The New Apple Vision Pro: A Spatial Computing Breakthrough

Den3...9PQb
3 Feb 2024
73

Apple Vision Pro review: magic, until it’s not


The Apple Vision Pro is the best consumer headset anyone’s ever made — and that’s the problem.
By Nilay Patel, editor-in-chief of the Verge, host of the Decoder podcast, and co-host of The Vergecast.


If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.


There’s a lot of pressure on the new Apple Vision Pro, Apple’s long-awaited entry into the world of computers you wear on your face. Apple claims that the Vision Pro, which starts at $3,499, is the beginning of something called “spatial computing,” which basically boils down to running apps all around you. And the company’s ads for it do not hedge that pressure even a little: they show people wearing the Vision Pro all the time. At work! Doing laundry! Playing with their kids! The ambition is enormous: to layer apps and information over the real world — to augment reality.
Apple has to claim that the Vision Pro is the beginning of something new because people have been building headset computers for over a decade now. I tried on a development prototype of the first Oculus Rift in 2013, and The Verge’s Adi Robertson, who edited this review, has tried basically every headset that’s been released since. All of that development means there are some pretty good products out there: that first Oculus evolved into the Quest line at Meta, which is now shipping the Quest 3 — a very good VR headset with a huge library of games and some AR features of its own, which costs $500.

7

Verge Score

Apple Vision Pro

$3499THE GOOD

  • Display is a technical marvel with the best video passthrough yet
  • Hand and eye tracking are a leap forward
  • Works seamlessly with Apple’s ecosystem
  • Fun to put windows all over space

THE BAD

  • Very expensive
  • Video passthrough is still video passthrough and can be blurry
  • Hand and eye tracking can be inconsistent and frustrating
  • Personas are uncanny and somewhat terrifying
  • It’s pretty lonely in there

$3499 AT APPLE
How we rate and review products
In the meantime, Apple, from Tim Cook on down, has largely insisted that augmented reality will be much more valuable than virtual reality. And it’s been building toward AR for a long time: developers have access to AR tools in iOS, and higher-end iPhones and iPads have had lidar depth scanners for a few years now.
The Vision Pro is Apple’s first attempt at building a computer out of all those ideas — a computer that works in the space around you. The goal is for the Vision Pro to be a complete device that can sit right alongside the Mac and the iPad in Apple’s ecosystem of devices and let you get real work done. You can use Excel and Webex and Slack in the Vision Pro, and you can also sit back and watch movies and TV shows on a gigantic virtual 4K HDR display. And you can mirror your Mac’s display and just use the Vision Pro to look at a huge monitor floating in virtual space.

Related

It sounds amazing, and sometimes it is. But the Vision Pro also represents a series of really big tradeoffs — tradeoffs that are impossible to ignore. Some of those tradeoffs are very tangible: getting all this tech in a headset means there’s a lot of weight on your face, so Apple chose to use an external battery pack connected by a cable. But there are other, more philosophical tradeoffs as well. 
As I’ve been using it for the past few days, I kept coming up with a series of questions — questions about whether the tradeoffs were worth it.
Is using the Vision Pro so good that I’m willing to mess up my hair every time I put it on?
Is it so good that I want to lug it around in its giant carrying case instead of my laptop bag?
Is it so good that I want to look at the world around me through screens instead of with my own eyes?
Basically, I keep asking if I prefer using a computer in there rather than out here. And as interesting as the Vision Pro is, there’s a long way to go before it can beat out here.
Photo by Vjeran Pavic / The Verge

Hardware

Apple doesn’t want anyone to think of the Vision Pro as a VR headset, but it’s a VR headset — albeit a VR headset that almost lets you pretend it’s not a VR headset. 
You put it on your head in a way that blocks out your vision entirely, and then it shows you a 3D video feed of the world around you passed through from the cameras on the front, as though you can see right through the device. But it can also put you in virtual reality, at various levels of immersion: I spent some time working entirely on the Moon and a lot of time in my kitchen with a bunch of windows floating around a portal into Joshua Tree. 
Defining “reality” is a messy business in computing. The past decade has seen a dizzying race to come up with new words for what head-mounted displays do, including disagreement over what each one of these phrases means. So here’s our interpretation of a few terms used in the review:
Augmented reality: virtual projections that are directly related to objects in the physical world, like an automatic translation of a restaurant menu or a virtual poster pinned to a real wall.
Mixed reality: a computing system that mixes the virtual and physical without direct interaction between the two, like an app window floating in your living room.
Virtual reality: a computing experience that immerses you in a fully virtual space, deliberately blocking out your perception of the physical world.
The Vision Pro is stunning compared to other VR headsets, which are largely plastic and often downright goofy-looking. The Vision Pro, by contrast, is built of magnesium and carbon fiber in an aluminum enclosure that feels like a natural extension of Apple’s familiar design language. There’s a little iPhone 6 in there, a little AirPods Max, a little Apple Watch. It is the cutting edge of technology in a package that seems instantly familiar. Almost everyone I’ve shown it to thinks it looks smaller in person than they expected, especially compared to some of the huge VR headsets we’ve seen over the past decade.
The front display on the Vision Pro is an attempt at keeping you from being isolated from other people while you’re wearing it. In Apple’s photos, it looks like a big, bright screen that shows a video of your eyes to people around you so they feel comfortable talking to you while you’re wearing the headset — a feature adorably called EyeSight. In reality, it might as well not be there. It’s a low-res OLED with a lenticular panel in front of it to provide a mild 3D effect, and it’s so dim and the cover glass is so reflective, it’s actually hard to see in most normal to bright lighting. When people do see your eyes, it’s a low-res, ghostly image of them that feels like CGI. The effect is uncanny — the idea that you’ll be making real eye contact with anyone is a fantasy. And there are no controls or indicators in visionOS for this external display, so you never really know what other people are seeing. Imagine looking someone directly in the eyes and talking to them without knowing if they can see your eyes — it’s weird!
EyeSight is so weird. So, so weird. Photo by Amelia Holowaty Krales / The Verge
That cover glass hides a huge array of cameras and sensors. There’s a pair of high-res front cameras for the video passthrough, cameras that face down and to the sides to track your hands, a lidar scanner and TrueDepth cameras for spatial tracking, and infrared floodlights so everything can work in low light. Underneath all that, you’ve got an M2 processor and Apple’s new R1 spatial coprocessor and a pair of fans to move the heat from all this tech out the top. The fans were never perceptible during my time testing the Vision Pro, but the heat was: after long sessions, the headset was definitely warm.
On the top edge, you’ll find what feel like larger versions of some familiar Apple Watch controls: a digital crown that adjusts both the volume and the level of virtual reality immersion on the right as you look through the headset and a button on the left that lets you take 3D photos and videos.
You get two headbands in the box: the solo knit band and the dual loop band. They both attach and detach easily; you snap them right on and then pull the little orange tab to disconnect them. The solo band is unquestionably cooler and messes up your hair slightly less — but they both mess up your hair, so if the dual loop fits you better, just go with it. I found the solo loop much more comfortable; I also wish the mounting points were on the outside face so I could pull the band around my head and clip it on instead of constantly pulling it over my hair.
The front of the Vision Pro is covered in cameras and sensors. Photo by Vjeran Pavic / The Verge
The solo loop is by far the cooler strap. Photo by Amelia Holowaty Krales / The Verge
The huge case costs an additional $199 and looks like a marshmallow from space. I love it. Photo by Amelia Holowaty Krales / The Verge
The other two pieces are the light seal, which comes in various sizes, and the headband, which comes in two thicknesses. (You get fitted for these in the store or by scanning your head with an iPhone if you buy online.) Both attach magnetically, which means they also detach magnetically — you want to pick this thing up by the frame because grabbing it by the light seal can lead to disaster.
The idea that you’ll be making real eye contact with anyone while wearing the Vision Pro is a fantasy
If you have glasses, you can click in custom Zeiss lens inserts — Apple sent us reader lenses to see what that process looks like, but I just used the Vision Pro while wearing my soft contacts, and it was fine.
The Vision Pro’s speakers are housed in the arms on the side, and they are good and loud and do a convincing job of rendering spatial audio. Things really sound like they are happening where they appear to be happening, which is a neat trick. The speakers are also pretty leaky, so everyone else around you can hear what you’re up to unless you use headphones. You can use any Bluetooth headphones you want, but you get a bunch of extra features if you use the latest AirPods Pro, like lower latency, lossless 48KHz audio, and Apple’s Adaptive Audio system, which automatically mixes in sound from the real world as appropriate.
Since you’ll mostly experience the Vision Pro in there, the most noticeable thing about the hardware after a while is that it’s just… heavy. You’re supposed to wear this thing on your face for long stretches of computer time, and depending on which band and light seal you use, the headset alone weighs between 600 and 650 grams. I keep joking that the Vision Pro is an iPad for your face, but it’s heavier than an 11-inch iPad Pro (470 grams) and pushing close to a 12.9-inch iPad Pro (682 grams), so in a very real way, it’s an iPad for your face.
All of the Vision Pro’s heft is totally front-loaded, too. Other big headsets like the Quest Pro (722 grams) have elaborate headbands to balance out their weight, but the Vision Pro just rests it all on the front. Swapping to the dual loop strap helps keep things more stable but doesn’t really reduce the overall sensation of having all that headset on your face. You’re just going to feel it after a while.
The Vision Pro’s external battery pack is fine, mostly because it’s such a stationary device overall. Photo by Vjeran Pavic / The Verge
Notably, the Vision Pro is substantially heavier than the familiar Quest 2 (503g) or even the heavier Quest 3 (515g) — headsets that have built-in batteries. Apple told me that it chose to use an external battery specifically to reduce the headset’s weight. The battery itself is barely worth talking about — it’s a silver brick that weighs another 353 grams with a USB-C port and a motion-activated LED that’s green when it’s charged and orange when it’s not. It connects to the headset with a satisfying twist connector, but the nice braided cable is permanently attached to the battery itself, so don’t break it. You can buy extra batteries for $199, but you can’t hot-swap them; disconnecting the battery from the Vision Pro cuts the power entirely.
I don’t really have strong opinions on this battery setup, which is mostly because nothing about the Vision Pro feels like you’re supposed to move around that much in it, so it’s more or less fine. It’s funny that Apple of all companies shipped this compromise, but it’s also very Apple that the battery is not actually bigger so it can provide more than two and a half hours of run time. (If you plug the battery in, the Vision Pro just runs on wall power for as long as you want.)
Setting up the Vision Pro is dead simple — the headband adjustments are the only manual adjustments involved. Everything else is motorized and sensor-driven. There’s no fiddly lens wheel to dial in; the headset asks you to hold down the digital crown when you first put it on to adjust the lenses to your eyes, and then you go through a pretty standard eye tracking setup. The rest of the setup is almost exactly like any other iOS device: if you have an iPhone, you can bring it close to the Vision Pro to send over all your settings, and you have to agree to some terms and conditions. If you don’t have an iPhone, you’ll have to enter your passwords and so on manually, but it’ll work fine as a standalone device. After all of that, you’re computing in the future.
See? It’s a VR headset. Photo by Vjeran Pavic / The Verge
Apple is very proud of the displays inside the Vision Pro, and for good reason — they represent a huge leap forward in display technology. The two displays are tiny MicroOLEDs with a total of 23 million pixels that are just 7.5 micrometers in size, which is about the size of a red blood cell. And each of those tiny pixels is composed of three RGB subpixels laid out in what Apple tells me is an S-stripe pattern. Just thinking about the level of precision required to make these displays and then make them work in a device like this is mind-blowing. 
They also look generally incredible — sharp enough to read text on without even thinking about it, bright enough to do justice to movies. Apple calibrates them for color at the factory so they are also vibrant and color-accurate without looking oversaturated or blown out. They are so small, but they work so well that they seem huge.
The displays are the main reason the Vision Pro is so expensive — they’re at the heart of the Vision Pro experience and what makes the whole thing work. You are always looking at them, after all. But for all their technical marvels, they are not without tradeoffs of their own when deployed on a device like this.
The speakers are nice and loud and convincing with spatial audio. Photo by Vjeran Pavic / The Verge
See this thing — a passthrough VR headset with a silly external battery pack and a display that shows ghostly images of your eyes on the front — is not the big goal. The big goal is AR, or augmented reality. In particular, the big goal is optical AR, where light passes directly through unobtrusive glasses to your eyes, with digital information layered over the top of what you’re seeing. AR is a technology with the potential to literally change humanity, and Apple CEO Tim Cook has been talking about how isolating VR headsets are and how important he thinks AR will be for years now. 

  • Tim Cook, 2016: “Few people are going to view that it’s acceptable to be enclosed in something.”
  • Tim Cook, 2017: “Unlike Virtual Reality which closes the world out, AR allows individuals to be present in the world.”
  • Tim Cook, 2017: “I also like the fact that [AR] doesn’t isolate [...] I’ve never been a fan of VR like that because I think it does the opposite.”
  • Tim Cook, 2020: “I think [AR is] something that doesn’t isolate people. We can use it to enhance our discussion, not substitute it for human connection, which I’ve always deeply worried about in some of the other technologies.”

You get the idea.
The problem is that the technology to build a true optical AR display that works well enough to replace an everyday computer just isn’t there yet. The Magic Leap 2 is an optical AR headset that’s cheaper and smaller than the Vision Pro, but it’s plagued by compromises in field of view and image quality that most people would never accept. 
So Apple’s settled for building a headset with real-time video passthrough — it is the defining tradeoff of the Vision Pro. It is a VR headset masquerading as an AR headset. And let me tell you: the video passthrough on the Vision Pro is really good. It works! It’s convincing. You put the headset on, the display comes on, and you’re right back where you were, only with a bunch of visionOS windows floating around. 
It sounds simple, but it is an astonishing engineering achievement to do that in real time, at high resolution, in a computer that fits over your eyes. Apple claims there’s only 12ms of latency between what the cameras see and what’s on the display, and that latency includes the exposure time of the cameras themselves. The actual data processing is faster. Do the math, and Apple says there’s not even a frame’s worth of latency; the next frame of video is ready before you’re done looking at the last one.
Still cameras, still displays. Photo by Vjeran Pavic / The Verge
You can also see Apple’s incredible video processing chops right in front of your eyes: I sat around scrolling on my phone while wearing the Vision Pro, with no blown-out screens or weird frame rate issues. I also worked on my Mac in front of a large window while wearing the headset, which is a torture test for dynamic range, and while it wasn’t perfect, it was still usable. It is the best video passthrough that’s ever shipped in a consumer device by far.
If you want me to perceive reality through something, I’d like to see all the colors of the rainbow
The problem is that cameras are still cameras, and displays are still displays. All cameras have motion blur, for example. In low light, cameras either have to increase exposure times at the cost of sharpness or increase ISO at the cost of noise, which then requires noise reduction, which makes things blurry and dull. And cameras and displays both have real limits in terms of color reproduction.
The Vision Pro cannot overcome the inherent nature of cameras and displays. You can easily see motion blur when you move your head in the Vision Pro — motion blur that increases in low light and leads to some weird warping of straight lines. Low light also causes the overall sharpness of the video passthrough to drop as noise reduction kicks in: my iPhone screen got noticeably blurrier when the sun set.
If you’re in a medium-lit room halfway immersed in a dark virtual environment with a bright window open — say, sitting in your kitchen at night with the lights on while writing a review in a Google Docs window floating on a dark beach — you will notice the display brightness slowly ramp up and down as the system tries to average out the brightness of everything you’re looking at. The LCD clock on my microwave flickers when I look at it through these cameras. And Apple’s specs say the display supports 92 percent of the DCI-P3 color gamut, which means the Vision Pro can only show you 49 percent of the colors your eyes can actually see.
Video

The vision pro's video passthrough did not like the clock on my microwave.
The displays have other limitations: the field of view isn’t huge, and the essential nature of looking at tiny displays through lenses makes that field of view feel even smaller. Apple won’t tell me the exact number, but the Vision Pro’s field of view is certainly smaller than the Quest 3’s 110 horizontal degrees. That means there are fairly large black borders around what you’re seeing, a bit like you’re looking through binoculars. 
On top of that, there’s a little bit of distortion and vignetting around the edges of the lenses, and you’ll see some green and pink color fringing at the edges as well, especially in bright environments. All of this makes the usable field of view feel even smaller. If you’re looking at something bright or otherwise high contrast — a white text window floating above a dark desert landscape, for example — you’ll see highlights reflecting in the lenses.
A screen capture from the Vision Pro and a simulated view of what it feels like to look in the headset.
I asked Apple about all of this, and yep — that’s how it’s supposed to look. Apple spokesperson Jacqueline Roy told me that a combination of hardware and software in the Vision Pro is specifically designed to minimize these various effects, but they’re definitely in there, and you’ll see them.
You’re constantly being reminded that you’re looking at video on screens, and reality is a lot more interesting than that
If you have been paying attention to VR for the past decade, you know that these are very familiar VR headset display issues. You’re passing light from a screen through lenses mounted on someone’s face and trying to line those lenses up with their eyes, which are notably in different spots on different people’s faces. (Our bodies are designed very badly when it comes to mounting hardware on them.) So a little weirdness at the edges of the displays is not a deal-breaker or even a surprise — except Apple is charging $3,499 for the Vision Pro and making it sound like these displays are perfect enough for you to casually wear this thing while folding laundry. 
I’m serious when I say the Vision Pro has the best video passthrough I’ve ever seen on the sharpest VR displays any normal person will ever come across. But you’re still constantly being reminded that you’re looking at video on screens, and reality is just a lot more interesting than that. There are vanishingly few contexts in reviewing consumer devices where anyone has to care about color gamuts — but if you want me to perceive reality through something, I’d like to see all the colors of the rainbow.
This is the best anyone has ever made in there look, and it’s still not nearly as good as out here.
Waiting. Watching. Ready to mess up your hair at a moment’s notice. Photo by Vjeran Pavic / The Verge

Controls

The other thing Apple is very proud of is the eye and hand tracking control system, which is light years beyond any other consumer hand or eye tracking systems out there. You look at things you want to control, you tap your fingers to control them, and that’s how you get around the entire interface. You’re not reaching out and touching things — it’s more like your eyes are the mouse, and your fingers are the button: you tap them together to click on what you’re looking at.
The first few times you use hand and eye tracking on the Vision Pro, it’s awe-inspiring — it feels like a superpower. The Vision Pro’s external cameras just need to see your hands for it to work, and they can see your hands in a pretty large zone around your body. You can have them slung across the back of the couch, resting in your lap, up in the air with your elbows on a table, pretty much anywhere the cameras can see them. It actually takes a minute to realize you don’t have to gesture out in front of you with your hands in the air — and once you figure it out, it’s pretty fun to watch other people instinctively reach their hands up the first time they try the Vision Pro.
But the next few times you use hand and eye tracking, it stops feeling like a superpower — and in some cases, it actively makes using the Vision Pro harder. It turns out that having to look at what you want to control is really quite distracting.


Think about every other computer in your life: the input mechanism is independent of whatever you’re looking at. On a laptop, you can click on controls and use the keyboard while keeping your focus on a document. On a phone, you can do things like drag sliders in a photo editing app while keeping your eyes focused on what those changes are actually doing to your photo.
The Vision Pro simply doesn’t work like that — you have to be looking at something in order to click on it, and that means you are constantly taking your attention away from whatever you’re working on to specifically look at the button you need to press next. I spent some time playing a lovely little game called Stitch that quickly became maddening because I kept looking away from the piece I wanted to move to the place I wanted to move it, which meant I wasn’t picking it up when I tapped my fingers.
Reach out and touch faith. Photo by Vjeran Pavic / The Verge
visionOS feels also designed for an eye tracking system that’s just slightly more precise than it actually is — a lot of controls are just a little too small and a little too close together to let you quickly bop around the system. You have to look, make sure you’re looking at the thing you want, and then tap, or you might end up clicking on the wrong thing. Sometimes the fastest way to select what you want is to look away entirely and try again.
It works until it doesn’t. It’s magic until it’s not.
Think about it like this: The keyboard and mouse on a Mac directly control the Mac. The click wheel on an iPod directly controlled the iPod. A lot of work has gone into making it feel like the multitouch screen on an iPhone directly controls the phone, and when it goes sideways, like when autocorrect fails or an app doesn’t register your taps, it’s not pleasant.
Your eyes and hands aren’t directly controlling the Vision Pro: cameras are watching your eyes and hands and turning that into input, and sometimes the interpretation isn’t perfect. The best example of this is the hilarious on-screen keyboard, which you use by staring at each letter and pinching your fingers to select it or henpecking with two fingers at the floating keys in front of you. It’s not worth using for anything beyond entering a Wi-Fi password — for anything longer, you’ll want to use dictation or connect a Bluetooth keyboard. Why? So you can directly control the input.
It’s not a given that the Vision Pro can always see your hands, either. There’s a pretty large bubble around the front of your body where the cameras can see your hands — it basically extends the length of your arms in a semicircle around the front of your body. But if you lean back in a chair with your arm at your side, it can’t see your hand. If you’re sitting at a table and your hands are on your legs, it might not see your hands. If you’re lying down in a dark room and the IR illuminators can’t reach your hands, the cameras might not be able to see them. If you’re simply standing up with your arms at your sides, it might not be able to see your hands if they drift too far backward.
Photo by Vjeran Pavic / The Verge
I recognize that it is fundamentally silly to complain about a hand tracking system needing to see your hands, and it is possible to navigate quite a lot of visionOS with Siri and dictation: you can manage apps, open different virtual immersions, and so on. If you squint, you can see a glimpse of how a computer that you use by talking to it and manipulating things in space might one day work. 

Get fast shipping, movies & more with Amazon Prime

Start free trial

Enjoy this blog? Subscribe to andremax

3 Comments