of data input that can be handled quickly by voice.
I was impressed by Siri’s ability to understand the context of conversations. It didn’t always work, but when it did, it was magical. I asked Siri for suggestions for places to have lunch, and it provided me with a list of nearby restaurants that serve lunch. I then specified that I wanted to eat downtown, and got a narrower list of places. This was so great, I tried to repeat the task later… and could never get it to work again.
Of course, talking to your phone is not much different from talking on your phone: it’s not appropriate in all contexts. If you’re quietly reading in the library and need to set a reminder, you should use the Reminders app, not Siri. And if you’re out in public, well, you can use Siri, but you do risk people looking at you funny.
Apple’s integration of Wolfram Alpha with Siri was a smart move. If you need answers to factual questions, like the speed of light or the number of days until Christmas, Wolfram Alpha can provide the answer. Unfortunately, these results come in the form of images, not parseable text, so Siri can’t actually read you the reply. I was also disappointed that when Siri gives up and searches the internet, it wouldn’t walk me through the search results and read their summaries. Most of my answers were in the first few search results summaries, but if I were driving they’d do me no good.
While Siri gets the bulk of the iPhone 4S feature hype, another speech-related technology may prove to be more important and a bigger boost to users’ productivity: dictation. Following the lead of Android, the iPhone 4S can now convert what you say into written text in any app.
Here’s how it works: On the keyboard you’ll see a new button in the bottom row, to the left of the spacebar, with the image of a microphone on it. Tap this button and the iPhone 4S will transcribe whatever you say. It sends the results over the internet to a server that analyzes your speech and converts it into text (meaning that if you’re not online, the microphone button will disappear). I was shocked at just how fast the results came back, especially over Wi-Fi. And they were generally, if not always, accurate.
To get the most out of dictation, you’ll need to start thinking in punctuation. For example, to construct a decent email message, I might say, “Dan comma new paragraph what do you think about writing a review of iOS numeral five question mark I think it might be right up your alley full stop new paragraph let me know what you think exclamation point.” The thing is, it works.
A big step forward
Speech-recognition powerhouse Nuance has offered this feature in an iOS app, Dragon Dictation, for some time. But being able to input text by voice in any iOS app is a big step forward. However, the existence of Dragon’s app makes me wonder why this is apparently an iPhone 4S-only feature. Older iPhones could run Dragon’s app. And, come to think of it, Siri was an iOS app before it was acquired by Apple.
I understand that Apple’s in the business of selling iPhone hardware, and that Siri and dictation are both features that will drive sales of iPhone 4S. But I admit that I wonder if the decision to have these features work only on the iPhone 4S is rooted more in sales strategy than in technology. I can almost understand the idea that Siri might require extra processing power, but speech-to-text? That’s a feature that should be available for iPhone 4 users as well.