Following Quest 2 almost three years to the day, Quest 3 is finally here. Meta continues its trend of building some of the best VR hardware out there, but it will be some time yet before the headset’s potential is fully revealed. Read on for our full Quest 3 review.

I wanted to start this review saying that Quest 3 feels like a real next-gen headset. And while that’s certainly true when it comes to hardware, it’ll be a little while yet before the software reaches a point that it becomes obvious to everyone. Although it might not feel like it right out of the gate, even with the added price (starting at $500 vs. Quest 2 at $300), I’m certain the benefits will feel worth it in the end.

Quest 3’s hardware is impressive, and a much larger improvement than we saw from Quest 1 to Quest 2. For the most part, you’re getting a better and cheaper Quest Pro, minus eye-tracking and face-tracking. And to put it clearly, even if Quest Pro and Quest 3 were the same price, I’d pick Quest 3.

Photo by Road to VR

Before we dive in, here’s a look at Quest 3’s specs for reference:

Resolution
2,064 × 2,208 (4.5MP) per-eye, LCD (2x)
Refresh Rate
90Hz, 120Hz (experimental)
Optics
Pancake non-Fresnel
Field-of-view (claimed) 110ºH × 96ºV
Optical Adjustments
Continuous IPD, stepped eye-relief (built in)
IPD Adjustment Range 53–75mm
Processor
Snapdragon XR2 Gen 2
RAM 8GB
Storage 128GB, 512GB
Connectors
USB-C, contact pads for optional dock charging
Weight 515g
Battery Life 1.5-3 hours
Headset Tracking
Inside-out (no external beacons)
Controller Tracking
Headset-tracked (headset line-of-sight needed)
Expression Tracking none
Eye Tracking none
On-board cameras 6x external (18ppd RGB sensors 2x)
Input
Touch Plus (AA battery 1x), hand-tracking, voice
Audio
In-headstrap speakers, 3.5mm aux output
Microphone Yes
Pass-through view Yes (color)
MSRP
$500 (128GB), $650 (512GB)

Hardware

Even if the software isn’t fully tapping the headset’s potential yet, Meta has packed a lot of value into the Quest 3 hardware.

Lenses

Photo by Road to VR

First, and perhaps most importantly, the lenses on Quest 3 are a generational improvement over Quest 2 and other headsets of the Fresnel-era. They aren’t just more compact and sharper, they also offer a noticeably wider field-of-view and have an unmatched sweet spot that extends nearly across the entire lens. That means even when you aren’t looking directly through the center of the lens, the world is still sharp. While Quest 3’s field-of-view is also objectively larger than Quest 2, the expanded sweet spot helps amplify that improvement because you can look around the scene more naturally with your eyes and less with your head.

Glare is another place that headsets often struggle, and there we also see a huge improvement with the Quest 3 lenses. Gone are the painfully obvious god-rays that you could even see in the headset’s main menu. Now only subtle glare is visible even in scenes with extreme contrast.

Resolution and Clarity

Quest 3 doesn’t have massively higher than Quest 2, but the combination of about 30% more pixels—3.5MP per-eye (1,832 × 1,920) vs. 4.5MP per-eye (2,064 × 2,208)—a much larger sweet spot, and a huge reduction in glare makes for a headset with significantly improved clarity. Other display vitals like persistence blur, chromatic aberration, pupil swim, mura, and ghosting are all top-of-class as well. And despite the increased sharpness of the lenses, there’s still functionally no screen-door effect.

Here’s a look at the resolving power of Quest 3 compared to some other headsets:

Headset Snellen Acuity Test
Quest 3 20/40
Quest Pro 20/40
Quest 2 20/50
Bigscreen Beyond 20/30
Valve Index 20/50

While Quest 3 and Quest Pro score the same here in terms of resolving power, the Snellen test lacks precision; I can say for sure the Quest 3 looks a bit sharper than Quest Pro, but not enough to get it into the next Snellen tier.

While the optics of Quest 3 are also more compact than most, the form-factor isn’t radically different than Quest 2. The slightly more central center-of-gravity makes the headset feel a little less noticeable during fast head rotations, but on the whole the visual improvements are much more significant than ergonomic.

Ergonomics

Photo by Road to VR

Ergonomics feels like one of just a few places where Quest 3 doesn’t see many meaningful improvements. Even though it’s a little more compact, it weighs about the same as Quest 2, and its included soft strap is just as awful. So my recommendation remains: get an aftermarket strap for Quest 3 on day one (and with a battery if you know you’re going to use the headset often). Meta’s official Elite Strap and Elite Strap with Battery are an easy choice but you can find options of equal comfort that are more affordable from third-parties. FYI: the Elite Straps are not forward or backward compatible between Quest 2 and 3.

While the form-factor of the headset haven’t really improved, it’s ability to adapt to each user certainly has. Quest 3 is the most adaptable Meta headset to date, offering both continuous IPD (distance between the eyes) and notched eye-relief (distance from eye to lens) adjustments. This means that more people can dial in a good fit for the headset, giving them the best visual comfort and quality.

I was about to write “to my surprise…”—but actually this doesn’t surprise me at this point given Meta’s MO—the setup of Quest 3 either didn’t walk me through adjusting either of these settings or did so in such a nonchalant way that I didn’t even notice. Most new users will not only not know what IPD or eye-relief really does for them, but also struggle to pick their own best setting. There should definitely be clear guidance and helpful calibration.

The dial on the bottom of Quest 3 makes it easy to adjust the IPD, but the eye-relief mechanism is rather clunky. You have to push both buttons on the inside of the facepad at the same time and kind of also pull it out or push it forward. It works but I found it to be incredibly iffy.

Field-of-View

In any case, I’m happy to report that eye-relief on Quest 3 is more than just a buffer for glasses. Moving to the closest setting gave me a notably wider field-of-view than Quest 2. Here’s a look at the Quest 3 FoV:

Personal Measurements – 64mm IPD
(no glasses, measured with TestHMD 1.2)

Absolute min eye-relief (facepad removed) Min designed eye-relief Comfortable eye-relief Max eye-relief
HFOV 106° 104° 100° 86°
VFOV 93° 93° 89° 79°

And here’s how it stacks up to some other headsets:

Personal Measurements – 64mm IPD
(minimum-designed eye-relief, no glasses, measured with TestHMD 1.2)

Quest 3 Quest Pro Quest 2 Bigscreen Beyond Valve Index
HFOV 104° 94° 90° 98° 106°
VFOV 93° 87° 92° 90° 106°

Audio

Another meaningful improvement for Quest 3 is improved built-in audio. While on Quest 2 I always felt like I needed to have the headset at full volume (and even then the audio quality felt like a compromise), Quest 3 gets both a volume and quality boost. Now I don’t feel like every app needs to be at 100% volume. And while I’d still love better quality and spatialization from the built-in audio, Quest 3’s audio finally feels sufficient rather than an unfortunate compromise.

Controllers

Photo by Road to VR

Quest 3’s new Touch Plus controllers so far feel like they work just as well as Quest 2 controllers, but with better haptics and an improved form-factor thanks to the removal of the ring. Quest 3 is also much faster to switch between hand-tracking and controller input when you set the controllers down or pick them up.

Processor

The last major change is the new Snapdragon XR2 Gen 2 chip that powers Quest 3. While ‘XR2 Gen 1’ vs. ‘XR2 Gen 2’ might not sound like a big change, the difference is significant. The new chip has 2.6x the graphical horsepower of the prior version, according to Meta. That’s a leap-and-a-half compared to the kind of chip-to-chip updates usually seen in smartphones. The CPU boost is more in line with what we’d typically expect; Meta says it’s 33% more powerful than Quest 2 at launch, alongside 30% more RAM.

Quest 3 is still essentially a smartphone in a headset in terms of computing power, so don’t expect it to match the best of what you see on PSVR 2 or PC VR, but there’s a ton of extra headroom for developers to work with.

Continue Reading on Page 2: Softwhere? »

1
2
Newsletter graphic

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. More information.


Ben is the world's most senior professional analyst solely dedicated to the XR industry, having founded Road to VR in 2011—a year before the Oculus Kickstarter sparked a resurgence that led to the modern XR landscape. He has authored more than 3,000 articles chronicling the evolution of the XR industry over more than a decade. With that unique perspective, Ben has been consistently recognized as one of the most influential voices in XR, giving keynotes and joining panel and podcast discussions at key industry events. He is a self-described "journalist and analyst, not evangelist."
  • ViRGiN

    NO DISPLAY PORT = INSTANT BUY !!

  • Nevets

    One of the more incisively-written reviews I’ve seen so far.

  • Lacking occlusion in a starkly promoted mixed reality headset is kind of telling the whole meta story from a software point of view. They really aren’t focused on making user frinedly software, imo.
    Either way, I‘ll be enjoying my new headset tomorrow fwiw.

    • That is a software failure as opposed to a hardware one. So, hopefully that will be sorted for the base OS engine for handling MR objects and occlusion.
      I imagine that individual apps that function in their own MR implementation could probably fix that issue. Or at least for recognized/tracked objects. Like for arms/hands it should definitely work. For other people or pets, maaaaaybe, maybe not.

      It is disappointing that they didn’t have that sorted for launch, but that is hopefully able to be sorted in the future.

      • Christian Schildwaechter

        That is a software failure as opposed to a hardware one.

        Unfortunately that pretty much sums up most of Meta’s ongoing XR issues.

      • That is indeed a software problem, as I wrote. The depth API will be rolled out by the end of the year. My point is, that a promoted mixed reality headset without proper occlusion is missing its mark.
        We all know that the Q3 will be a different beast by the end of next year and in the coming months ….

    • kraeuterbutter

      well: they have the best AR out there in the moment, with most features..

      and that at a price-point starting at only 500
      HTC, HP, Microsoft, Valve, Pimax, Bytedance, Samsung …
      ????
      so its hard to tell that they arent focused when there is NO competition at all in the moment from all the other headset-providers

      • Fair enough. Although they should promote it as a VR headset first, imo and not promoting it as a mixed reality headset when in fact its more of an add-on.
        Maybe they do feel Apple breathing down their neck…

      • Christian Schildwaechter

        TL;DR: the Quest 3 doesn’t offer any AR, MR is still just VR with passthrough and clever tracking tricks; it will never support anything but light AR, if at all, the actual AR in the new Ray Ban glasses relies on powerful remote servers.

        Sorry, but you triggered my main pet peeve with MR: MR is not AR. The borders between them are fuzzy, and superficially MR does something similar, but is kind of faked. The spatial hooks allowing to pin objects in MR on Quest 2 don’t really recognize the wall your virtual picture is hanging on. During setup the walls were either marked manually or partially recognized to construct a 3D map of your space. But during use the tracking mostly uses easy to detect edges to determine how much you have moved/rotated. Based on that it shifts your position in its 3D map, and with it the virtual picture that really is just hanging somewhere in space. The Quest 3 so far improves the 3D map creation and (maybe now) allows for dynamic changes, but still only pins objects into the air inside the room.

        The Ray Ban glasses that Meta also showed during Connect actually feature AR, they recognize the actual objects and can provide extra information to them, replace or enhance them or do lots of other things. Which requires a lot more processing power, so it is most certainly not done by the rather weak X1 SoC powering the glasses. Instead somewhere remote at a data center via the permanent internet connection this requires, with the resulting privacy issues. Probably less of an issue with sun glasses that you’d usually wear outside, compared to a Quest 3 you mostly use in your private environment.

        I have always been annoyed by the term MR, as it implies something like AR, and Meta marketing most certainly wants people to assume exactly that, but actually doesn’t support AR’s main features. MR is still VR with passthrough and clever tracking, and so far Meta uses the new depth sensor in Quest 3 mostly to improve hand/controller and room tracking. They haven’t even implemented a guardian that would allow extending the play space above low objects like couches, they are just “evaluating the option”. And with no access for developers to the cameras (with reasonable privacy arguments), it isn’t even possible to implement some light AR or fiducial markers.

        So: Meta doesn’t have the best AR out there in the moment, with the most features. They don’t have any AR out there in the Quest, and the only AR they offer is on Ray Ban that works pretty much the same way as Google Lens, sending everything to a remote machine for interpretation. Others actually allow(ed) access to the camers for AR and thereby object recognition and interaction, so Pico 4 Enterprice, HTC Vive Pro (with old SDK), HP Reverb G2 Omnicept Edition and everything from Varjo offer better AR than Quest 3, although all of them come at a higher price.

        Every iPhone is leagues ahead of the Quest 3 regarding AR features. The Quest 3 has the hardware to do light AR tasks, so hopefully more is coming, though the lack of occlusion culling at launch is not a good sign. But I’d expect Apple to beat Meta fiercely in AR, with the AVP focused that way from the start, using the much more powerful M2, and also adding the closely related R1 just for the processing needs of all the sensor/camera data that is required for more detailed object recognition.

    • Lucidfeuer

      They’re just bad a conceiving software because they have no vision. The results being clunky assemblages of softwares that never culminate in a stable, smooth and optimised usage.

      • Thats where Apple will destroy them uI/UX wise. Meta should focus on gaming imo.

      • ViRGiN

        You sometimes forget you’re a nobody of the worst kind, a parasitic shill, so your remark is funny.

  • Xron

    I wonder how far behind XR2 Gen 2 is from gtx 1060 n terms of power o.O
    Wanted to preorder quest 3, but after reading reviews I guess it still lacks apps to fully show its potential.

    • ViRGiN

      RTX4090 is unable to show potential of pcvr, cause it’s not about rendering power but quality software

      • kakek

        It’s about both.
        PCVR lack good games. There’s been like 4 or 5 games that truly showcased what was possible on PCVR that wasn’t possible on mobile VR.

        And while I have you attention, is there still enough players on contractors on Quest ? Specially on EU servers. I’m wondering if I should suggest it to a friend looking for an alternative to pavlov.

        • ViRGiN

          Contractors is booming as always, not slowing down. Still top seller daily for Quest, and even on Steam it’s rather improving (its crossplay anyway).
          Last night there were 25 pages of servers when I got on. Not sure how many EU servers. I just join whatever is on the list which will be mostly NA as that’s majority of VR market, but I don’t have any issues with connection. I usually hear it NA->EU complaints, but not the other way around.

          Skip pavlov altogether, it’s a giant mess, not a shooter game.

    • T-ReX

      Should be around 1050ti

  • Dragon Marble

    I would say give the soft strap a chance and don’t buy accessories on day one. You have nothing to lose. My personal experience with Quest 2 is a cautionary tale. I bough elite strap on day one, later tried Bobo, but eventually went back to the default strap.

    People hate the soft strap but seem unable to explain why. Does it slip off? Pull the back down more. Does it then start to push against your ears? Tilt the front up a little. I know people have different head shapes. However, precisely because of that, aren’t you more likely to find comfort with a soft, flexible strap than with a hard, rigid one?

    • Bob

      Those with smaller heads tend to find the soft strap more uncomfortable. With larger heads, there is generally more “mass” to hold the device in place better so this type of strap works much better for them. If you have a rather big head, than the soft strap will suffice. Likewise, if your head is on the smaller side, get the Elite Strap accessory to go along with the Quest 3.

      • Dragon Marble

        I have a normal head as far as I can tell. The soft strap does require some fiddling, which may be time someone reviewing this in 4 days does not have.

    • Christian Schildwaechter

      I still assume that whether somebody loves or hates the Quest 2 soft strap mostly depends on how well the facepad matches their bone structure. A good match will distribute the pressure very evenly, while a bad match will lead to high and very uncomfortable pressure at just a few points, driving those users to 3rd party head straps that remove pressure from the face through stiffer frames and better balance.

      Ben’s review doesn’t mention the facepad, but I’ve seen at least one review (don’t remember by whom) reporting that the facepad is much improved. So even if the strap is the same as on a Quest 2, there is still a chance that it might provide a better experience thanks to a better matching facepad. This again depends on the individual facial structure and won’t work for everybody, as evidenced by Ben’s “its included soft strap is just as awful” conclusion. Basically your chances of winning the facepad anatomy lottery may have somewhat improved, so first trying the combination of new facepad and soft strap instead of getting a replacement immediately is sound advice.

    • trmn8r

      Bobo is best. They are doing a conversion kit too woot! Default strap was terrible for me.

    • I’ve done the same. When Q2 was released, the elite strap with battery just made sense, and fortunately mine never cracked as some did.

  • CrusaderCaracal

    Can’t wait to buy one

  • Ivan Mike

    Justo wait to Pico 5 pro

    • ViRGiN

      Uncompressed Display Port SteamVR Gorilla Tag Player alert!!

      • Garhert

        He didn’t say Pico 5 Pro Max, the Pico 5 Pro doesn’t have a DP. I mean, if there would be good SW, then the eye-tracking of the Pro could indeed be interesting. But I don’t think there will be any cool SW for this feature in a year.

        • ViRGiN

          Well, it’s safe to assume he meant the Max or whatever. There is no reason to buy Pico at all. All the hype I’ve seen is for that rumoured DP port.
          Almost 2024 and we still have PCVR elitism.

          • Garhert

            Well, Pico 4 visuals and comfort (without accessories) are better than Quest 2. And if you found some games optimized for Pico 4 that you like, then I wouldn’t say there’s no reason at all. But in general I agree, after almost 1 full year with the Pico 4, I’m happy to switch back to the Oculus Store.

          • ViRGiN

            Pico 4 came out 2 years after Quest 2. Of course it should be better. Is it truly better? I don’t think so if you aren’t using it mainly for pcvr. The same games aren’t running better on pico, and has missing features like lack of dlc or mod support. And there isn’t anything to look forward to on it.
            Pico is basically a headset for Meta haters, and those who wanted pancake pcvr before Quest 3 materialized. Now it’s obsolete.

          • kraeuterbutter

            i have (still) all Meta Headsets (go, Quest1, two Quest2) – so iam not a meta-Hater

            my Quest3 will come today
            still i love(d) my Pico4 and the last year it replaced my Quest2 completly
            its much more comfortable (only 290g in front instead of 450g of the Quest2), very good balance (you dont need to “press it” to your face for that reason
            i can use it very good without facegasket as well (my prefered mode for the moment)
            better E2EC than the Quest2
            much better FOV (Index-like without gasket)
            i like the UI better than on quests.. Quests feel meanwhile overloaded..
            easier APK support,
            you can extend the storage bei 1TB with no problem, not (easily) possible with the quests
            better buttonlayout (in m y opinion)
            for my hands: the pico4 controllers feel better
            for my usage without gasket: of course no fogging lenses
            the binocular overlap is better on the pico4 than on any other headset on the market right now
            i liked the pico4 and it increased my VR usage-times over the last years..
            so it was a very good (for me better) competitor to the Quest2
            nevertheless: i hope the Quest3 will easily replace it for me…
            lot of improved features over the Quest2 and also over the Pico4
            the lenses should be by fare supperior (some qualit issues with the Pico4 lenses, warping, glare)
            the only two points i think the pico4 will still be better than Quest3:
            binocular overlap: already reportet, that it is small on the Quest3 and unfortunatly you can see it more than on for exmaple INdex
            so here the pico4 will win easily
            and comfort: the balance, the battery on the back: the pico4 is like a hat: i can without strapping it to the face just “lay” it on my head..
            even with money, elitestrap etc. this will not be so good with the Quest3..
            with elitestrap i think it will be around 800g
            but everything else – i hope – will be better compared to Pico4
            so: the quest3 should win

          • Gonzax

            I would use the Pico as a PCVR headset but I don’t see the point in buying one to be used as standalone. I prefer a Quest 3 for that because of the exclusive games: AC, Asgard’s, etc.

    • Christian Schildwaechter

      This requires a little bit of context. There are rumors about Pico releasing the Pico 5 in three different versions (Standard/Pro/Max), all using the same XR2 Gen 2 SoC and a depth sensor as the Quest 3, and very likely pencake lens, as the Pico 4 already featured those. Standard and Pro with 2560*2560 per eye, Max with 3840*3840 per eye and DisplayPort input, as that resolution will only make sense when driven by a rather powerful PC. The Pro and Max would also include eye and face tracking.

      The rumored prices are USD ~490 for Standard, USD ~600 for Pro and USD ~1110 for Max when converted from Yuan. Which puts the Pro at less than the cost of a Quest 3 plus Elite strap with battery, (assuming Pico sticks to a Pico 4 style strap with the battery as counter balance) with eye and face tracking and about 43% more pixels. That would make it a pretty good deal, esp. for those that were/are disappointed by the lack of eye tracking on the Quest 3.

      Meta’s cooperation with Qualcomm includes exclusive access to the XR2 Gen 2 for six months, so a Pico 5 based on it couldn’t ship before April 2024, which is the rumored release date for the standard version. The Pro and Max are supposed to ship in October and December 2024. By then Meta may have released a USD 200 lower end Quest to replace the Quest 2, and the AVP will be in the wild, either showing that eye tracking isn’t all that important or pretty much essential for a good XR UI/UX.

      Combined with the rumors being in no way reliable, waiting on a Pico 5 Pro delivering all the goodies at a lower price than the Quest 3 is a somewhat risky bet, as Pico may change its strategy. It is pretty safe to assume that there will be a Pico 5, and Bytedance makes enough money from TikTok to allow them to underbid Meta with the higher end model, while ignoring the entry level price, as they already did with the Pico 4. And there is a chance that Pico will count on the political climate in the US being less hostile next year and start their delayed attempt to enter the US market. And then try to pound on the Meta Quest 3 by providing an elegant UI driven by eye and hand tracking that will basically be a copy of the AVP UI, something Meta cannot replicate with their current generation of HMDs.

      • Garhert

        Don’t forget the controller that could provide an even better tracking than the ones shipped with Quest 3 and the 16MP RGB cameras that would provide a higher resolution in passthrough mode. So on paper it sounds really nice. But only future can tell and I don’t wanna wait another 6 month. I have the Pico 4, will hopefully receive the Quest 3 early next week and might end up buying the Pico 5 Pro in the future. If you wanna have the best standalone VR experience right now (and probably for quite some time) then there is no way around the Quest 3.

        • kraeuterbutter

          hmm.. i would not expect toooo much from the passthrough of the Pico5
          1.) 16MP RGB.. the Pico4 has also a 16MP RGB camera.. still the resolution, how you can read text, … is definitly less and worse than on the Quest3 with “only” 4MP cameras
          2.) the two images must also be processed by the xr2gen2 (which has to do much other things as well)
          latency is very important to “feel in the moment” for a passthrough
          so – with the same processor like the Quest3 – processing two 16MP images could be challenging when you still want to have low latency

          • Garhert

            Good point, I was just counting pixels here and thought that it’s a shame that the Quest 3 uses 4MP cameras that cannot even provide enough pixels for the full resolution of the displays. Can we assume that what Meta does with passthrough is pretty much the limit of the xr2gen2 since the GPU isn’t really underclocked, or at least not by much? Could additional CPU power help?

            But this is why I said the Pico 5 specs look good on paper. The end result counts.

          • Christian Schildwaechter

            It’s very likely that tracking and passthrough on Quest 3 are handled mostly by dedicated subprocessors instead of CPU and GPU, so the limit may be there. Meta already did all the tracking on the DSP of the Quest 1/2, which initially limited its quality compared to e.g. Ultraleap handtracking also utilizing the CPU. But over time the tracking quality on both Quest 1 and 2 improved significantly, so I’d expect Quest 3 passthrough to also get better with future software updates.

            Image processing for passthrough could be done on a PC, but you’d add a lot of latency, as due to the limited bandwidth, the sensor/camera image would first have to be encoded as a video stream, transferred via a limited high latency connection to the PC, decoded, analyzed and the results again sent back via the same limited connection. Apple said the AVP is the first that manages the camera-to-display passthrough in just 12ms thanks to the dedicated R1 sensor data processor, so @90Hz passthrough introduces only a single frame of delay compared to not wearing the HMD, significantly improving comfort. The Varjo XR-3 offers similar low numbers, but there the cameras connect directly to the PC, and the PC also renders the images and shows it via DisplayPort, which removes most of the latency issues that would occur with a PC doing the processing for a Quest 3.

          • Christian Schildwaechter

            The Pico 5 would also include a depth sensor, allowing for geometry correction like on the Quest 3, there leading to much improved passthrough compared to the Quest Pro, which tries to correct the color data based on the stereoscopic image from the b/w tracking camera at high computational costs. The Pico 4 doesn’t even try to apply geometry correction, leading to a different kind of artifacts, and depending on who you ask, either the “no correction” on Pico 4 or the “approximated correction” on Quest Pro is better or worse. The resolution of the color sensor is less of an indicator for usable passthrough than the extra depth sensor, and much of the pixel processing will be done by either the DSP or maybe even a dedicated Spectra image processing unit the XR2 Gen 2 may have inherited from the SD8 Gen 2.

      • Gonzax

        I don’t see a headset like that costing only 600 or even 1110 for the Pro, that’s a lot of great specs and features.
        But if it happens I might be getting the Max as my second headset for PCVR, unless Valve finally comes up with an Index 2 or whatever they’re calling it.

  • Christian Schildwaechter

    TL;DR: massively underclocking the CPU hints that the Quest 3 is not supposed to ever replace the Quest 2, but that both will run the same software at different graphics quality levels until their whole “generation” is replaced by a Quest 4.

    The CPU boost is more in line with what we’d typically expect; Meta says it’s 33% more powerful than Quest 2 at launch […]

    That is not at all what I expected. We know that the Quest 2 CPU was underclocked significantly to allow for giving more power to the GPU without the whole SoC getting too hot and then having to throttle. And one of the changes in the Quest Pro was moving the RAM that was placed on top of the SoC with the XR 2 Gen 1 to the side, allowing better cooling.

    Comparing benchmarks for the SD865 (XR2 Gen 1) und SD8 Gen 2 (XR2 Gen 2) shows about 50% more CPU performance for single core and about 60% more in multi core. So if the Quest 3 CPU performs only about 33% better than the already underclocked one in the Quest 2, when it should be about twice as fast just from the improvement of the SoC, this means that the Quest 3 CPU is even more aggressively underclocked, to allow for the 2.6x speedup of the GPU.

    And that is relative to “Quest 2 at launch”. But Meta released the Quest software update v55 in June, saying that it provided up to a 26% CPU performance improvement for Quest 2/Pro, and up to 19%/11% for GPU. So in the best case 26% scenario, the Quest 3 CPU performance is only 6% higher than on the Quest 2 with the latest software. As little as 6% compared to the 60% the SoC according to benchmarks could actually offer is not what I expected.

    It made sense to underclock the Quest 2 CPU, because mobile SoC are optimized for phones with much lower resolutions. The Adreno GPU in Quest 3 has to drive 4.4x the pixels as a phone with 1080p, so underclocking the CPU again makes sense. But underclocking a lot more than on Quest 2 hints that Meta expects both to run basically the same games, only the graphics will be a lot better on Quest 3. There is no extra CPU budget to allow for better physics or larger numbers of NPCs in a more advanced type of game, esp. since preparing more complex graphics for the GPU also takes some extra CPU performance.

    That (again) hints that Meta doesn’t see the Quest 3 as a successor to the Quest 2, instead as a more powerful sibling. Like PS4 Pro to PS4, and both Quest 2 and 3 together form one generation. Significant improvements or technical changes like eye tracking/ETFR may therefore have been pushed to a Quest 4 that would start a true next generation and allow for a new type of games, with a similar six year cycle and a more performant enthusiast device mid cycle.

    This also matches rumors about a USD 200 headset for 2024, most likely replacing the Quest 2 as a cheaper to manufacture device. It would probably keep most of the specs like CPU/GPU performance and Fresnel lenses, at most updating the cameras to allow slightly better passthrough in color, which should be rather cheap to do and help Meta’s MR ambitions. Such a Quest 2 lite would probably be sold up to the release of a Quest 4, meaning that, unlike with Quest 1, Meta would not at one point drop the base model for lack of performance. Instead developers would continue to create games that run on both Quest 2/lite and Quest 3, just with significantly improved graphics on the “enthusiast” Quest 3.

    As a strategy, that makes quite a lot of sense. It’s not the Nintendo way, where a console completely replaces its predecessor, nor the Apple way, where the previous top model is sold in parallel to its successor as the now cheaper version. It’s the Sony way of long generation cycles with a hardware refresh and enthusiast option in the middle. It leads to less fracturing of the software market, quite important for developers, and keeps the already large Quest 2 user base in play, as it is quite unlikely that the Quest 3 at a 67% higher price will ever get close to Quest 2 sales numbers, even less if in 2024 Meta releases a Quest lite that pushes the Quest 3 premium to 150%. And we have been told this before, only nobody assumed that Meta’s “long” really meant “console generation long”. Andrew Bosworth in 2021:

    For those who are curious, Quest 2 is going to be in the market for a while – for a long while, and it’s gonna be, you know, I think the best bet for the most accessible way to get into VR and have a great experience.

    • Isaac

      The XR2 Gen 2 is not the same as the 8 gen 2. The former has 6 CPU cores instead of 8, all of them being medium performance cores which are not underclocked. There are no cortex X2 cores. The 33% improvement is over the upgraded Quest 2.
      The Quest 3 is still bottlenecked by the GPU in graphically intensive games, so CPU increase will not benefit them. This was further exacerbated by AppSW which cuts CPU requirements by nearly half and increases the GPU bottleneck.

      • Christian Schildwaechter

        Wasn’t aware of the changed core configuration, so thanks for mentioning it, though it hasn’t really cleared up yet.

        The XR 2 Gen 2 does indeed seem to have 6 CPU cores, but in a 2/4 setup, so (according to Digitrends) 2 of these are performance and 4 efficiency cores. The SD8 Gen 2 has 8 Cortex cores in a wild 1/2/2/3 setup, with 1 X3@3.2GHz, 2 A715@2.8GHz, 2A710@2.8Gz and 3 A510@2Ghz. Which is only one (high) performance core, but clocked much higher at up to 3.2GHz for burst activities, compared to two on the XR2 Gen 2, which according to a slide from Meta only clock up to 2.05Ghz for CPU heavy and 1.6Ghz for GPU heavy tasks.

        Meaning I still don’t really have a clue, and so far it seems rather hard to compare the two, as they use very different core counts, clocks and configurations. Also not sure if running the XR2 Gen 2 faster cores at just 64% the frequency or less (if they are X3) or 73% (if they are A715) as on the SD8 Gen 2 still even counts as underclocking. Low power cores are useful to extend the runtime on phones which mostly idle with background tasks like checking for notifications, something HMDs just don’t need, so maybe the 2/4 on XR 2 Gen 2 are 2 A715 performance/4 A710 economy?

        Unfortunately trying the obvious and checking the specs at Qualcomm didn’t work, as their XR2 Gen 2 page doesn’t even bother mentioning the CPU and only emphasizes the 2.5x GPU and 8x AI performance. If someone has a reliable source for the specs or even benchmarks that make this less of an apples vs oranges comparison, I’d be very interested. Even clearing up the names could help. Is the SD8 Gen 2 1/2/2/3 config 1 high performance/ 2 performance/2 efficiency/3 low power, and does XR2 Gen 2 use only performance/efficieny cores?

        For the 33% I may have fallen for the ambiguity of the English language, and the Meta slide before only list the speedup over the Quest 2 as 34% for CPU/GPU neutral, 16% for GPU heavy and 44% for CPU heavy, without mentioning if this relates to the Quest 2 at launch or today. To add even more confusion, there is a thread on Reddit discussing that the XR2 Gen 2 CPU is again massively overclocked, but that may be partly due to confusion about (high) performance core burst clocks, while Digitrends quotes Qualcomm in the opposite way: “The Snapdragon XR2 Gen 2’s CPU features two performance cores and four efficiency cores, outpacing even the speedy Snapdragon 8 Gen 2, according to Said Bakadir, Qualcomm XR product and business lead.”

        Mood: confused

        • MasterElwood

          With the 8x improvement in AI – they probably moved some background tasks like parts of tracking from CPU to AI = more headroom for the CPU even it doesn’t show in classic benchmarks.

          • Christian Schildwaechter

            The tracking on Quest actually was never done on the CPU, it was running on the Hexagon DSP (digital signal processor) part of the Snapdragon 835 on Quest 1 and it counterpart on the XR2. Google already used the Hexagon for signal fusion to improve the 3DoF tracking on DayDream, so it is very likely that Meta also already utilized the DSP on the Oculus Go.

            AI/machine learning is still based on math, and the big improvements in numbers currently come partly from making the math “more stupid”. Traditional calculations on the CPU are done in 64bit, giving you about 16 decimal digits of accuracy for FP64. Graphics don’t need this, so they use 32bit, allowing them to place two calculating units in the same space as one on 64bit, doubling the speed. The 8x improvement on the XR2 Gen 2 is only achieved at Int8, so at 1/8th the CPU accuracy with no fractional values. But that happens to be just fine for the math needed for the neural networks driving most of AI today, while allowing to squeeze eight parallel calculations into 64bit instead of just one, leading to the impressive speed ups.

            You can place these “dumb” calculators anywhere, CPU, GPU or DSP. And usually they are an extension of the already existing math units there. In the Snapdragon family, they are placed on the DSP, so the gains in AI sort of come at cost of signal processing power, as you (simplified) can do either, but not both at the same time. Meta now uses more machine learning, e.g. for the estimation of leg positions, so the AI performance improvements will be beneficial, but typical apps will still rely mostly on CPU and GPU directly. And the benchmarks I quoted were explicitly only for the CPU, as the driving question was whether the CPU part in the Quest 3 is significantly underclocked compared to what the XR2 Gen 2 specs would allow.

          • XRC

            Daydream tracking (had pixel XL) was very performant and surprisingly robust. Hexagon DSP, Qualcomm 821.

            And Google’s Worldsense tracking on mirage solo further impressed, “Bladerunner Revelations” using Seurat graphics technology and 6 DoF head tracking was a stunning achievement for mobile VR. Though the heads/hand tracking mismatch was reminiscent of a Lazer pointer fixed to the hip.

          • Christian Schildwaechter

            TL;DR: Google messing up Cardboard/Daydream was a tragedy and a huge loss for humanity and VR

            Google trying to fix Cardboards worst issues in software instead of relying on extra hardware like GearVR was right. But by implementing the improved tracking only on Hexagon, thereby limiting Daydream to a few certified/expensive SD82x phones, they condemned it to a tiny niche, and ultimately killed the platform. Not allowing other phones to run Daydream apps with e.g. an AppLab like warning was a very bad choice, that option would have incentivized developers to swap the Cardboard SDK for Daydream, and phone manufacturers could have gradually updated to full compatibility.

            I still wonder what opportunities have been lost by Google badly handling phone based VR. It couldn’t compete with dedicated standalone HMDs with better tracking, controllers, resolution etc., and restricted input severely limited the uses cases. But today we have powerful 120Hz phones with decent voice recognition, and if Daydream had been a success, more of them would have at least 1440p screens, power modes for sustained performance and reduced screen latencies, solving many of the comfort issues.

            For me the interesting aspects of Cardboard/Daydream weren’t the games, but the educational and cultural experiences. Esp. on Daydream there were a lot of these, much more about exploration that didn’t need complex input or being usable for hours. Those were pretty much lost. There are some on Quest, but they are mostly ignored, as it is seen/used mainly as a gaming device for USD 299+.

            And this is where phone based VR still really could and should shine, lowering the entry bar to getting a tiny viewer that can even be made out of cardboard, though there are many better solutions for less than 10% the price of a Quest. For a long time I carried a foldable “Pocket 360” VR viewer from I AM CARDBOARD, doing a lot of impromptu VR introductions, with a lot of WOW instead of the decried “poisoning the well” nonsense. It’s a tiny 70x70x20mm incl. case at 60g (10g for the case), and using Fresnel lenses and some clever Origami, it could be reduced to a flat sheet that is part of a phone cover, so you’d always have it with you.

            By the end of this decade, there will be more active smartphones than humans, each faster than your already usable Pixel XL. It’s such a shame that Daydream never grew into a market with decent VR viewers available everywhere for USD 10, and lots of smaller, but valuable VR experience about art, music concerts, travel, nature, complex science visualizations, stereoscopic 360° video, and even some games, for the devices we bring with us everywhere. With widespread ARKit/ARCore support, even 6DoF head tracking is now possible, and hand tracking just a question of time, so todays experience would be quite different from 2014 Cardboard.

            IMHO this would have a much bigger chance to go mainstream, and push the metaverse into existence, than either trying to grow Quest out of its gaming dominated VR enthusiasts niche, or lure more users into VR with smartphone versions of Horizon Worlds, Rec Room or VRChat, or hoping for true AR glasses to become technically feasible anytime soon. With billions of potential users already owning a smartphone, phone based VR going mainstream would also give a gigantic boost to user numbers and hardware/software development on dedicated VR HMDs. But unfortunately we get none of that, because Google fucked it up and dropped both Daydream and Cardboard a few years ago.

          • XRC

            The death of Daydream was a tragedy, as you mention some incredible experiences, experiments, educational, cultural software. Google stories were excellent, YouTube VR provided me with hours of content and increased resolution was noticeable compared to Rift CV1 and Vive headsets of that time.

            WebVR support in Chrome and Firefox, tight integration with other Google apps. Going into the street to shoot 360 with sound, taking that straight into Daydream headset, fantastic stuff. That little handheld remote worked ridiculously well, and developers used it very successfully.

            Met some of Google’s AR/VR London team who worked out of Tottenham Court road. All owned Rifts and vives at home, two team members visited my HTC Vive “Roomscale Plus” experiments in a vacant retail space in Canary Wharf Jan 2017. Wasn’t long until they were disbanded, leader going to California to work for X division.

            Thankfully some Daydream games escaped to other platforms like virtual virtual reality and eclipse edge of light; very pleased to see Google’s Audio Factory and Google Stories all made it to Steam for 6DoF PCVR consumption.

            What could have been…

        • Isaac

          The XR2 Gen 2 almost certainly doesn’t have 2 X3 cores. Those are huge in terms of die space and don’t do well when underclocked.
          The 2 performance cores are likely similar architecture as the efficiency cores, but with higher clocks and cache.

          • Christian Schildwaechter

            Which would support a 2 A715 performance/4 A710 economy configuration. That way the Qualcomm statement regarding XR2 Gen 2 outpacing SD8 Gen 2 would sort of make sense, as sustained performance may be higher with 2/4 performance/economy cores on XR2 compared to 2/3 on SD8, where the extra high performance and low power cores only come into action during either burst or idle activities.

            The remaining question regarding “XR2 outpacing SD8” would be clock rates. Both types of cores run at 2.8GHz in SD8 Gen 2, but Meta only lists 2.05Ghz in CPU heavy and 1.6GHz in GPU heavy. I doubt that the default clock for sustained performance is so much lower on the XR2 Gen 2, so the CPU is most likely again significantly underclocked on the Quest 3, but we would need actual specs incl. core clocks for the XR2 Gen 2 to finally determine that.

        • T-ReX

          There is little difference between perf/efficiency of a710 and a715. The reason two a710 are used in 8g2 is its the last arm core with 32 bit compatibility.

    • Ben Lang

      Interesting thoughts. I would guess that the CPU clocking is more about heat and battery life than trying to restrict Quest 3 to play in the ‘Quest 2 generation’, but you could be right.

      • Christian Schildwaechter

        I’m now wondering if I have misinterpreted your sentence about the performance compared to Quest 2 at launch, and I cannot find other references, so hopefully you can clear this up:

        Meta says it’s 33% more powerful than Quest 2 at launch,

        Did you mean that
        a) the launch performance of the Quest 3 is 33% higher than (current) Quest 2 performance, with “at launch” meaning Quest 3 performance will probably improve further later
        or
        b) the current performance of the Quest 3 is 33% higher than the Quest 2 performance was “at launch” as in 2020, with the Quest 2 performance improved since.

        I read it as b), but now realize that the sentence could actually mean both.

        • T-ReX

          It probably is the heat and battery life concerns, the gpu efficiency has gone up much more than cpu efficiency, at 4w 8g2 is only about 30% faster in cpu multi, while gpu is more than twice as fast in gfxbench.
          socpk(dot)com/cpucurve/

  • Bumpy

    Maybe now we’ll finally get more quality VR games on PCVR from Q3 trickle down. The wait continues.

    • ViRGiN

      Crossfire is a high quality game and it’s a complete flop, 8 players right now.

  • JanO

    @Ben, how are the display brightness and color reproduction?

    • Ben Lang

      Brightness and color are fine, but it’s LCD and not HDR. I’m pretty happy with the blacks, but still not true black.

      • ViRGiN

        You should start calling Quest 3 high-end headset, cause that’s what it is, highest end possible headset.
        Abandon your stupid definition “if it doesn’t cost at least $800 for full set, it’s not high end”

        • Mike

          Lol. Highest-end for a standalone? Sure. Overall? Hell no. I still choose my 2018 Odyssey+ over this.

          • ViRGiN

            LMAO. So what are you representing yourself with this 5 years old headset? Preference for piss poor sound, extremely horrible controllers, non-existeng tracking, tethered experience with no way to replace a cable when it inevitably goes bad, unconditional love for pentile 1440p display, driven by long abandoned and unsupported windows mixed reality platform?

            The only redeeming quality of it is AMOLED, I’m a sucker for xOLED. But claiming you would choose it over high end Quest 3 rather shows that you want to troll, or your case is extremely specific like those mysterious “simmers” who think they represent and influence the market in any shape or form.

          • Mike

            Not 1440p. 1600p, with an optical filter that makes it look over 4K downscaled to 1600p rendered resolution.

            Sounds seems standard to me.

            Contrast and black levels are high-priority for immersion.

          • ViRGiN

            You can render 4K on DK1 headset.
            You seem to be coping hard and shilling for something probably even you aren’t interested in buying, all in turn to dismiss Quest 3.

          • Mike

            Lol, you completely missed what I was describing. In regards to SDE, it appears as ~2200 lines (as explained by Samsung in a presentation, and as appears accurate in usage). What I said about “rendered resolution” is in regards to the fact that it only actually has 1600 pixels. It basically creates a secondary, transparent display of “virtual pixels” at 2200p, which averages the light values at each point to create each virtual pixel.

            And yes, I preordered the Beyond. Projected to arrive this month.

          • kraeuterbutter

            realy ? i had the Odyssey+ as well…

            didnt like the colors.. was to “Samsung phone-like”.. to cartoonish, too saturated
            then it had mura
            comfort for me was terrible
            tracking was good enough, but the the area where the contollers were tracked too small, controllers bad/cheap
            WMR-Software: not comparable to a native steamheadset or Oculus/meta
            many problems with WMR (also with the Reverb G2)

            so: if “black blacks” are the only thing what bothers you.. yes.. than you can take the Odyssey+
            my index replaced my odyssey+ then
            the odyssey+ was not a bad headset at its time.. the best WMR at that time
            but never would i want it back

          • Mike

            Hmm. Real-life is often saturated. The O+ saturation varies depending on how a scene is designed. Good for saturation to be possible.

            Yes, controller tracking takes getting used to. And it’s not comfortable. But if you use a VRcover and wear it just right, it’s no more uncomfortable than the average headset.

            But comfort is the biggest reason I’m excited for the Beyond. First VR headset that’s LEGITIMATELY comfortable, and it also greatly improves on the most important specs of the O+.

      • JanO

        Thanks! Coming from Quest 2, I can deal with LCDs, at least until micro OLEDs are more affordable, but the lack of true HDR for the price of Quest 3 is yet another factor that makes me hesitate…

        With Meta’s misplaced focus on MR, the Q3 might be better than Q2 in every aspect, it just doesn’t feel like it’s all it could have been had they focussed solely on VR and kept MR for the Qpro line for another gen or two…

  • mbze430

    Mine isn’t coming in the mail till tomorrow (10/11). Not sure why that is, when it supposed to be here today. Anyway, I originally wanted the Quest 3 to replace my Oculus Rift. Now I am really interested in Mixed Reality. Although knowing that it now support AV1 I can’t wait to try PCVR on my 4090 and 4090 Laptop

    • Gonzax

      Same here, I pre-ordered on Amazon but it’s been delayed until Thursday. My girlfriend’s nephew bought it from the official store and it’s just arrived, literally 10 minutes ago even though the official date they gave him was November!

    • Totius

      Mine has just arrived! but my new laptop has not..

  • Klaus

    I would be interested in a comparision between Quest 3 and Reverb g2 in terms of clarity. Reverb 2 is somehow limited in terms of tracking and i´d like to go standalone/pcvr streaming in some cases. But the main focus would is still on clarity in simracing

    • ViRGiN

      No brainer. Quest 3 all day every day.
      HP G2 was an absolute mistake, and people bought it cause of MRTV shilling and paper specs.

      • eadVrim

        People bought it for racing and flight simulator not cause of MRTV, I have it and I am very happy with it

        • ViRGiN

          Of course you are happy with it.

      • Mike

        The only major thing wrong with the G2 is its black levels and contrast.

        • kraeuterbutter

          no… E2EC is the real bad thing with the G2…
          there are few headsets worse than the G2 in that field…
          colors and Black-Level are indeed for a LCD-Headset more on the good side…
          better tahn Pico3, Pico4, Quest2, Index, old PImax-Headsets, ….
          colors and blacks were one of the best in LCD when the G2 was released

          of course. not comparable with OLED
          but oled has other downsides…
          Mura, black crush, loosing details in dark areas, ….

          • Mike

            What is E2EC?

          • kraeuterbutter

            Edge to Edge Clearance…
            in former times called “Sweetspot”.. but nowadays sweetspot means: how big is the area you can place your eyes and image is as it should be… some headsets have small sweetspot.. you move a little out and it gets blury fast.. the quest3 has very very big sweetspot.. even when you move the headset some mm you can still use it
            E2EC means: how fast does the sharpness degenerate when going to the borders of the lenses/displays…
            the G2 is very sharp in the middle.. but just a little bit to the side you notice already a considerable downgrade in Sharpness..
            in the HMD-Testtool there is a spiraltext.. about 13 lines of text in a spiral written.. with the G2 only the first inner line is sharp.. the second line is already less sharp..
            the 5-6th line is readable and than it becomes bad
            on the pico4 for example line 10 is better readable, sharper, less degeneration than line 3-4 on the G2
            the Quest3 tops now everything (dont know how it realy compares to Aero and Crystal) in regards of e2ec and sweetspot

            very userfriendly and now: you can just move your eyes when you want to read subtitles of a movie an dont need ot turn your head

          • Mike

            Interesting.

          • Mike

            Mura isn’t inherent to OLED, and the Rever G1 had it.

            Assuming you meant “black smear”, what can be avoided by very-slightly limiting black-levels. All OLED headsets limit it by default, except the Odyssey, which can be limited to-taste using OpenVR Advanced Settings.

          • kraeuterbutter

            i know…
            with black crush i mean: in dark scenes, black is black – great – but: some details are missing because it gets too dark..
            i noticed that back than when i replaced my Odyssey+ with the Index (Pentile Oled vs. Index)
            the experienced SDE was very comparable maybe even better on the odyssey+ thanks its anti-SDE-layer.. but of course also less crisp for that
            balck crush: i played MOSS back than, and compared A/B the Odyseey+ with the Index.. and in the dark scenes of Moss i could see textures on doors, and trees with the index, i could not see with the Odyssey+
            i had the same effect when comparing some years later Quest1 (Oled) with Index and Quest2 (pale colors) when playing VAder immortal
            with the Quest1 – yes, the atmosphäre was greater, dark scenes …
            but: ont he other hand i could not see details i could see with the index in the same scenes
            with my Pimax 5k OLED: the same thing

            so.. well.. i was never “hunting” this black-blacks..
            maybe because my TElevision is LCD, my computermonitor is LCD, notebook has LCD .. so iam “used” to not THAT dark colors and so they dont bother me so much

      • kraeuterbutter

        i have the G2 – first i nearly send it back because of the bad E2EC
        but you get used to it..
        and image-quality in all other aspects: the G2 spits the quest2 out of the water
        so quest2: used it only for standalone purpose
        for PC: INdex (sporty games), G2 (visuals) and Pico4 were used in the last years in parallel…
        the Quest3 hopefully will replace the G2, the P4
        the Index: for sportgames i think the index will still be king, latency.. will see

    • Gonzax

      I’ll let you know how it goes when mine arrive (they were supposed to be here today from Amazon but have been delayed til Thursday) but I doubt simracing via wireless is really a good option. Wired is still better; at least the quality on my Index, even with the outdated lenses, is much better than what I got on my Quest 2 via VD or Airlink and it’s more noticeable in fast games so racing is not the best option IMO but there should be some improvements on Q3 so hopefully it will be better than before.

    • Dragon Marble

      In terms of the overall clarity across FOV, Q3 blows G2 out of the water. I just got my Q3 but I have been using Quest Pro for some time. Compared with Pro, G2 still has the advantage of higher resolution at the center. Q3 has the same level of resolution as G2, but with much better lenses. Much of G2’s pixels are wasted due to the small sweet spot.

    • david vincent

      Every simmer will tell you that display port is pretty mandatory for simulations as you want the greatest image clarity and performance and a latency within the industry standard (<20ms).
      The Pico Neo3 Link could be a great update to your G2, I use it for race and fly simming, the image clarity and lack of SDE are excellent.

  • MosBen

    Great review. I’ll admit that since we had a daughter a couple years ago I really haven’t had time to play with VR much. And while I found myself considering whether I’d pick up a Quest 3 while reading the review, if I’m being honest with myself I still probably don’t really have the time to fit that into my life. But it certainly is a neat signpost on how far VR has come since the first Oculus Kickstarter 11 years ago. And while I’m sure that it will be another three years before we get another update, it is interesting to think about what the Quest 4 will bring to the table. Hopefully we’ll start to see wider field of view, smaller form factors, and more. Should be an interesting few years.

    • kraeuterbutter

      i feel you – two kids, 2 and 4 years old
      for that reason a standalone-Headset (Quest2 and Pico4 for me, tomorrow Quest3) is ideal…

      you can use it just for 10min
      for such short periods i would not bother to use my gaming-rig with index or reverb g2 or the pimax
      so the most used headsets are the standalone-ones

      even when using it with wireless pc-vr: its nice not to have to bother with a cable when storing the headset

      and: you can paus the gaming at each point with single button-press, put the headset on your shelf and continou in seconds with a single button press the next day

      and: you can use the headset on the balcony, in the garden, in bed… so even when your regular lighthouse-playspace in the livingroom is blocked by a big pillow-castle the kids build

  • Dragon Marble

    Just got mine and tried First Encounter. Default strap works fine for me. You can adjust the eye relief one side at a time, so you don’t need to push both buttons at the same time.

    • Hussain X

      Yes you can adjust one side at a time. I do one notch each side at a time.

  • Bart

    Tested quest 3 as well and i left a bit disappointed.
    – Fov is wider, but there are some cut out corners now, so it’s only really wider in the middle.
    – In contrary to most reviews i noticed a massive amount of glare while watching movies, certainly not less than quest 2.
    – a lot of aliasing problems all over
    -you can still see the screendoor effect.
    -the brightness across the lens is not consisten. it’s way darker in the corners.
    – the blacklevels, just as grey as the quest 2 and it’s preventing feeling really immersed.
    Will not buy an lcd headset again. micro-oled it should become.
    having said all this, yes it’s better than the quest 2, but the jump in visual quality isn’t as big as suggested by many reviewers.

    • Mike

      There will never be an LCD VR headset with good contrast. They could have at least had good black levels using a dynamically-adjusting backlight – but nope.

      No point in buying this over a Beyond unless you really need all-in-one / wireless.

      • MasterElwood

        Beyond 3x price (999 + base stations + index controller)

        Beyond way smaller FOV

        Beyond no MR

        • Cl

          According to the chart in this review the beyond doesn’t have way smaller fov. It does have worse edge to edge clarity, but that’s a tradeoff of better resolution, comfort and contrast.

          If you already have basestations and stuff the price isnt really much different if you buy a few accessories for quest 3.

          Depends on use case.

        • Mike

          Maybe smaller FOV, but still higher than on Quest 2.

          Not 3x price for people who already have base stations and controllers (many people do).

    • Muhammad Jihad ✓ᵛᵉʳᶦᶠᶦᵉᵈ

      I wasn’t expecting to get one, I’ve had every Oculus unit to date since DK1, but wasn’t impressed by the spec sheet on this one. However, after getting it and using it, I’m liking it quite a bit.

      It’s possible you got a dud of a unit, because most of the stuff you’re whining about isn’t a real issue – except for black levels, and I already know the black levels were going to be garbage. Just saying it should be OLED, when you already knew it’s not, is a dumb point.

      I haven’t noticed any issues with glare. The fake ‘screendoor’ nonsense is still not an issue, never has been. And for the corners, again, I’m not noticing anything like that.

      As for the leap in visual quality – what have you actually played? Let me guess, a bunch of stuff that was made for the Quest 2, that run the exact same is what.

      Overall, what a whiny, worthless post – and based on your comments, I’m thinking you didn’t actually try it, or you did so for like 5 minutes. Because anyone who has actually tried one of these would absolutely talk about how good the mixed reality stuff is. I didn’t think I’d care in the least about it, but my goodness, is it pretty amazing.

      Also the Quest 3 is slightly more comfortable, and literally everything about it is better than the Quest 2. Only real issue is the price, in that outside of Mixed Reality, it’s really not a huge upgrade, but it’s clearly an upgrade.

  • Mike

    This is a joke of a headset for late 2023.
    Still LCD? 515g? Only 2064 lines? Not worth $500. Maybe $300 at most.

    The Bigscreen Beyond is out already.
    OLED, 127g, 2560 lines.

    The PSVR2 is better – been out a long time.

    • ViRGiN

      You’re masturbator of anything OLED, and still fail to mention that Beyond isn’t 2560 natively. For such a “quality” lover you show complete ignorance.

      • Mike

        Contrast and black levels are high-priority for VR immersion.

        It’s native 2560 if you use the lower refresh rate setting. But I thoroughly researched the compression used for full-refresh-rate 2560 – everyone agrees you can’t see any difference whatsoever from uncompressed.

        • ViRGiN

          Oh so refresh rate isnt a high priority for VR immersion?

          • Mike

            90 Hz is enough that other things are much higher priority. It’s into “diminishing returns”.

          • ViRGiN

            Ah, yes, especially the resolution.

    • kraeuterbutter

      depends on what your preferences are..

      the PSVR2 has a tiny sweetspot
      E2EC is not good, just like on other “old-generation”-FResnel-Headsets
      comfort is for many people not that good as well
      the lenses of the Q3 blow the PSvR2-lenses out of the water
      mura is for many people big issue with the PSVR2 and ALL Oled-Headsets out there suffer on that mura-Problem
      (micro-oled should solve this issue)

      if blacklevels and hdr is important for you: there the psvr2 can shine
      for me – owned 5 OLED-Headsets and 5 LCD – iam ok with LCD and in the moment: all put together – LCD gives me – for now with the headsets i own – a better over-all experience (more resolution, pancake Lenses)

      hope to check out the beyond soon

      • Mike

        Mura isn’t specific to OLED. The LCD Reverb G1 had really bad mura – main reason they made the G2.

        The O+ doesn’thave actual mura. It has small leftover hints of the pixels only visible on bright & uniform surfaces, after going through the optical filter that makes 1600 lines look like ~2200 lines (rendered at 1600). Looks like wearing imperfect glasses, rather than a pixel array.

    • CrusaderCaracal

      Bigscreen beyond costs $1000, NOT INCLUDING controllers (cheapest new are vive pro wands, 2x$200= $400) and 2.0 base stations (2x$212= $424).
      Add that up (not even including shipping) that’s $1824 USD. A staggering $1174 more than the 512gb Quest 3 at $650. So of course the Beyond is going to be better when it’s more than a grand more expensive than the Quest 3.

      PSVR 2. $550 USD NOT INCLUDING a Ps5 ($500 USD at best buy).
      Also, the PSVR 2 suffers from a limited game library, incompatibility with PCVR and general issues with the design (bad tracking for controllers, glare)

      Therefore, your cherry picking isn’t going to help you in this case. Whilst some features of the Beyond or the PSVR 2 may be somewhat better (which is negated by both of the headsets having issues or not being ideal as a first vr headset setup) they are MORE EXPENSIVE.

      Maybe do some research before you decide to bash the Quest 3.

      • Mike

        Many people already have base stations and controllers. Not fair to include that in the price without that caveat.

        • Andrew Jakobs

          But also many don’t. If you never had an Index, Vive Pro or regular Vive, you probably won’t have basestations and controllers.

        • CrusaderCaracal

          People don’t want to have to buy all of that though, they just want a headset

    • Andrew Jakobs

      Beyondscreen is very costly, is fixed to one user, is cable only, no controllers and requires the lighthouse basestations, not included. So, no, pricewise the Quest 3 is much better.

      • Mike

        Not exactly fixed to one user. They’ve stated that people within several mm of the setting should still be comfortable. And there was a product teardown showing that it’s easy to open and replace the IPD setting plates. And it can be worn without a custom facepad (using the rigid strap), and modders like VRcover might make a universal facepad.

    • Garhert

      Why do you even care (and complain) about standalone HMDs then? Taking out the SoC alone would probably bring down the price by $150+. This is not made specifically for PCVR. If you wanna use it then fine, but don’t complain.

  • Chris Meeks

    Enjoy mostly Quest 2 games until it’s discontinued.

    • CrusaderCaracal

      Quest 2 isn’t going to be discontinued for a while. Even if it does get discontinued, it doesn’t really matter. My Vive Pro bought in 2019 is still kicking strong.

      • Andrew Jakobs

        Also still using the Vive Pro(1)/wireless module with great pleasure, but the Quest 3 is really getting interesting as I do get more and more annoyed with the godrays/fresnellenses, and ofcourse the SDE is pretty visible, but luckily I haven’t used another headset yet, so I’m not spoiled yet, haha. But the OLED panels/colors of the Pro is what I think also still makes me hesitant to upgrade. Ofcourse I would be using the Quest 3 mainly as a wireless PCVR headset.

        • CrusaderCaracal

          Didnt ask

  • kraeuterbutter

    john carmack:

    havent tried for long time: whats with the Netflix app in the Quest STore?
    is it still – like on the Oculus Go – limited to 480p as it was when programmed in 2 days by john carmack?
    now with the quest3, good lenses, higher resolution it would be fine when the Netflix app finaly reaches fullHD
    (as the amzone Prime Video-App does)

  • MasterElwood

    U2VR mentions in their review a 50 percent better contrast?

    Is this true?

  • ZarathustraDK

    Paying to be the MR-canary in the coal-mine, no thanks. People’s fascination with this is ridiculous. Everybody is like “dude, it’s the future!”, yet no-one is capable of coming up with anything beyond breaking virtual holes in the wall and playing Demeo on a coffee-table, whoopdee-fucking-doo. You’re not going to have a use for MR before it can be shrunk to sunglasses size and the battery-life allows comfortable evening of use without the worry of running low; and that future is far away.

    • gothicvillas

      I kinda agree with you… but I find it pretty cool to play Demeo and watch news on telly in the background. I could have a game while following Gaza/Israel conflict. On the other hand am not sure if passthrough will be good enough for me. I hear some worrying reviews and I just cant stand subpar experience.

      • CrusaderCaracal

        Jesus christ that’s overstimulation to the max. Maybe you should also get an monitor playing family guy clips aswell

    • ViRGiN

      Who the hell are you reading/watching?
      It’s a stepping stone towards MR.
      Look at 2016 VR. WTF was that? Even valve wasn’t able to come up with anything beyond a bow game where you shoot cardboard cartoons and a bunch of other gimmicks. There is no future for VR before it can be shrunk to sunglasses size and the battery-life allows comfortable evening of use without the worry of running low; and that future is far away.
      Explains valve lack of any interest.

  • Lucidfeuer

    Very good, exhaustive review, minus the passthrough FOV° missing. Haven’t tried yet, the passthrough doesn’t look bad except the incorrect near-object warp correction (which I was pretty sure they would never nail). I will wait for a decent strap replacement and see after I test it at office if I can see a use for it.

    • ViRGiN

      You sometimes forget you’re a nobody of the worst kind, a parasitic shill, so your remark is funny..

      • Lucidfeuer

        So CrusaderCaracal is your alt account?

        • ViRGiN

          Are you at least paid? You know you could be and are missing out if not…

    • CrusaderCaracal

      Maybe meta should hire you to fix their headset

      • Lucidfeuer

        Already have a job

        • ViRGiN

          A job? Lmao. That’s so low level. I thought you would own a business that makes money on it’s own. Pathetic. I held you in higher regard, but not any longer. Learn how to business

        • CrusaderCaracal

          Yeah, proffesional hater

          • Lucidfeuer

            Almost literally, but with a much more productive, specialized and informative approach. I’m not gonna boast more

          • ViRGiN

            You sometimes forget you’re a nobody of the worst kind, a parasitic shill, so your remark is funny.

  • Lucidfeuer

    Then you’ll be disappointed and have to tolerate it.

  • david vincent

    You forgot again to mention the lack of display port, which is a no-go for some users.

    • CrusaderCaracal

      Pretty obvious there wouldn’t be one, meta hasn’t had a displayport since the rift s

      • ViRGiN

        And VR has only got better ever since.

      • david vincent

        Not everyone is as informed as us, you know

      • david vincent

        Not obvious for everyone

  • wowgivemeabreak

    Got my Q3 today and there is absolutely a mura-like effect that shows up on solid light colour or white areas/backgrounds when you move your head. Enough people have commented on this at the various reddit subs to make me pretty confident it is there on every Q3. If it isn’t and this is not normal then that is some super shoddy quality control on their part.

    I also find glare to be worse than all these “pro” reviews indicated yet still not as bad as Q2 god rays.

    FOV to me was very nice with more vertical than I was expecting based on reviews and measurements. It is absolutely increased over the Q2, and I was using a VR Cover slim pad on my Q2 so eyes fairly close to the lenses. I first noticed this while playing an experience in Synth Riders.

    Colours are better than on the Q2 and contrast ratio does seem improved. Seeing this makes me think the panel the Q2 uses wasn’t even getting 100% rec.709 or srgb coverage while this definitely is.

    Overall, it seems like a solid headset with great clarity that is especially beneficial for PCVR. Just wish that mura-like or dirty screen effect wasn’t there since it slightly takes away from how great the clarity is.

  • Great review, as always!