Now that iPhone X is in our hands, how does Apple’s engineering match up to its marketing? Can it really justify a £1K starting price tag? Read on to find out!
The original iPhone remade the world by making the Internet pocketable and above all accessible. Its design partnered a then huge 3.5-inch multi-touch screen with a single iconic home button.
iPhone X ditches the home button in favour of an all-screen design marked with a distinctive sensor notch. The physical button gives way to sensors that look for you and enable X to react to your attention.
If the first ten years transformed the world within the bounds of the mobile screens we all walk around staring into, the next ten years will be about smartphones getting smarter and more responsive to us.
Face ID in operation
It’s faster and simpler to set-up than Touch ID is for a single finger print. Simply move your face within the target circle twice and that’s it. Touching the screen anywhere brings it to life and, by default, attention awareness is on which means seeing your face isn’t enough, it also needs to detect you looking at it.
Typically, you pick up the phone swiping upwards to open it with the expectation Face ID will do its thing without you even being aware of it.
The infra-red cameras means this works at night as well as day. I set it up while wearing my glasses, but it didn’t matter when I removed them or if was brushing my teeth. It’s like not having any security at all, mostly.
If the iPhone is flat on a table, then the acute angle can defeat it. So you either lift it up or tap in your passcode.
Since the experience of using X begins with the upward swipe, you soon get into the habit of picking it up or leaning closer.
Occasionally it’ll ask for a passcode if uncertain, but the overall experience is so new and magical it’s easy to forgive occasional foibles.
It’s cool to see your notifications expand simply because the X recognises you looking at them. Visit a website that requires your password and there’s no need press your finger onto TouchID. iPhone X knows it’s you, calls up your password and enters it with a slick little FaceID glyph animation. Put on your Apple Watch and it unlocks from your glance into FaceID.
It’s unclear whether Animoji actually require the full FaceID sensor suite, but the way in which various cartoon characters mirror your facial expressions with zero lag is both entertaining and fascinating.
How long before an iPhone is scanning your face and reading your emotions? We so easily anthropomorphise online chat bots, what if they could soon respond to your smile or frown?
Prior to iPhone X’s launch, skepticism abounded about its speed and accuracy. A few weeks later, rumours abound of all next year’s iPhone’s being all-screen designs with sensor notch. The 2018 iPads will also have notches, even some future iteration of the delayed HomePod speaker system. Perhaps.
TouchID was the gold standard for biometric authentication, but that’s all it ever was going to be, whereas FaceID is potentially much more.
AI & A11
The near instantaneous reading of your 3D face is dependent not just on multiple sensors, but also Apple’s fifth generation SOC - the A11 Bionic processor. Design work began over three years ago and included the neural engine that FaceID relies upon.
Neural networks or Deep Learning software excels at tasks such as visual recognition, speech recognition and speech synthesis. It’s at the heart of FaceID, but is also well suited to Siri and AR applications. Neural network code isn’t written by humans, but is instead machine driven based on programmed constraints or weighting's intended to drive the code to the desired goal.
Deep Learning code first appeared in iOS 10 for face detection in the Photos app and to match what Google was doing with cloud-based services demanded cutting edge on-device processing. Face ID for the first time shows Apple moving ahead of Google in this area.
Other than the all-new Neural Engine, the A11 builds upon the A10 Fusion by adding two more cores, for a total of six with support for asymmetric processing. The A11’s two high performance cores, four high efficiency cores, can be called on whatever permutations suit the application at hand.
On some benchmarks, the A11 can now outperform a 13-inch MacBook Pro with an Intel Core i5-7267U. Benchmarks are not real life, but iPhone X feels blazingly fast. Swiping through a deck of typical apps can take you from watching HDR video in Netflix to advanced image editing in Pixelmator to editing 4K video in H.265 using Lumia Fusion.