The iPhone 5S and 5C

Tuesday, 17 September 2013

The 5C

Let’s consider the new iPhone 5C¹ first, because it’s easier to understand.

On the inside, it’s an iPhone 5, with a few relatively minor upgrades. (To wit: the cellular antenna now supports more LTE bands and faster LTE speeds, and the front-facing FaceTime camera is better.) Otherwise, it’s the same A6 processor, the same iSight camera, the same everything as in the iPhone 5, so far as Apple has claimed and I can tell. I’ve run a few benchmarks (see below), and on each, my 5C review unit scored nearly identically to my personal iPhone 5. And the 5C camera, to my eyes, produces identical results to that of the 5.

That is not to say the iPhone 5C is uninteresting. It simply is not interesting in terms of tech specs.

Any praise or criticism of the 5C is thus about fashion, branding, and marketing. These are perfectly valid criteria. Fashion and branding are about emotion. The 5S is another engineering triumph for Apple (and no slacker in the fashion/branding game either). The 5C, though, is purely an emotional play — and, I think, a winning one.

If Apple had stuck to its playbook from the past few years, the 5C would not exist, and instead, the year-old iPhone 5 would have hung around for another year, at $99 on contract, with 16 GB of storage. Engineering-wise, the afore-noted exceptions aside, they’ve stuck to the old plan. But marketing and branding-wise, they’re in all new territory. The mid-range iPhone model is no longer an afterthought, but instead a full-fledged family member, with its own TV commercials, two storage/pricing tiers (16 GB for $99, 32 GB for $199, on contract), and most importantly, a distinctive new appearance and brand.

Apple has made colorful devices before. Other companies have made colorful touchscreen smartphones before. But there’s never been a colorful iPhone before. There is now, and people are noticing.

A few times this week I’ve repeated the notion that in marketing, what looks new is new. The technically-minded may well view the 5C as a “rebadged iPhone 5”, but the overwhelming majority of people are not technically-minded, and to them, if it looks like a new iPhone, it is a new phone. To paraphrase Stanley Kubrick, the truth of the iPhone 5C is in the feel of it, not the think of it. Most people make purchasing decisions based as much or more on emotion than logic.

My review unit from Apple is the pink one, with a blue case, with 32 GB of storage and running on Verizon’s network. (I did not ask for any specific colors or network.) Pink is the last color I’d choose for myself, personally, but I think it’s a fine pink, and likely to prove incredibly popular. Like all the 5C color choices, it looks bright and friendly. It feels slick but not slippery. Feel-wise it’s not too dissimilar from the old 3G and 3GS (both of which I still have sitting here in my office), but it presents a far more premium overall effect than those previous forays into plastic iPhones. The 3G/3GS had more seams (because of the metal bezel between the plastic back and front touchscreen), and those seams were more noticeable. The 5C has just one seam, between the plastic and the glass, and that seam is very tight. The 5C is not as thin as the 5 or 5S, but it’s so much thinner than the 3G/3GS it’s not even funny. Side-by-side it’s hard to believe the 3GS is only four years old. The 5C buttons — power, ringer toggle, volume, and home — all feel good, with nice crisp clickiness.

There’s not much more to say about it. I predict it’s going to be a huge hit, and that if anything, the iPhone lineup was overdue for something other than monochromatic color options. (It is interesting to me that Apple went with a black face for all the iPhone 5C colors, but last year went for a white face for all but the gray iPod Touches.)

And if I’m right, and the 5C proves popular, there’s an operational win for Apple as well. The 5C’s components are mostly those of the year-old iPhone 5 — a phone Apple already knows how to produce in very high volume. Though the case is new, surely plastic cases are easier (and cheaper) to fabricate than aluminum ones. The iPhone 5C could be the first new iPhone for which Apple has no problems meeting demand — not because demand is low, but because supply is easier to achieve.

With the iPhone 5C Apple may well have created what will prove to be the most popular smartphone in the world, based almost entirely on year-old technology, distinguished only by its colorful plastic casing — yet still sold at premium prices compared to the rest of the industry. Not bad.

Defining Innovation Down

You can’t swing a stick and not hit some Apple bear beating the drum that Apple is no longer capable of innovation. They could do it under Steve Jobs, can’t do it without him, and the proof is in the pudding. Under Steve Jobs, Apple released innovative products like clockwork: iPod, iPhone, MacBook Air, iPad. Under Tim Cook, nothing but the same products, slightly improved.

What a pile of crap.

Here’s the thing. The arguments of the “Apple can no longer innovate” bears could be applied just as aptly to the products Apple released under Steve Jobs. Imagine if Jobs had left the company in late 2007, after the release of the original iPhone — the one product that at least in hindsight everyone can agree was truly revolutionary. Now further imagine that Apple had then gone on, in Jobs’s absence, to do everything else exactly the same as they did in the real world, where Jobs remained CEO until 2011.

The iPhone 3G in 2008? Nothing but the original iPhone with a plastic case — an aesthetic step backward from the gorgeous metal of the original iPhone — and the addition of 3G cellular networking — which these naysayers all argued the original iPhone should have shipped with in the first place back in 2007. The 3GS in 2009? Just a slightly faster 3G. No innovation in two years. The original iPad, now hailed as one of the Jobs-led Apple’s historic innovations? In reality, it was widely panned upon release as “just a big iPhone”. Take that response to the original iPad and multiply it by jackass-finity to measure the response in my hypothetical scenario where Jobs had left Apple in 2007.

It might have been hard for anyone to argue that the retina display in 2010’s iPhone 4 wasn’t innovative, but Apple’s naysayers — in the real world, not my hypothetical one — looked right past it as they worked themselves into a frenzy over Antennagate. Now imagine how Antennagate would have played out if Jobs had then been out of the company for three years. These critics would have apoplectically jumped from “Apple can’t innovate without Steve Jobs” to “Apple can’t even make a functioning cell phone antenna without Steve Jobs”.

The square-pegged facts have been hammered to fit the round-hole narrative, and there’s no better example than Antennagate. The GSM iPhone 4 remained on sale until this week, with no changes to its antenna, and no ongoing complaints from users.

Refinement, in the eyes of these naysayers, does not count as innovation. Only revolution counts. But the iPhone needs no revolution. It continues to sell better year-over-year, year after year, without lowering its prices. Every step of the way between 2007 and that lone original iPhone — running an OS with no third-party apps, no multitasking, not even copy-and-paste — and today’s world, where Apple is on the cusp of selling its 700 millionth iOS device and the lineup ranges from the iPod Touch to the iPhones to two sizes of iPad, has been about just that: refinement.

The iPhone 5S shows that there remains much room for refinement.

The A7: ARMv8 and 64-Bit Architecture

First, performance. Apple claims this is the biggest year-over-year improvement in computing performance in the history of the iPhone, and in both my day-to-day experience² and some benchmark testing over the past week, I have no reason to doubt them. The iPhone 5S is fast.

If anything about the 5S has proven controversial, it’s the move from 32- to 64-bit. I drew quite a bit of ire this week, from the following brief comment responding to a piece by AllThingsD’s John Paczkowski:

Indeed, there’s little to be gained from slapping a 64-bit chip into today’s smartphones, aside from being the first to say you’ve done so.

“Adding 64-bit processor capabilities adds nothing to the user experience today, as it would requires over four gigabytes of memory,” Patrick Moorhead of Moor Insights and Strategy, and a former executive at AMD, told AllThingsD. “Most phones today only have one to two gigabytes of memory, and it will be years before the norm is four.”

That’s nonsense. There are serious performance gains by going 64-bit. Addressing more than 4 GB of memory is not the only advantage.

I stand by that, but this deserves a full explanation.

Switching from 32- to 64-bit processing, in and of itself, on a hypothetical generic computing platform, does not inherently improve computing performance. It might, in fact, make things slower — integers are bigger, pointers are bigger, so there’s more data to move around for everything. And on some actual specific platforms, that’s been the case. PowerPC got slower when it first transitioned to 64-bit computing. The x86 platform, however, got faster, because with the move to 64-bit processing it also gained numerous other nice elements of a modern computing architecture.

ARM is not a hypothetical computing platform. It’s a specific one. And ARMv7, the 32-bit platform upon which Apple’s previous generation A6 CPU was built, is, in some ways, similar to the old 32-bit x86 architecture: an amalgamation of cruft, some of it dating back decades.

ARMv8, the architecture upon which Apple’s new A7 is designed, is a clean break. ARM’s previous instruction set dated back 20 years. ARM has always been designed for low power consumption, but 20 years ago is forever in this industry. Rather than simply adding 64-bit instructions to the old ISA, ARMv8 is a clean break designed for today’s world — and the future. From an ARM white paper introducing ARMv8:

Fundamental to ARMv8 has to be the new instruction set, known as A64; the encoding of instructions to enable an application to utilize a 64-bit machine. ARM took the decision to introduce 64-bit through a new instruction set rather than extension of an existing instruction set for many good reasons. Most notably, and probably as no surprise, because we could develop a new independent instruction set to execute code in a lower power manner than by adding instructions to the existing instruction set. Of course, for compatibility reasons, we still support the entire ARMv7 machine in the new ARMv8 architecture, but when running 64-bit software, this part of the machine is not being used, and the area of complex legacy it had built up does not need to be active when running in the 64-bit ISA, unlike other architectures where 64-bit extension was simply added to the historical complexity and legacy of their 32-bit mode. The new ISA drew upon the years of experience of building different micro architecture implementations, so again it was defined so that these new processors can be more easily optimized for low power operation — an opportunity not really offered since the first ARMv4 machine that resulted in the now legendary low power ARM7 processors.

What does this mean? It means, for one thing, that the biggest reason for the performance and power consumption improvements going from the A6 to A7 is the switch from the ARMv7 to ARMv8 architectures, not 32- to 64-bit. ARMv8’s improved instruction set alone has resulted in 15-20 percent performance gains while simultaneously using less power, from what I’ve been told by informed sources. And though Apple could have gone to ARMv8 while remaining 32-bit only, it made no sense not to go 64-bit.

There are applications today — imaging, gaming, cryptography, video and photo filters — that will benefit from 64-bit despite the fact that the iPhone 5S has just 1 GB of RAM. It should prove faster overall, even if only slightly, than a hypothetical A7 that had switched to ARMv8 but remained 32-bit only.

But the big win is laying the groundwork for the future. iOS developers should have few problems recompiling their apps for 64-bit. (My Q Branch colleague Brent Simmons on how long it took to get Vesper to compile cleanly for 64-bit: “Just a few minutes.”) Apple’s been through this transition before, with Cocoa on Mac OS X,³ and any iOS developer who didn’t see this transition coming sooner or later simply wasn’t paying attention. Many apps should be native 64-bit binaries soon. By next year, when the A7 works its way down to the mid-range iPhones, most will be. And two years from now, it’s almost certain that all new iOS devices being sold will support 64-bit. It won’t be long until Apple can consider dropping 32-bit support and going 64-bit only. By the time it becomes feasible for iOS devices to have more than 4 GB of RAM, iOS will have already been a native 64-bit platform for several years.

The idea that Apple switched to and is promoting 64-bit processing as a marketing gimmick simply doesn’t hold water.⁴

(Note too that Apple generally, and admirably, eschews spec-based gimmicks. Look no further than the 5S camera. Apple increased the area of the image sensor but kept it at 8 megapixels, choosing to make the pixels bigger, which results in better actual photos, despite the fact that megapixel count is (wrongly) considered a measure of image quality in the consumer space. Competitors jumped on the 5S megapixel count immediately.)

So, how fast is the iPhone 5S? I ran a few popular web browser benchmarks to see how the 5S compared to the iPhone 5. (I tested the 5C too, but as stated above, it performed almost identically to the 5, as expected.) The results (average of three runs, after a device reboot):

• WebKit Sunspider 1.0.1: 1.8× faster (408 vs. 716, smaller is faster)
• Google Octane 1.0: 1.9× faster (5,576 vs. 2,984, bigger is faster)
• Mozilla Kraken 1.1: 2.4× faster (5,948 vs. 13,937, smaller is faster)

To put that in context, the iPhone 5S beats my 2008 15-inch MacBook Pro by a small measure in the Sunspider benchmark (with the MacBook Pro running the latest Safari 6.1 beta). The iPhone 5S is, in some measures, computationally superior to the top-of-the-line MacBook Pro from just five years ago. In your fucking pocket.

I also ran the general-purpose Geekbench 3 benchmark. (Again, average of three runs, and not including the 5C because it was statistically so similar to the 5.) On the iPhone 5S, Geekbench 3 was running in 64-bit mode. (Given that Apple has recompiled all of iOS 7 for 64-bit, including MobileSafari, the iPhone 5S presumably benefits at least somewhat in the above web browser benchmarks as well.)

Geekbench 3 Results

Test	iPhone 5	iPhone 5S	Improvement
Single Core	722	1,416	2.0×
Multi Core	1,298	2,557	2.0×
Integer	1,465	2,876	2.0×
Floating Point	1,164	2,637	2.3×
Memory	1,235	1,761	1.4×

Primate Labs, the company behind Geekbench, publishes user-submitted results. A perusal of their Android results page shows that the iPhone 5S handily beats every Android device on the market, including tablets. Certain models of the Samsung Galaxy S4 come closest, but the iPhone 5S is almost twice as fast as the (fastest) Galaxy S4 on the Single Core test, and still 1.2× faster on the Multi Core test, despite having only two cores to the S4’s four. I asked Primate Labs’s John Poole for his take on the relative merits of the single and multi core benchmarks. He replied, via email:

The single-core score is useful for estimating the performance of non-threaded tasks. Single-core performance is arguably the most important score Geekbench 3 provides for estimating real-world performance, especially on mobile devices. Just about every task will benefit from improved single-core performance, but only properly threaded tasks will benefit from improved multi-core performance.

That said, there are certainly benefits to having more than one core even if your tasks aren’t threaded (e.g., keeping the foreground application responsive while other applications or tasks run in the background).

What I find remarkable about the 5S’s benchmarks is not that they’re the current top scores in the mobile world,⁵ but rather that they’re at the top despite the fact that Apple famously values the ratio of performance-to-power-consumption far more than performance in and of itself.

And so how is the battery life on the 5S? It’s great. I did not run any battery life tests, but I’ve used it for seven days, including several days while I was traveling and it served not just as my primary iPhone but as my primary computer. The battery has yet to fall below 20 percent. (To be clear, I charge my iPhone overnight while I sleep.) My guess is that actual battery tests will show that it lasts as long or longer than the iPhone 5, despite the 2× CPU performance gains.

One last novelty in the A7: the M7 motion co-processor. I found no good way to test it yet, though. Apple claims some uses for it built in to iOS, such as checking the network for background updates less frequently when the device hasn’t been moved in a while (e.g. if you’re sleeping), and switching between driving and walking directions in Maps when it detects that you’ve parked your car and have gotten out — but mostly the M7 seems to me an opportunity waiting for third-party developers to take advantage of it.

Touch ID

My family went to Disney World this summer, and the parks there switched recently to new RFID-based entry passes. You pair them with a fingerprint scan the first time you use them, and subsequently, you press the pass against (or just near) a sensor at the park gates, then confirm your fingerprint, and you’re in. The fingerprint scanning worked reliably, but it also consistently took a few seconds and some back-and-forth rolling on the scanner to register.

Touch ID on the iPhone 5S is nothing like that. Adding a new fingerprint (iOS allows you to save up to five, enough for both your thumbs, both pointer fingers, and one more for, say, a trusted family member) takes upwards of a minute. It’s a borderline tedious process of repeatedly lifting and placing the finger on the sensor. But it’s easy enough, and the on-screen instructions nicely guide you through it.

Once you’ve added a fingerprint, subsequent scans of that finger are nearly instantaneous. Touch ID is way faster than “fast enough”. I’d call it “I can’t believe it works this quickly” fast. It’s also very accurate — only a handful of times over the past week have I had to try a second time, and each of those times, I hadn’t really squared up my finger with the sensor. (The key is simply to make sure your skin covers the surrounding metal ring. Do that and it seemingly never misses.)

You know how iOS touch latency and scrolling performance have always been far ahead of its competition? The way you could just tell that internally, Apple had uncompromising standards for how responsive these things needed to be? That’s what Touch ID is like — it’s to all previous fingerprint scanners I’ve seen what the original iPhone was to previous touchscreen computers.

If there’s a single fundamental knock against Apple that rings true, it’s that Apple does not do well with online services. I think Siri is good and getting better — on a whim I asked her the other day who was pitching in the Yankees game that night, and she replied with the correct answer very quickly — but there’s no question that Google Now is faster. There’s too much latency with Siri.

Touch ID’s extraordinary performance and accuracy fit right into that story. There’s nothing “online” about Touch ID. Instead it’s what Apple does best: a complete experience hosted entirely on the device. Your fingerprint data is not just “not stored in iCloud yet”, it is not stored in iCloud by design, and according to my sources, never will be. iOS, even the system itself, cannot read from or write to the secure storage location where fingerprint data is stored — only the Touch ID hardware sensor itself can. And what is stored in that secure location are not fingerprint images, but cryptographically hashed values, unique both to your finger’s biometric data and the device itself on which you scanned it. Even if someone does figure out how to obtain the fingerprint data from the secure storage on your iPhone, that fingerprint data should prove useless anywhere but on the unique Touch ID sensor on that iPhone itself — which device would have to be in the possession of your attacker/adversary in the first place for them to read the data.

I’d be far more concerned about a nefarious criminal (or, let’s face it, nefarious snooping government agency) decrypting the passwords I’ve saved on my iPhones ever since 2007 than their obtaining the fingerprint data stored by the Touch ID sensor.

In practice, Touch ID has proved incredibly convenient. I always knew that entering my lock screen passcode and App Store password dozens of times a day, every day, was a hassle, but I had no idea how much of a hassle it was until I didn’t have to do it any longer. It’s quickly habit-forming, too. After just a few days testing the 5S, whenever I go back to my personal iPhone 5, I inevitably try to unlock it with my thumbprint without thinking — then silently curse its old dumb home button as I peck my passcode.

Simply as a home button, the 5S Touch ID sensor is an improvement too. As an essential component to the iOS user experience, I’ve felt ever since 2007 that Apple’s home buttons weren’t good enough. They felt nice, but nice isn’t good enough for a button you click so often. (And on a few of my iPhones, they’ve had a tendency to get squishy or even stuck after a few months.) The iPhone home button should feel great. The 5S home button does. You can tell by touch that it’s not plastic, and its clickiness is superb. It is now flat, not concave, but it is recessed from the front face of the phone — the metallic ring is sloped downward. The new home button is an upgrade in every way, even putting Touch ID aside.

The Camera

Technically, the 5S camera appears to be an improvement, but a relatively minor one, over that of the 5 (and 5C). The sensor’s pixels are slightly bigger, and the lens has gone from an aperture of f/2.4 to f/2.2. I definitely see some improvement in low-light situations, but not radically so. The new True Tone flash sounds great, and in my testing really does noticeably improve the quality of photos taken using the flash — but personally I almost never take photos with flash, and True Tone won’t change that.

But yet in practice I’ve found the iPhone 5S is clearly a significant improvement over the 5 for photography. One reason is the new Slo Mo video mode, which is very fun. The second, and I suspect more useful reason: burst mode. Burst mode is terrific not just because the camera hardware is capable of shooting 10 photos per second, but because of what happens after that. In your Camera Roll, each burst appears as a single item, with the appearance of a stack of photos. Hold down the shutter button for 3 seconds and take 30 photos, and they’ll appear as a single stack in Camera Roll. Open a burst stack and you can pick which one (or ones) you want to keep. Those selections then become first-class photos in your Camera Roll. Then — and here’s the genius — you can delete the burst stack, getting rid of all the other photos from that burst in a single action.

I’ve been shooting photos with an SLR for over 10 years, and especially after switching from film (expensive!) to digital, I’ve often shot in continuous/burst mode. The reason: with moving subjects, shooting in a burst increases your chances of getting at least one really good exposure, where your subjects are in focus, well-framed and well-lit, with their eyes open, etc. The problem: cleaning up all the rejected photos from the burst. With the iPhone 5S, Apple has solved that problem. Burst mode alone might be the single biggest leap forward ever in iPhone photography — it’s a feature that puts multi-thousand dollar SLRs to shame.

The Slo Mo feature is also enabled as much by software cleverness as hardware capabilities. Slo Mo appears as a separate mode in the camera app; the regular Video mode still shoots at 30 fps, to keep file sizes down. When you shoot a clip in Slo Mo, after you’re done, you get a simple editing timeline where you can adjust which portion of the clip you want to appear in slow motion. It couldn’t be simpler.

The main thing that occurs to me with the iPhone 5S camera is how clearly its new features are designed for the mobile-first world. We, as a collective whole, no longer shoot, sync to a computer, edit, and then upload. We shoot, edit, and upload right from our phones (and, let’s face it, tablets). The standalone consumer camera is dying a quick death — an always-on network connection and apps for editing/filtering, and most importantly, instantly sharing are now essential components for consumer photography. The iPhone 5S embodies this better than any other device in the world.

If there’s a problem, it has nothing to do with the iPhone 5S and everything to do with iCloud. Last month I linked to a piece by Bradley Chambers, wherein he argues:

Also, photo stream needs to be reversed. Apple should store all photos/video taken with your iPhone and just store the most recent 1000 (or 30 days) locally on the device.

I agreed with him then, and agree with him now. Not a word from Apple, though, about any sort of improvement to Photo Stream.

“Can’t Innovate, My Ass”, Indeed

So what has Apple delivered with the iPhone 5S?

With the A7 they’ve doubled CPU performance in exactly one year, at no apparent cost to battery life. They’ve potentially obviated the need for standalone motion trackers like Fitbits and Nike Fuelbands. And they’ve started a transition to platform-wide 64-bit computing years ahead of their competition.

With Touch ID they’ve eliminated the need to enter a passcode to unlock your phone and a (one hopes) complex password to download apps and media from the iTunes Store. Mere conveniences, yes, but very nice ones indeed. They’ve also potentially set the stage for numerous future conveniences. Imagine Touch ID integrated with the upcoming iCloud Keychains.

With the camera they’ve created a simple, intuitive interface for taking bursts of fast action or won’t-stay-still kids and pets. It’s a wonderful hardware feat to put a 10 fps still camera in a remarkably thin mobile phone (a phone with no inelegant bulge to accommodate more distance between lens and sensor, at that). But the real innovation — there’s that word — is software, right there on the device itself, that makes it easy to select only the shots from those bursts that you really want to keep, and to throw away the rest.

This is what innovation, real innovation, looks like.⁶ It’s like the Thomas Edison quote, “Opportunity is missed by most people because it is dressed in overalls and looks like work.” Innovation is missed by most people because it is so often incremental.