Metaboba 3 Prototype Hands-On: An Uncompromising Extra-Long View

Table of Contents

At Siggraph 2025 this week, I practiced Boba 3, a meta prototype for ultra-width view PC VR in a form factor similar to Quest 3.

I have long argued that vision is the fundamental driving force behind VR immersion, and that the industry should prioritize increasing it.

Until Boba 2 and Boba 3 were revealed, it was unclear whether a significant trade-off would always be required to form factor and weight by significantly increasing the field of view. If that were the case, it might have been decades since we saw big players push their vision once again, given the industry’s legitimate focus on making headsets smoother and lighter.

After trying out the Boba 3 and seeing a bigger perspective with a viable form factor for my own eyes, I am now more optimistic than ever about the future of VR.

Stagnation for 10 years

When I tried out the original Oculus Development Kit 1 (DK1) in 2013, like many others, I was blown away.

Today, DK1 is in most respects a dinosaur. Its resolution was less than 640 x 800 per eye (approximately 6 pixels per degree, legally blind), and the refresh rate was just 60Hz. There was immediately a full pergence display with noticeable motion blur, no position tracking, and only 3DOF rotation.

Still, despite all this, most people wearing DK1 cried out that they felt they had been transported elsewhere.

Before the DK1, the only headset that could be purchased for less than tens of thousands of dollars had a field of view of about 45°. All I could do was peer at the slivers of the virtual world. However, with DK1, it’s finally here in That virtual world. Therefore, the consumer VR revival of the 2010s began.

*Note that these values are approximate as vision varies from person to person based on face shape.*
	Horizontal Field of view	vertical Field of view	release year
DK1	90°	90°	2013
DK2	93°	99°	2014
lift	87°	88°	2016
Quest	93°	93°	2019
Quest 2	96° (IPD 1 and 2) 86° (IPD 3)	96°	2020
Quest 3	108°	96°	2023

Today we have a wireless standalone headset of 4K per vision, a 120Hz refresh rate, low persistence, and ultra-low latency markerless inside-out head, controller and hand tracking. Not only VR, but also mixed reality color pass-throughs, headsets no longer require gaming PCs.

But most of the time, their vision is barely upset. Since then, almost all of the headsets have landed between 90° and 110°. For better or worse, the focus is not on vision, but on making the headset smaller and higher resolution.

There are exceptions. Pimax shipped a horizontal field of view of about 150° in 2018, but sacrificing “hammerhead” form factor and fish-level geometric distortion. These days, the Somnium VR1 and Pimax Crystal Super have reached a horizontal level of around 130° for some, with a very high price and trade-off for some.

However, most of the time, the field of vision of the headset people actually buy and use Masses are stagnant. VR has become sharper, clearer, and much more interactive, but not immersive. That’s why the Boba 3 lens is one of the biggest VR breakthroughs in a decade.

For years, the Display Systems Research (DSR) team at Meta’s Reality Labs Research Group (originally Oculus Research), led by Douglas Lanman, publicly presented and sometimes demonstrated research prototypes that promote the state of art for head-mounted displays.

This includes a variety of optics, retinal resolution, ultra-brightness, and undeniable pass-through.

Back in 2018, DSR showed off their original half-dome headset.°. However, subsequent versions of the Half Dome have a smaller field of view, with Half Dome 2 focusing on compactness, while Half Dome 3 moved to an electronically fluctuating approach without moving parts. In many ways, this reflected industry-wide priorities.

Until recently, we had no broad-view metaprototypes at all. Then in October 2024, Meta CTO Andrew Bosworth shared a photo of himself holding the headset in a very wide curved body, suggesting an ultra-wide vision.

Last week, Meta revealed that the prototype Bosworth held was called Boba. But more interesting than this name, it also revealed that there is a stack of displays and optics from the early “about nine years ago” of Oculus research.

The Boba 1 achieved an ultra-width horizontal field of view of 210° and features a key form factor trade-off, with its vertical field of view only 70°. Considering the low density of displays at the time, the angular resolution was also very low.

But many things have changed since then. So, recently, DSR has reconsidered what is possible with regard to field of view.

Enter Boba 2 and Boba 3

The Boba 2, first revealed in a paper abstract paper three weeks ago, was developed last year. DSR optics scientist Yang Zhao has realized that recent advances in pancake lenses, driven by a massive investment in Meta’s technology, can be adapted to achieve a very broad field of view in a practical form factor.

Additionally, VR’s commercial LCD displays have reached density, allowing them to achieve at least three levels of angular resolution in a very field of view. Specifically, the Boba 2 used a 3K display and reached 25 PPD, the same as Quest 3.

The Boba 3 is a new version that will reach 30 PPD using 4K LCDS from this year, and is a sophisticated lens production process that improves quality and clearly reduces the “peeling” seen at the edge of the lens of the Boba 2 display unit on the Siggraph stand.

Both have a horizontal 180 degrees and a vertical 120 degrees field of view.

For comparison, Quest 3 has a field of vision of less than 110° x 96°, while Quest 3 is less than 96° x 96°, and human vision is very about 200° x 135° depending on the shape of a person’s face. Simply put, Meta says Boba 3 covers about 90% of the human field of vision, compared to 46% in Quest 3.

It has an ultra-wide vision in form factors like quests, and a higher angle resolution than quest 3. In theory it seems great. But does it stand up to real-world demonstrations? And what is the catch?

Boba 3 and the joy of complete immersion

Wearing a boba 3, I had no choice but to smile. What I saw was an image as clear and sharp as Quest 3, but it’s both horizontal and vertical, and really surpasses a huge field of view. It involved most of my vision, with only small chunks around extreme horizontal periphery, not being done vertically beneath me.

Unlike most other attempts in field of view so far, Boba 3 has very slight edge distortion, comparable to what you’ll find at the edge of Quest 3.

In VR mode, the Boba 3 gave me a deeper sense of being truly somewhere else than the other headsets. And isn’t this the real purpose of VR? The senses were so intense that after returning to Quest 3 I was keenly aware of the vision I had now lost and I couldn’t stop looking at it.

And this strong immersion is not the only benefit of an ultra-wide vision. It also improves object persistence. This means you are continuously aware of the virtual objects that are next to you, not just at the front. The virtual meeting continues to watch tracked body and facial movements of all participants. In the game, you will see approaching enemies much faster. Alternatively, when using the triple monitor virtual workspace, you can view the side monitor with your eyes, rather than rotating your head. You don’t need to move your head when watching a 180° video.

Reducing the need to pan along with your head is another comforting advantage, and it’s more subtle than just the weight of the headset. Certainly, many future headsets may be slim and light. But is it really comfortable to have to constantly turn your neck?

With a camera on the front, the Boba 3 variant supports pass-through mixed reality not only in VR but also in full ultra-width field of vision. Meta is technically impressive as he developed a custom FPGA pipeline chip with a fiber optic PCI-E interface to carry a dual 20 megapixel camera at a high frame rate (USB isn’t enough). However, the Boba 3’s passthrough was not depth corrected, and I moved my head to see the geometric distortion. Plus, I already have a peripheral headset and accessories that I have opened at home, so I can already see the real world around me. Certainly, it does not contain virtual objects. But nonetheless, this made Boba 3’s mixed reality experience slightly less impressive than VR.

That pass-through distortion, and lack of depth accuracy, exists because Boba3 uses “direct pass-through.” This is a flashy term that does not use dynamic reinjection, but simply displays the camera view with a matching lens but static projection. This is a limitation of time and scope of the project than some inherent technical obstacles, Lanman told me, and in theory could be improved with future developments.

With headsets, the difference is more This flat screen projection is dramatic as it distorts the periphery.

The most notable thing about the Boba 3 demo was that there was no “catch” to break the deal. The other three DSR prototypes I have tried include a prominent and obvious trade-off to drive the frontier of certain aspects of the HMD display system.

The Starburst provided a spectacular 20,000 NIT brightness, but with a headset, it had to be hung from top to top with a very large and heavy headset.
Butterscotch Varifocal provided nearby “retinal” resolution and mutation optics, but with a thick, heavy headset of 50° Horizontal field of view.
TiramisuI quickly share my impressions and combine high brightness and contrast beyond the resolution of the “retina”. But it is thick and heavy, with only 33° field of view.

That’s not to say that other prototypes aren’t impressive in themselves. But the Boba 3 is particularly different. That felt completely. If I was allowed to take you home or pulled away from a bold robbery, I would use it as a PC VR Daily driver. The Boba 3 is the best overall headset I’ve ever worn, with no 2 seconds.

But to be clear, the Boba 3 is not a product, and despite its sophisticated appearance, it is not a prototype of the product. Meta researchers repeatedly explained many of the Siggraphs to attendees. More than any other question.

“We are researchers. We don’t make decisions about what to ship to our products.”

That said, the most promising thing about the Boba 3 is that if Meta wants to ship it as a product, it’s probably possible.

The Boba 3 isn’t like Orion, the great AR glasses of Meta. The lens uses no exotic materials and the display is neither curved nor tiled.

In fact, the Boba 3 lenses are built in a process that is very similar to the Quest 3. They are dual-element pancake lenses, with new additions being a high profitable reflective polarizer layer. The display is a regular 4K LCD similar to the ones already shipped with the Varjo XR-4 and Pimax Crystal Super.

Conceptually, the Boba 3 is relatively simple. Certainly, it’s significantly more expensive than the Quest 3. But it doesn’t cost thousands of dollars.

That’s not to say that the Boba 3 was easy to develop, as DSR optical scientists want to be aware of. There were many smart optical and mechanical engineering involved here. But the key is that the difficult part was not theoretical manufacturing possibilities, but the early design.

The Boba 3 is upside down (this is a VR variant without passthrough).

There are no flaws in Boba 3 for me to consider deal breakers, but there are still trade-offs to watch out for.

For one, Eyebox is smaller than Quest 3. This means that you need to align your eyes with the center of the lens to see an acceptable image. Interestingly, what you see when your Boba 3’s lens is off-eye is a dark image rather than a blurry image. The IPD is aligned to one eye, but not the other, which can lead to an unpleasant brightness discrepancy.

This is one of the reasons why the Boba 3 is designed to be highly adjustable. It features IPD alignment for each eye, eye relief adjustment, and visor tilt control.

And when you say that the Boba 3 has a “similar” form factor to the Quest 3, you need to note that the lens is a little heavier because it is much larger. Given the increased weight and size of the lens, as well as the need to secure everything firmly on all possible adjustments and stress, the Boba 3’s visor is reinforced with aluminum rather than all plastic design. This is part of the Boba 3’s weight is 142 grams more than the Quest 3 with the elite strap, despite the fact that it doesn’t include a battery.

That aluminum could theoretically be replaced by a magnesium alloy, Meta researchers told me to lose weight. But this will increase costs.

	Field of view	Optical stack	Weight
Quest 3	108° x 96°	63 x 60 x 26 mm	698g (with elite strap)
Boba 1	200° x 70°	140 x 78 x 45 mm	874 grams
Boba 2	180° x 120°	80 x 80 x 42 mm	845g
Boba 3	180° x 120°	75 x 75 x 42 mm	840g

Incidentally, if you’re wondering how the Boba 3 is being tracked, the answer is that it used the original Oculus Rift Constellation tracking system, and the demo had two Rift sensors on the desk in front of me. However, if Meta wants to ship the Boba 3 as a product, this can be very easily replaced with inside out tracking.

Given how complete the Boba 3 feels, how few defects it has, and its apparent manufacturing possibilities, you may wonder why it is just a research prototype.

The decision on what to actually ship is completely out of the hands of the DSR team. Instead, there’s Meta’s product team, and ultimately executive Mark Zuckerberg and his CTO Andrew Bosworth.

Given Bosworth’s repeated statement, the decision not to ship products using Boba 3 lenses is probably dependent on calculations.

When asked about Boba 1 he held in an October photo, Bosworth reiterated his long-standing view that trade-offs on “weight, form factor, computing, thermal” are not worth it.

The Boba 3 certainly proves that it can achieve an ultra-wide vision without significantly changing the form factor. If it’s streamlined and refined for shipping, I think it’s only going to add about 100 grams of weight. It’s not a trivial thing, but the right head strap can be alleviated and is worth more.

But Boba 3 does nothing to address the calculation problems that Bosworth often raises. The wider field of view of the headset, the more pixels each frame needs to be rendered to achieve the same angular resolution. Also, performance costs are worse than increasing the resolution of the same field of view. This is because more geometry, entities, and effects are displayed and therefore need to be rendered and simulated. This even more overhead may be a bit more manageable in a virtual scene with few objects beside you, but in dense scenes such as open worlds, it can be devastating to performance.

Remove Boba 3 VR, no passthrough variants, side blockers.

The Boba 3 demo is powered by Nvidia’s RTX 5090. This is a $2,000 2-kilogram PC graphics card that draws around 600 watts of main power. Meanwhile, Meta has made a promise to ship the last PC-only VR headset in 2019, and not to ship another PC with the release of Quest 2.

Meta focuses solely on standalone headsets with mobile chipsets, pulling around 10 watts from a relatively small battery, with orders of magnitude less performance than a gaming PC. Within these constraints, the resolution, refresh rate, or field of view fields are traded strictly against the already limited graphical fidelity of standalone VR. As much as I hate to admit it, Bosworth is right when it comes to vision for a standalone headset.

However, this does not mean that Meta will continue to ship the same view of the same time period forever. There is a lot of space between 110° and 180°, and the basic approach of the Boba 3 lens can be adapted to gradually increase the field of view over time.

Tiramisu: “Hyper Rear VR”

The Boba 3 is pushing through the field of view, but still has similar angular resolution, with the same low brightness and contrast of a headset like Quest 3.

I also practiced tiramisu, a prototype of the “hyperrealistic” VR meta, combining high brightness, high brightness and high contrast beyond the retina despite its narrow field of view.