Vision Pro teardown part 2: What’s the display resolution?

https://www.ifixit.com/News/90409/vision-pro-teardown-part-2-whats-the-display-resolution

by mzs • 2 years ago

171 232 2 years ago

kj997• 2 years ago

“friend of iFixit Karl Guttag has a blog post explaining why using the Vision Pro as a replacement for a good monitor is “ridiculous.”

…

This, says Karl, makes for a virtual Mac display with low enough PPD to see individual pixels—not quite the standard desktop display”

So rather than just trying it for themselves and realizing Karl is wrong, they quote this uncritically?

For what it’s worth, there’s nothing wrong with Karl’s math. He’s just not taking into account the fact that the Vision Pro isn’t mounted in a fixed position and that your brain integrates information over time.

In practice it certainly is as clear as my 4K display when I use it for mirroring.

It’s definitely not as good as a 5K studio display, but then Apple never makes that claim.

lukev3• 2 years ago

Eh... it's extremely good and more than I expected, but it's not quite 4k quality IMO. Not because I can see the pixels (I can't), but just because of very slight inconsistency across the field of vision, and motion blur when I move my head.

WhatIsDukkha2• 2 years ago

I've wondered how it looks with a very large spreadsheet filled with figures?

That big span of numbers should be fairly revealing of its display capabilities, with all the lines and columns.

How is the fringing and aliasing as you look at different angles?

Is a large pdf filled with math and tables comfortable to spend a few hours with taking notes?

steveklabnik4• 2 years ago

One of my friends who got a Vision Pro is disappointed in its use as a monitor, but only because he has multiple large monitors on his Mac, yet you can only set up one 4k display inside the Vision Pro, so it ends up being effectively much smaller and lower resolution for him.

Slightly different problem, but related. I'm sure it'll get there over time.

ballresin2• 2 years ago

This was going to be my complaint with the headset as well. I use 2x 49" super ultrawide displays, as well as 2x 27" 4k displays. Lots of real estate.

But then I started to think harder about what I was consuming display space with, and started to peel those things out into their Vision or iPad app equivalents. I'll not get into the obvious trade-offs of using iPad apps, but it largely suits the need.

Once I had Slack, Outlook, a Terminal, and TablePlus floating separately from the Mac display, I started to see a path toward potentially not caring if I even had the Mac display.

kj997• 2 years ago

Apple only ever promised it would mirror one 4K display. They announced that clearly when they introduced it.

It does that as well as the promised, but nothing more.

threeseed4• 2 years ago

Apple is rumoured to be testing multi-monitor support:

https://www.macrumors.com/2024/02/06/vision-pro-virtual-disp...

danpalmer• 2 years ago

Karl Guttag has a history of getting a lot of excruciatingly difficult details right, and missing the literal big picture right in front of him. I think he makes great technical contributions to the field (at least from my perspective as a layman), but every one of his analyses that I have read ends with something to the effect of "and this is why it is physically impossible for AR/VR to ever work", which is far beyond skepticism. It makes it hard to take the rest of his analysis seriously because it shows an irrational bias against the technology.

kj997• 2 years ago

Yeah. It’s weird. Even the analysis of passthrough is so off. If I look out of my window with AVP on, it’s obvious that I’m looking at a comparatively low res screen, but if I look at digital content, passthrough seems perfect because it’s not in my fovea, and given that the point of the device is to interact with digital content, passthrough is definitely at the level of being good enough.

threeseed4• 2 years ago

Karl Guttag: "As per last time, since I don’t have an AVP."

For a device like this where there is so much more than just raw hardware impacting user experience anyone who doesn't own one should have their opinion immediately discounted.

phonon4• 2 years ago

He owns one now.

https://kguttag.com/

LegitShady• 2 years ago

saying 4k and 5k displays while talking about PPD, without mentioning display size, is sort of useless. a 22" 4k display is going to look different than 85" 4k display.

jkubicek• 2 years ago

The Vision Pro comes in at a stunning 3,386 PPI.

Sweet jesus.

I know that PPD is the actual important measurement here, but still, a display with 3K PPI is outrageous.

michaelbuckbee• 2 years ago

Agree, but also the closer the screen gets to your eyeball the more it matters.

nomel5• 2 years ago

How so? The light coming into your eyes is taking the exact same physical path through your lens, and inside, as if the virtual display actually existed*. It's a physical requirement that the optics achieve this, for you to be able to see the image.

* Assuming the virtual display is virtually positioned at the focal plane of the lens, which is usually around 3m.

mciancia• 2 years ago

So, given their comparison with iphone/65 inch TV we would need 10k PPI to get similar PPD.

I wonder if that's possible with todays technology

Edit: Seems like there are some news from last few years about samsung having 10k PPI displays

Gare• 2 years ago

Seems like even 60 PPD in the central region would already be "good enough" for most.

fsh• 2 years ago

The DLP/LCD chips inside 4K video projectors have about the same pixel density.

huslage• 2 years ago

Reflective displays like that are much simpler to make than an OLED.

newzisforsukas• 2 years ago

https://news.stanford.edu/2020/10/22/future-vr-employ-new-ul...

modeless4• 2 years ago

they also offer bifocals, and progressives

Wow, really? I don't see how that makes any sense. Bifocals are for things that vary in distance, but the displays are at a constant distance from your eyes. Are people really so used to looking through bifocals that they need them in a headset just to feel comfortable even though they aren't necessary to focus?

I'm disappointed there is zero detail about the eye tracking cameras. Where they're located, what kind they are, how they see around/through the lenses, where the infrared illumination is. Has anyone else been investigating this?

pram• 2 years ago

I’ve found a lot of recent teardowns from them are light on interesting details. They seem to be making short YouTube videos now that don’t really cover everything.

The screen analysis was interesting but they rushed through the rest, and didn’t even look at any of the cameras???

daemonologist2• 2 years ago

I'm not aware of any information on camera details, but you can see the locations of the cameras and illuminators in this image from Apple: https://web.archive.org/web/20240206041049/https://www.apple...

The tracking system is inboard of the lenses, so it doesn't need to look through them. There are two cameras per eye, below and on the inside (near the nose), and infrared illuminators (not sure if they're just LEDs or something more complex) surrounding the lenses on all sides.

modeless4• 2 years ago

Thanks for that image, I can see the locations in the teardown now as well. But it couldn't be more clear that the cameras do need to look through the lenses. And furthermore, the IR LEDs also go through the lenses. They are all on the display side of the lens. It must have been tricky to avoid reflections from the illuminators going through those very fancy lenses.

outworlder2• 2 years ago

but the displays are at a constant distance from your eyes

Nitpick, but the focal point is about 6 feet away, which is the important thing for this discussion.

jrockway4• 2 years ago

I don't see how bifocals could possibly work in this; the display is always a constant distance away. What I bet they do is interpolate between your distance and close-up prescription to guess what prescription you need to view the constant-distance displays.

You can ask your optometrist for "computer glasses" if you want to see this in action.

heisenzombie• 2 years ago

That brings up a question for me: Are there people with wildly different near and far prescriptions whose vision would be improved by looking at the world through this device which reprojects it at a fixed optical distance such that a single prescription can be used?

banana_giraffe• 2 years ago

Not mentioned, but Apple also has help pages for what to order if you only wear/need reader eyeglasses. I'm equally confused why this would be necessary.

thomastjeffery• 2 years ago

It's a contextual communication choice: the intended reader is someone who is concerned about their bifocal prescription. I would assume that whoever makes the lenses will simply choose the correct focal measurement corresponding to the Vision Pro's fixed 6ft focal distance, and manufacture a lens for that specific focus. What needs to be communicated to the intended reader is that a bifocal prescription will be properly accommodated.

modeless4• 2 years ago

I'm pretty sure you're right. I'm not the intended reader. I'm wishing for a much more technical teardown while iFixit is going for a general pop sci audience.

xnx• 2 years ago

Whatever the strengths and weaknesses of the Apple Vision Pro as a product, this is a marvel of engineering. The APV seems to include the most advanced version of the output from several industries (semi conductors, display, materials, assembly, etc.)

mastax2• 2 years ago

Also textiles. The solo knit band is a technical marvel. A single 3D knit with embedded mechanical parts and hidden cables to adjust tension. It’s not my wheelhouse but I’d love to read a textiles expert’s take on it.

jsheard3• 2 years ago

It's a shame the solo knit band seems to be form-over-function, it looks cool and a ton of work clearly went into developing it, but nearly everyone seems to find the much simpler dual loop band to be more practical.

e: I stand corrected, the reviewers I've watched mostly preferred the dual loop but there's plenty of solo knit fans here. I guess that YMMV factor is why they include both.

maxglute2• 2 years ago

Can you link to some expert take? Sounds interesting.

namlem• 2 years ago

I think preference ultimately depends on the shape of your head. It's nice that they included both for that reason.

KRAKRISMOTT• 2 years ago

Same as juicero, many people bash the product but the engineering is world class.

cqqxo4zV46cp• 2 years ago

Like, maybe, not same as.

jrockway4• 2 years ago

I am not sure that I agree. Engineering is a world of tradeoffs. Juicero said "we are going to overbuild every component even if it means the company goes out of business". Welp.

Most impressive to me is when every unnecessary component has been removed, and every remaining component is as simple as possible. (The example that sticks in my mind is the humble $5 drip coffee pot. No moving parts except the switch to turn it on!)

nicolas_17• 2 years ago

Speaks really badly of the Vision Pro if there's anything that can be compared to Juicero.

LASR2• 2 years ago

Same as juicero, many people bash the product but the engineering is world class.

Absolutely not. Juicero was absolutely terrible engineering. Any mechanical engineering intern can whip up a big chonking hunk of a juicer or any other kind of product in SolidWorks.

Filleting every edge does not mean "engineering".

The vast majority of the work and difficulty in engineering is NOT getting something to work, but doing so while also working within constraints. There were no constraints in the Juicero. It was impressively manufactured. But terribly engineered. Which is why the entire thing went down despite an absurd level of fundraising.

babypuncher• 2 years ago

Even if the engineering was "world class", it was completely pointless seeing as people could get most of the juice out of the packs by just squeezing them with their bare hands.

Though I get the impression they put most of their real engineering effort into their obnoxious juice drm scam, and everything else was just overbuilt to make it feel premium.

antman123• 2 years ago

I prefer the solo band

mrcwinn3• 2 years ago

I agree. As much as tech communities tend to dump on achievements, we really should applaud this work. I can't wait to see breakdowns of versions 2, 3, 4. As you begin to master your own creation and manufacturing process, it will be neat to see how much simpler it becomes.

monkeynotes3• 2 years ago

I do wonder about the actual material constraints. My thought is, will it even be possible to make something so slight that it becomes an everyday piece of tech like the iPhone, or is it destined for entertainment and casual work?

Batteries are heavy, the optical aspect is bulky, heat dissipation is challenging. I haven't seen any tech on the near horizon that can answer those challenges. I wonder what Apple has in their R&D plans.

seydor2• 2 years ago

Didn't all VR devices before it do the same?

pornel• 2 years ago

Similar, but all the components in it are a notch better — higher resolution, better pass-through video, more precise hand tracking, and much faster CPU.

However, my impression from iFixit's teardown is that there's a ton of tiny boards, screws, and ribbon cables. I bet in v2 they'll figure out how to package it even more densly.

0x457• 2 years ago

Early VR of this wave had some remarkable displays, not in a sense that "wow they are so good", but "wow HMD displays were pure trash before". Quest 1 while it was a very silly device, it was a pretty good inside-out HMD. Q2 was a lot better.

All of those devices still suffered from the same issues: they can put only whatever Qualcomm can make for them. XR2 Gen 2 is slower than Snapdragon from the same generation. This is due to how "inefficient" Qualcomm SoCs compared to what Apple makes for themselves. (not M2 is meant for desktops and XR2 Gen2 is based on SoC used in mobile phones, so they were created with different power and thermal budgets in mind)

I find it interesting how apple went about the "number crunching" part of AVP by making a separate coprocessor just for that. Yeah, XR2 Gen2 has similar capabilities on SoC, but I want to see benchmarks.

MBCook• 2 years ago

Boy when they showed the shot of the Vision Pro Display density vs the iPhone 15 I literally laughed out loud.

That screen is simply incredible as a piece of technology.

Krasnol• 2 years ago

I'm just going to leave this material review here: https://www.youtube.com/watch?v=LmcWMjBpYBU

JKCalhoun2• 2 years ago

we estimate the Vision Pro to have an average of 34 PPD

I'm not sure that is interesting though. I would expect tradeoffs on PPD for my peripheral vision, higher density for the center of my vision.

I want PPD numbers like perhaps an average PPD for the center 30 degrees FOV.

hnuser123456• 2 years ago

Issue is that you can look around with your eyes at any part of the display, so the whole display needs to be able to show very high PPI/PPD, if you look at it. They can save on compute by detecting where your eyes are looking and reducing render resolution on the edges of your eye's FOV though.

https://en.wikipedia.org/wiki/Foveated_imaging

JKCalhoun2• 2 years ago

True, but once you glance to the side and something catches your attention it's not unreasonable to turn your head to get more detail/resolution.

akamaka2• 2 years ago

The potential of the display is what convinces me that Vision will be a breakthrough device.

If Apple has a roadmap to keep scaling up the resolution, while only incurring a fixed manufacturing cost of producing a small 2cm display, there will soon be no other device category that will be able to offer such image quality at so low a cost, not laptop screens or desktop monitors or TVs or movie theatres.

holman• 2 years ago

After using it since launch, the displays are still startling every time I step back to really look at them. They look so good. There's a few things that could be improved upon — let me look at my phone, or project a digital display over my phone screen — but the UI and virtual displays you see in Vision Pro feel just like reality. It's startling, and bodes well for future Vision Pro devices (and other VR handsets, too).

losvedir• 2 years ago

I agree they feel like reality, but I don't know how much of that is the display quality vs the absolutely astonishing way they stay fixed in space.

esskay• 2 years ago

Long term I dont think they _want_ Vision to be a breakthrough device. It's only supposed to be a stepping stone towards a true AR/MR device. The ultimate and likely not possible for decades goal is a contact lense based solution. The main goal is a normal looking pair of glasses capable of true AR/MR (not Google Glass or Nreal style attempts with hardware thats not really capable of a truly usable experience).

Long term wide adoption cant really happen with a VR device, the pitfalls (Weight, long term pain, hair being messed up, red marks on your face, poor battery life, etc) are too high of a barrier.

steveklabnik4• 2 years ago

Tim Cook is famously pretty negative on VR, but positive on AR. I would agree with you.

akamaka2• 2 years ago

I agree 100%, I think the potential is in a future Vision device which is smaller, lighter, and cheaper, while having an even better display.

TillE• 2 years ago

That's been their plan in the past, but I wonder if they'll reconsider. You can't do stuff like virtual environments in AR glasses. Apple may never care about games, but there are other VR features which users might really love.

I wouldn't be surprised if the headset continues even in the future where amazing perfect AR glasses are possible.

threeseed4• 2 years ago

The main goal is a normal looking pair of glasses capable of true AR/MR

You can't consume media e.g watch sports, movies etc or do any real work with those glasses which is a huge part of why Apple built Vision Pro in the first place.

And the battery technology doesn't exist for them to last anything longer than a few minutes in such a form factor unless you have it tethered.

gfodor• 2 years ago

You're confused. Vision Pro is a true AR/MR device. It uses AR passthrough. A lot of people for a long time have assumed there's some inherent reason people would never accept passthrough AR devices. A few people like myself have been consistent saying that's not the case, that passthrough AR is actually the right solution. Vision Pro is just the first version of what I expect to be a long line of passthrough AR devices from Apple.

paganel4• 2 years ago

I don't think the physics will ever be there in terms of battery/energy consumption of said future devices, meaning that I don't think Apple (or any other company) will ever be able to build a processing-device that won't have to depend on a big enough battery.

You could try to miniaturise it all (up to a point) and include it as part of the device itself, similar to Google Glass, but who would want such an energy "bucket" so close to one's eyes and face? That's just disaster (and a big lawsuit) waiting to happen.

mastax2• 2 years ago

If I understand your point, I don’t think it makes as much of a difference as it may seem. Display panels are typically made from large “mother glass” panels out of which smaller display panels are cut. So you can make many different sized panels from a single type of mother glass (all with the same PPI of course). And the TV/monitor industry has delivered remarkable improvements in quality and cost very quickly, so it’s not as if our current methods aren’t working.

It’s an interesting thought though, let’s see how it develops.

jsheard3• 2 years ago

The process used to make Micro-OLED displays is very different to the processes used for large OLEDs or LCDs, it has more in common with silicon manufacturing than conventional display manufacturing. AIUI that means the panel size will always be limited by the maximum reticle size of the silicon process, which is usually at most ~800mm².

That lines up with the 27.5x24mm size of the panels in the Vision Pro - that's 660mm², already close to the largest silicon die you can make.

megaman821• 2 years ago

I thought Sony makes the display (and cameras for that matter). It's odd that there are two critical pieces of tech that need a lot of R&D for the Vision Pro to keep progressing, and Apple has nothing to do with them. The screens need to get brighter and have higher resolution. The cameras need to work faster in low light and capture more accurate colors.

phonon4• 2 years ago

It's Sony's display. You can see a very similar version here. [1]

[1] https://www.sony.com/content/sony/en/en_us/SCA/company-news/...

jijijijij• 2 years ago

The potential of the display

... will never be manifested, because Apple won't let you have it. Use your imagination to tease you, but you won't ever get it. Lock'n'lol. iPad paperweight. Maybe next time, tiger.

dyno12345• 2 years ago

discussion of Part 1: https://news.ycombinator.com/item?id=39246664

robg2• 2 years ago

Can someone explain why they think the AVP version in 2034 will be light years ahead of v1? I see amazing engineering but not many places to shrink everything into a lightweight pair of glasses.

ceejayoz• 2 years ago

What other physical tech product doesn't meaningfully improve in a decade?

(Maybe the TI-83 calculator?)

nomel5• 2 years ago

You should ignore anyone claiming unquantifiables like "light years ahead".

In 10 years, my hopeful list is:

  * More custom/specialized silicon for XR. There's aren't many players here currently.
  * Retina/~60 PPD.
  * Micro LED (not OLED), making varifocal and HDR possible.
  * Varifocal lenses, which are just as important for comfort as high res.
  * Cameras optimized for passthrough, rather than borrowed from cellphones.
  * Lighter weight. They're almost tolerable, rather than acceptable, as is. This almost certainly involves offloading some of the compute.

But, a powerful untethered XR system and a lightweight set of untethered glasses will necessarily be very different, in capabilities. I imagine I'll end up with a tethered device, as a display replacement, in the next 10 years.

robg2• 2 years ago

Thanks for this list, doesn’t seem like the pair of glasses that people are envisioning, more like iPhone X, lots of incremental improvements but not a major difference from v1 to v10 or even v15.

grecy• 2 years ago

Can someone explain why they think the AVP version in 2034 will be light years ahead of v1?

Think back to when the iPhone first came out - could you have predicted how it would improve over the next 10, or 15 years?

Sure some things are obvious like faster networking, higher quality screen, better camera and just faster processor. But in all honesty I'd say some of the biggest improvements are apps that now define how we use it and what we use it for. It's obviously so much more than a phone.

I'm not sure anyone quite knows what the future of VR/AR truly is, but I'm betting it will apps which will change how we use it and what we use it for that will make it "light years ahead" of where we are now.

mschuster912• 2 years ago

Looking like a bulked-up first-gen iPhone, the case is milled out of a single chunk of aluminum, and the lid snaps into place with firm perimeter clips, leaving little to no seam for us to pry at. We needed a hammer and chisel to open it up! Adhesive also lines the lid, just to make sure you get the message: this pack is not designed to be opened.
Currently, iFixit engineering intern Chayton Ritter has Apple’s proprietary battery cable splayed out on a breadboard, and is trying to determine what sort of electronic handshakes are required to make the Vision Pro accept battery packs.

Apple really dropped the bar on this one. It's high time the EU forces companies to design their hardware in a way that allows the user (or a recycling company!) to replace the batteries inside without destroying it in the process, and for fucks sake what happened to the USB-C requirement for electronic devices?

lawik• 2 years ago

The brick/battery charges via USB-C from what I have seen. Casey Neistat ran that into a battery pack for extended use. I think that makes them plenty compliant.

I asusme the custom connector on the head unit is better fit for purpose (comfort, not getting snagged) than USB-C would be.

duskwuff2• 2 years ago

Or, to put it even more simply -- the battery pack is the replaceable battery. Which is fine; it's little different from the similarly sealed battery packs on electric power tools, for instance.

ribosometronome2• 2 years ago

The battery charges via USB-C. The battery connects to the headset with a weird side-snap connector that's rather flat. As designed, there's not anywhere obvious that a USB-C port could really fit on the headset and assuming they hope to make it smaller in future years, the problem only amplifies. I think other headsets have cables coming out of them when they need to connect to other things, rather than relying on a port. It's not particularly aesthetic.

threeseed4• 2 years ago

USB-C makes zero sense on the headset which is why it's on the replaceable battery.

You need a connector that can tightly lock and be capable of withstanding significant angular loads on the cable ends.

Twist locks have been proven to work well.

outworlder2• 2 years ago

It's high time the EU forces companies to design their hardware in a way that allows the user (or a recycling company!) to replace the batteries inside without destroying it in the process

Why? The battery pack is external. I have not ever seen a requirement that battery packs themselves need to be repairable. Not many devices come with external batteries these days, but do you really want to open them up to replace the cells? Where are you going to source reputable prismatic cells to use with such a device? They are also pretty delicate, you don't want most people handling them. If it used standard 18650 cells I could kind of see the point, but it would be much larger.

A recycling company can destructively open it up to recover the material.

A better 'mandate' would be to require it to accept third party power supplies.

for fucks sake what happened to the USB-C requirement for electronic devices

USB-C would suck for this application. It already does for headsets like Meta's Quest, where you can use a link cable to a PC. You need to secure any USB-C cables pretty well if you don't want to risk damaging the port.

mschuster912• 2 years ago

I have not ever seen a requirement that battery packs themselves need to be repairable.

Indeed it is, but it would make the life of recyclers way easier. Having to use a hammer and pin to open it risks puncturing the pouches.

Not many devices come with external batteries these days, but do you really want to open them up to replace the cells?

Uhhh... yes? That was the state of the art for decades before removable batteries were sacrificed for optics. You can even do watertight devices that way, Samsung's Active Tab 3 is watertight and supports swapping batteries.

Where are you going to source reputable prismatic cells to use with such a device?

The aftermarket supply chain for devices has gotten pretty well in serving that market, although I'd also love some standardization effort in that space.

kccqzy• 2 years ago

You already can replace the batteries. Apple will gladly sell you another battery: https://www.apple.com/shop/product/MW283LL/A/apple-vision-pr... But there's no use case for a consumer to disassemble the battery or to replace parts inside the battery.

fumar3• 2 years ago

Living with the Vision Pro is much easier than I expected. I thought it would feel bulky to manage with the external battery — it hasn't. I put the battery in my pocket and pop into visionOS without any hassle. I found the max time is about 1.5 hours before I feel slight eye and physical pressure fatigue from the headset.

nomel5• 2 years ago

and physical pressure fatigue from the headset.

Is this double loop, or cloth band?

fumar3• 2 years ago

Solo band, not the double loop.

joshstrange• 2 years ago

The displays on the AVP are truly amazing, the best I've ever seen. In fact I noticed zero issues relating to the displays themselves. Where this device stumbles is the passthrough (cameras) and the Mac Virtual Display (AirPlay). If/when those get better I'll buy another one but for now I'm planning on returning mine before the 14 days are up.

dwallin2• 2 years ago

One thing the article doesn't get quite right is you can't compare ppd of a virtual display directly to an IRL screen. The virtual pixels are not 1:1 to device pixels. Your perspective is constantly shifting slightly and the virtual pixels being sampled from is constantly changing. This is providing your eyes and brain a lot more data to sample from and means the perceived resolution can be higher than the actual resolution.

garbageman• 2 years ago

This is an interesting perspective. Genuine question: would it not be the other way - seeming lower than actual resolution when slightly moving? Why?

Could this be used to simulate a higher pixel density? Maybe doing something crazy vibrating the display in some kind of circular motion or something silly?

axoltl• 2 years ago

There are projectors that do this trick! The technology is called XPR: https://www.projectorjunkies.com/4k-dlp-projectors-is-it-rea...

layer8• 2 years ago

The issue is that the original Mac pixels get 3D-projected and thus interpolated onto the VP panels pixel grid. This invariably loses some sharpness and detail. Even just 2D-rotating a Mac screenshot by some non-orthogonal angle will make it more blurry.

monkeynotes3• 2 years ago

I am not convinced. Your eyes are literally magnifying the pixels compared to being at a distance from a modern 2k display. You may have dense PPD in AVP but being that close to them doesn't do it any favours.

I mean, if you can see a pixel you can see a pixel, there is no getting around that. It's like once you notice a blemish you can't not notice it anymore. From what I hear from users is you can absolutely see the pixels.

ninkendo• 2 years ago

My 2¢, not an expert, etc etc...

I think the key detail here is that each eye is getting its own set of pixels for looking at the same virtual content. This can lead to more detail, because the pixels aren't 100% redundant (they're both looking at the content from a slightly different angle.)

If I'm looking through the AVP at a "square inch" of my virtual Mac display, it may be that each eye is getting say, a 100x100 grid of pixels in this angular area. But since each eyepiece is giving me a slightly different projection of that same inch of space, the pixels themselves are going to be of subtly different values, essentially contributing more information to what I'm looking at. This is a lot different from a "real" display, when both of my eyes are staring at the same physical pixels. I think the idea is that my brain will be combining this information into a perceptual image in a way that will appear slightly more detailed than the equivalent-sized pixels in meatspace.

It's like once you notice a blemish you can't not notice it anymore

Interesting that you say this... because when you move your head towards a virtual object to inspect any pixels you saw, the image literally gets clearer (because you're getting "closer" to it) and you don't see the pixels any more. IME this goes a long way towards tricking my brain into thinking the pixels aren't real and therefore aren't there. (I've been using my AVP for real work all day today and I've been mostly very happy with it. The resolution is absolutely the least of my concerns, and it looks subjectively phenomenal to me.)

ShamelessC• 2 years ago

I feel like the article covered this.

crazygringo• 2 years ago

I've noticed this as well. I kind of think of it like when you look through a chain-link fence, there's a lot of visual data that gets blocked.

But if you're in a car driving past, the chain link fence gets blurred out and you can see everything again.

Something similar happens with the constant resampling happening in a VR headset. The hardware pixels are constantly shifting over the image pixels, as your head constantly moves just from breathing and whatnot. So not only does the signal-over-time counteract any blurriness from resampling in a single frame, but I find myself wondering if it enhances the resolution slightly.

(After all, you can buy 1080p projectors that "jiggle" the image by half a pixel diagonally, several times a frame, to effectively double the resolution and claim it's halfway to 4K. And what is the constant movement of pixel resampling if not a similar jiggling?)

I'm curious if there's an image processing term for this effect? The effective signal resolution from a constantly shifting pixel grid over time, and how to calculate it mathematically.

CarVac• 2 years ago

Multi-frame super-resolution.

What matters is illuminated pixel aperture size, in the end. The smaller the pixel aperture, the less smearing you get from overlapping.

dwallin2• 2 years ago

I put together a simple gif to roughly demonstrate the effect: https://giphy.com/gifs/JImONk8V5YK6iGBFkH/tile

While it still can't fully resolve shapes smaller than a pixel, it definitely helps with identifying forms.

CobrastanJorji• 2 years ago

This article does a fantastic job of explaining how good the screen is in clear, easy to understand language. Good job!

Is it a good display? It's incredibly good, far beyond the DPI of your phone or monitor. Does it look as good as your phone or your computer monitor when used? No, at the distance you'd look at them from, it looks much worse.

mrcwinn3• 2 years ago

The display may be 3,386 PPI, but wow my laptop screen has never looked sharper after taking the AVP off my head. Screen sharing from my Mac looks distractingly bad, which is a huge bummer. (AVP apps, by contrast, look fantastic.) No matter how I scale or where I place the space, it just never looks very good.

I'm really hoping this improves because it's nice to turn my 14" MacBook Pro into a theater screen of sorts alongside native apps. For the moment, though, AVP doesn't seem very useful apart from watching movies (which is amazing).

dvdhsu• 2 years ago

Agreed. I find my AVP actually quite bad for using my Mac. I use a MBP instead of a MBA because 120hz makes a big difference. The MBP screen within AVP is probably something like 30hz (likely because it's using Airplay, and it doesn't have enough bandwidth). And I can't change the resolution either! It's kind of like working on a TV — I don't see why it's any better than the 14" MBP screen.

Overall, the AVP is a disappointment for me. Most of the new UX patterns I find far worse than keyboard shortcuts. (For example, a window manager is _much_ easier to use than having to pinch to drag windows around. For example Vimium is much easier to use than looking at elements and pinching at things.) I don't consume much content (e.g. TV), but of the content that I do consume, it feels lower resolution to me. The demos they have (e.g. the Alicia Keys video) feels nothing like real-life to me. As a parallel to what you said — real life has never looked better (after taking off the AVP), haha.

bookofjoe• 2 years ago

I agree with much of what you wrote — except I found the Alicia Keys video staggeringly good. It felt like I was in the studio with her and her band.

Sephr• 2 years ago

You can combine paradigms for the best of both worlds (e.g. physical keyboard + your apps). I imagine that IDE experiences will only improve as time goes on.

sjm• 2 years ago

You can change the resolution though? If you go to Displays on your mac you can change the resolution of the AVP mirroring. I find it works much better at "1080p" than the default since it makes small text much more readable.

zerohp• 2 years ago

How is your wifi network? Have you tried changing the display resolution on the Mac while you're connected to the Vision Pro?

The virtual display doesn't look as good as the laptop screen, but it looks as good or better than the 1440p monitor I normally use at my desk.

mrcwinn3• 2 years ago

Thank you, I’ll fiddle with that. WiFi is good and my 14” M3 is using its native resolution, whatever the default is.

gpm4• 2 years ago

For comparison purposes can you share the size of your monitor and how far away you sit from it? Trying to decide if it's worth getting one of these (which would be a bit of a pain since I'm in Canada).

jolt42• 2 years ago

Disappointing to hear. I was hoping laptop work to be on par with a 27" 4k monitor, otherwise can't see I'd use it "enough".

mil2210• 2 years ago

Matching the 60° FOV of a 27" monitor at close distance, it's equivalent to 1080p or less. It gets even worse at FOVs equivalent to more reasonable distances.

matsemann• 2 years ago

Given that the screen inside the AVP is basically 4k, your screen would then have to fill your entire vision to have the same resolution. So almost like having your nose touching the screen.

mil2210• 2 years ago

It's not surprising. The AVP has a PPD of 34. A 14" MBP screen with a resolution of 3024x1964 already has a PPD of 66 at a close 14" viewing distance and it gets higher as you move further away [1]. Retina is considered 60+ at 20/20 vision. Of course a real display has no resampling or micro-OLED smearing to overcome either.

To achieve the same effective pixel count as a MBP, a virtual screen at the AVP's PPD would have to be resized to 90° horizontal FOV, which is just too big to be comfortable, equivalent to sitting 6" from a 14" MBP screen.

The technology just isn't there yet. We need at least another doubling of PPD. See you in 10 years.

[1] https://qasimk.io/screen-ppd/

jsheard3• 2 years ago

That's the great irony of Apple trying to make AR work as a monitor replacement, their existing users will be some of the hardest to please in terms of resolution since they've been spoiled by retina displays as standard on every Mac for the last decade.

Ericson23147• 2 years ago

Nice website. I got a 27" 5K UHD LG monitor at 90 PPD, so I'll never be able to use these things. So sharp I have to wear glasses not contacts.

ionwake• 2 years ago

i dont know why it took a week but i finally have you to confirm my suspicions based on the nebulous " its a bit out of focus i dont know" crap from every review I came across.

keep it up good man.

Dave_Rosenthal• 2 years ago

I think we need to withhold analysis based on the 34 PPD number. In the article you can see that the rendering projection is quite distorted, with the apparent effect of using more pixels per degree for the center part. They don't account for this when calculating the 34 number. Of course a "fisheye" lens would also be "distorted" and allocate an equal number of pixels to each angle, so it's hard to tell by just eyeballing things. I wouldn't be surprised if the actual PPD number in the center, where it matters, is higher.

fouc2• 2 years ago

I guess I was assuming that based on the AVP virtual desktop native resolution of 4096 x 2304 that it would be about 86PPD in practice

lukev3• 2 years ago

I for one don't notice any quality/resolution difference between native apps and a mirrored Mac screen.

I agree it feels "worse" than a physical screen - it is certainly less crisp. But I ran an experiment, and created a doc with font sizes from 1-15 pt, then viewed it on my physical 24" 4k monitor, and on a mirrored Mac display in the AVP set to the same (virtual) dimensions and distance away.

Interestingly, text became illegible (without leaning closer in either case) for me at around the 6pt mark for both of them. Objectively speaking, the size of font I can read is pretty equivalent, although I find myself preferring fonts about 2pts larger on the AVP than on my physical monitor.

I suspect this is because the physical display is able to benefit from precise pixel alignment and subpixel rendering, which the AVP can't. But the "average density" feels pretty equivalent... it somehow feels more an issue of optical sharpness than of "pixel size" (even I know the two are related.)

cubefox2• 2 years ago

I for one don't notice any quality/resolution difference between native apps and a mirrored Mac screen.

I wonder how so different opinions exist on things that should be pretty much the same for everyone.

jeppester2• 2 years ago

Screen sharing from my Mac looks distractingly bad, which is a huge bummer. (AVP apps, by contrast, look fantastic.) No matter how I scale or where I place the space, it just never looks very good.

I think a big part of the difference is that your feed is rotated and skewed raster graphic. That might work fine for photographs and movies. But UIs and text will quickly suffer from washed out edges when you apply transformations to otherwise "perfect" pixels.

The resolution of the pre-transformed image would have to be vastly higher to counteract the fuzzyness.

An even better solution would be if the UI could be rendered directly to the AVP rather than being a screen capture.

lancesells• 2 years ago

The thing I wonder is can't you fix the screen to be straight relative to your head at all times? Or is it always fixed to the environment?

madeofpalk• 2 years ago

There's so many layers of compression and scaling and transforming going on that it's never going to look Great.

macOS renders to 5120x2880 'display', then it's downscaled to 4k, which is then streamed to the Vision Pro with whatever compression or bitrate, where it's scaled to whatever size you display it as.

pciexpgpu• 2 years ago

How much of that is compression artifacts of streaming via wifi?

Hiko0• 2 years ago

Maybe it‘s linked to streaming resolution and bitrate.

seec• 2 years ago

Of every major issue with the AVP the fact that they not only avoided but made impossible to have a wired connections to anything even though they laughably made the battery pack external is really the most stupid thing ever. As a display replacement for more powerful hardware, we will never overcome the bandwidth problem (heck, it's already hard enough to do 5k at high fps with a wire) no matter how good the compression is.

It makes no sense that they built and advertise this "feature" when it could have been so much better with a thunderbolt/usb c input on the battery pack...

It doesn't spell much good for the future of the device, because it seems it will be largely locked down and limited like the iPad but even less enticing uses cases...

theogravity• 2 years ago

The Vision Pro has another small problem for spectacles wearers. Contrary to some reports, Apple says that corrective lenses are available for most conditions, including astigmatism (which we weren’t sure about in part one), and they also offer bifocals, and progressives. But if you have a prism value included in your prescription, you’re out of luck. Prism correction is used to correct for diplopia, or double vision. The easiest way to see if your vision prescription is supported is to use ZEISS’s online tool.

As someone that has prism in their glasses, that sucks. I do have custom lenses with prism for my Quest 2. Not sure why apple can't offer it as well, especially if the lens comes from ZEISS.

mac-mc3• 2 years ago

I think it's because of eye tracking issues. If you need a prism then you probably have other issues with how your eyes move around and they haven't developed for that yet. They are thinking about it although as shown with the accessibility options. I think they will figure it out eventually.

LoganDark• 2 years ago

I suspect I might need prism lenses, but there's already an eye tracking option that tells it to only use my left eye. Used that option during the in-store demo, and it worked pretty well. If the image in the right eye were adjusted to match my condition then that would've been even better, but my brain already mostly ignores the image from the right eye so it didn't matter that it wasn't.

seydor2• 2 years ago

Perceptible resoultion is what 's missing from all these VR goggles. The density of the screen doesn't matter much if the angular resolution isnt great. I was hoping Apple would be the first to rival our actual Retina

FredPret• 2 years ago

It will be so cool once these are miniaturized to those big square 70's glasses or smaller. Maybe contact lenses in some sci-fi future.

I imagine popping on a pair and suddenly you have a huge wrap-around workstation in front of you.

LASR2• 2 years ago

I've said this before, but until we have gimballed screens that can reposition a very high-density portion of the display to always be on the eye's fovea, we will not have a good productivity tool headset.

AVP v1 does the eye tracking quite well. I bet real $ that v2 will have motors internally to reposition the screens.

hartator• 2 years ago

I will take that bet. $5?

jxy• 2 years ago

So the PPD means we only need a factor of 2 denser pixels and we can throw all other 4K screens away. Wait for a factor of 3 higher PPD (~ 100?) and we may get rid of all of our single-person-use monitors. I guess I'll wait for Vision Pro 4 to get the "retina display."

gpm4• 2 years ago

Disappointing that more analysis isn't put into the lense effect on PPD.

SimulaVR (a competitor) has an extremely good blogpost about this [1] where they argue that the PPD in the central 30 degrees is far more important than the average PPD over the whole headset (because you very rarely turn your eyes more than 15 degrees off center, and your eyes have the highest resolution right in front of them), justifying the same sort of lenses that Apple appears to be using. They claim that this allows them to reallocate pixels for ~45% higher "foveal PPD" than would otherwise be expected.

[1] https://simulavr.com/blog/ppd-optics/

zaptrem• 2 years ago

Does Apple focus pixels in the center?

gpm4• 2 years ago

Yeah, there's a picture of the lense above a grid of lines demonstrating this in the article (which is the first time I've seen that confirmed).

gattr• 2 years ago

I'm waiting for a "monitor-replacement glasses" - show me a device which simulates an equivalent of my current 24"/4K/60 Hz monitor (same amount of pixels, same angular size in my FOV). Since I can touch type, I don't need AR/motion sensing, just the display.

bredren• 2 years ago

Vision Pro is not capable of delivering the level of low-strain fidelity needed to work long hours on anything related editing text. I think it is fair to say a 4k display or near equivalent is the minimum generally needed for this now, particularly for programming. [1]

Spatial Computing is not about arranging windows and using interfaces designed for flat surface devices.

It is about addressing problems using a three dimensional space. Instead of dashboard panels showing graphs in a window floating in the air, there are zones laid out in front of you showing data illustrated using 3D infographics

Apple has had to provide many examples and artful ways to work with windows because it is the only way people understand multi-tasking applications today.

"This device has to be good at window management or it isn't a useful device" is the reasonable thinking of a person who will continue to spend most of their time in flat-land, relating to others who also live in flat-land computing.

I think Apple knows very well that traditional application presentation is not actually where visionOS / spatial computing aspires to be. You can almost derive this from the types of games they put in the Apple Arcade of the app store. These do more than anything else to show where productivity will go.

Even for programming, spatial computing will shine by providing intuitive and interactive visualizations of software systems and infrastructure. Not because it renders an IDE in a window really well. That's what monitors are for!

In some way, how we solve problems has to evolve to make use of what spatial computing can provide.

Focusing on how the Vision Pro falls short of presenting the old interaction model served via hardware designed for that model is missing the true intent of the product and where it is all headed.

Apple is in a tough spot with this, because without handling some of the old way, the Vision Pro doesn't solve any "real world" problems. However, spatial computing solutions to old problems do not really exist yet. Most VP apps are just rebuilt iPad stuff in windows.

I think this is why the home screen apps are locked and Apple is so careful to subjugate apps not designed for VP. Also, why the App Store doesn't just show you a list of all apps published with VP entitlements.

I suspect they already have internal disagreement between product and marketing on what the low bar should be to qualify for an app getting to be listed as a Vision Pro-ready at all.

[1] I use an XDR Pro at my home office and a Dell U2723QE at my coworking space and the 4k is cringe-worthy in comparison. The VP comes nowhere near the experience of the Dell.

paganel4• 2 years ago

Has anyone done any studies on what the corresponding carpal tunnel-like diseases will affect the eyes of the people that will start using this device quite a lot.

Meaning that our eyes (and the associated "mechanics" behind them) for sure weren't meant to fiddle with computer windows on a "screen", the same as our hands weren't meant to type on a keyboard/click on a mouse for hours on end, that's why I'm curious about the potential side-effects.

fckgw• 2 years ago

How would it be any different than the normal eye strain from looking at a computer screen? It's essentially the same movements.

paganel4• 2 years ago

Never used this thing but as far as I understood you have to "move" your eyes in order to drag a window, or to click on something, or to do anything of the sorts, while when I'm looking on a screen I generally read stuff (which doesn't involve that "much" of an eye movement) or watch a video (which doesn't involve eye movement almost at all).

it• 2 years ago

Isn't it going to give people a terrible case of myopia to have a screen this close to their eyes?

gpm4• 2 years ago

The actual distance doesn't matter at all, only the apparent distance after taking into account the optics, your eyes can only see the latter. I don't know the number for the Vision Pro, but it's probably similar to most VR headsets at ~2 meters away, far better than most computer screens that people work at all day.

holoduke• 2 years ago

You need 2x 32k screens per eye to have no pixel visibility.

weltensturm• 2 years ago

which is why your phone’s screen needs a higher resolution than your TV, and why a movie at the theater looks fine, even though it may only be 2K

They should have measured the angular screen area for those scenarios, because this is entirely backwards.

atombender• 2 years ago

Anyone know how the Vision Pro handles focus distance? I use multifocal contact lenses, but even with the contacts, I can't focus on anything closer than 25-30 cm. I assume VR headsets like the AVP — where the screen sits much nearer the eyes — use lenses that move the focus distance, but does it work well even for people like me?

askonomm• 2 years ago

3000+ PPI, yet here I am, unable to get a decently priced 200+ PPI 27" monitor.

hotdogscout• 2 years ago

Does anyone know how the finger tapping tracking work? Is it just the Lidar sensors?

fouc2• 2 years ago

This, says Karl, makes for a virtual Mac display with low enough PPD to see individual pixels—not quite the standard desktop display experience.

I haven't seen any reviewers saying they can see pixels in AVP, in fact they usually say the opposite. So this part is surprising.

There's also a strange complaint about "not real 4K" but I'm far more interested in if it's a retina display or not.

i.e. if it visually manages to look like 227 PPI or better, then I'm happy.

In Apple world, retina 4K is 4096×2304, which is 9,437,184 pixels.

Vision Pro teardown part 2: What’s the display resolution?

Related Stories