Hacker News Reader

> At this point it's clear that Facebook have "won" VR.

Eh, it doesn't look clear at all to me. It looks like its main advantage is slightly higher resolution than Valve Index and wirelessness. On the other hand from what I read Quest 2 only now will finally get 90Hz - this is what even old Vive provided. And Index goes beyond that - up to 120 and 144Hz. Also I wonder how Quest's screen quality compares to Index, besides resolution and refresh rate - couldn't find any information about Quest's field of view, pixel persistence or fill factor. And with Facebook behind Quest, I consider Index to be cheaper - I'll take $999 any day over $299 + my data.

I'd say it's hardly a clear "win".

> And with Facebook behind Quest, I consider Index to be cheaper - I'll take $999 any day over $299 + my data.

All that really means is you have enough disposable income that it's actually a decision worth even considering. For most people it's aimed at that's not the case.

This isn't Gmail for free vs some paid service for $5/mo, this is more than three times the cost for something that already costs multiple hundreds of dollars.

I'd say all it means is I value my data above $700, which I understand is not the case with many people. If I would not have enough disposable income, then I wouldn't buy Index nor Quest.

Your comment was about whether facebook "won" the VR segment, not whether facebook won over you specifically.

I'm sure many people in hackernews value their privacy above $700, me included, but that's clearly not the point here. We're just a blip in facebook's radar.

A squawked and ident'ed blip, linked to a/s/l and ML estimates of personal data self valuation and price sensitivity on VR headsets.

True. Then going by Steam's VR statistics from August (https://store.steampowered.com/hwsurvey):

46% for all Oculus types 36% for Vive+Index

Well, Oculus does seem to be more popular and perhaps that is what the parent meant by Facebook "winning" VR segment. To me "winning" sounds more like total domination, where any competition is near a statistical error. This doesn't seem to be the case here (though I admit Steam statistics may be skewed towards Vive/Index).

Oculus Quest in particular is in majority a non-steam headset. If you want to use it with steam, you need to buy an additional cable (Oculus Link).

When I see people talking about Oculus dominination technology-wise, they mostly talk about the Oculus Quest. More specifically, a wireless standalone headset that doesn't require static captors to have 6 degrees of freedom.

Someone correct me if I'm wrong, as I only started looking into VR recently.

Oculus Quest works with stock USB cable. Even via a passive, copper, extender to make it longer.

Oculus desktop software gives you a yellow mark, but games still work.

I'm not familiar with the Facebook integrations on the Oculus, can you clarify how the use the these devices would force you to give up your data?

Classic knee-jerk reaction. Don’t use Facebook and they literally can’t do anything with their data. I’ve stopped using FB beyond messenger and I doubt my data helps them in anyway shape or form.

Well, if you use it only for their messenger, then they can still for example track webpages you visit (either via cookies or some finer fingerprinting based on e.g. user agent).

It may also be a bit of a knee-jerk reaction, but in a world where even seemingly innocuous data is used to find unexpected correlations and identify people, I certainly prefer to keep a safe margin of error and steer away from companies which live on our data.

Quest 2 requires to have a Facebook account linked. And to have a Facebook account you need to provide phone number and real name (as per their TOS). If I provide fake information, they may terminate my account. And if they do so, I fear it would make that $299 piece of hardware useless.

Besides, knowing Facebook's hunger for data, I certainly don't trust them enough to install any piece of their software on my PC. Nor do I trust them enough to use a device which could potentially identify some unique traits of the way I move in VR (akin to gait analysis).

Not to mention you need a relatively beefy PC with Windows installed to use the Index.

The Index works just fine under Linux.

You're describing specs, not market share.

And comparing a 1000 dollar headset which requires a 1200 dollar computer to drive it, to a 300 dollar all in one unit.

It's akin to suggesting that Toyota haven't done well in the car market because they don't make Bugattis.

Facebook absolutely have won this generation in terms of market adoption. They've strategically withdrawn from competing at the higher end of the market. Which, at least today is a relatively tiny luxury segment.

> I'll take $999 any day over $299 + my data.

As someone already pointed out, that's a nice problem to have! For me, I considered the Index but the cost of upgrading my PC + the Rift S still came out a couple hundred £ cheaper plus I don't have to dedicate space to it.

>Is anyone interested?

Honestly, I feel like you could position such a project as "Deconstructing Facebook's business models one step at a time: The corporation" and hire a lot of anarcho-hacker types while positioning the company as a social-good venture. I personally would love to get paid to gradually undercut and destroy facebook.

I worry that 2 things would happen:

1. The company would become what it set out to destroy (evil)

2. Anarchist hackers are passionate but you'd have a hard time getting funding from any of the big reputable players.

Still, that would be an awesome company.

Open-firmware has had a long history, with varied degrees of success. From RockBox for the iPod to the ReactOS open re-implementation of Windows.

I think now is different. I think crowd-funding has changed the economics of such projects from a few passionate hackers working in their spare time to now being able to fund people working on such a project full-time.

There's probably tens of thousands of users willing to pledge $100 into escrow on a Kickstarter-like service, paying out when somebody to creates an open Oculus Quest firmware capable of 6 degrees of freedom using inside-out tracking.

This funding model and skillsets are suprisingly close to what's required to provide Android security updates to older phones and tablets, so there's a broader business to be built.

Many of those users are probably willing to pay a few dollars a month so such software gets continued to be developed.

Maybe, but I think the market for a Facebook-free Oculus is smaller than one might think. Your average xbox dudebro probably uses instagram quite prolifically and isn't in the "protect my privacy" camp.

I would 100% Kickstart a Quest-like (wireless, 6DOF, inside-out tracking) competitor that isn't wrapped up in surveillance capitalism.

I for one, would support that patreon!

Whenever you talk about startups with people outside the HN-like crowd, they always ask you about "your ideas", particularly if you assume you have all the funding lined up no problem.

My go-to response is always "put Facebook out of business".

I've been thinking about it for probably 10 years, I still have no idea what that means, but your post here got me a little stoked it could be possible

It's unclear how creating a what-sounds-like a rather expensive open source VR system could possibly put Facebook "out of business", considering probably 99 out of 100 people have no idea FB even owns Oculus.

Also, FB's core business is ads sold to businesses trying to reach users doing social networking stuff...how does this even touch that revenue?

At this point, I don't think FB will ever be put out of business given how old their core user base is, knowing that older people loath tech changes.

Why do you think the bulk of the Oculus innovation is deep in the stack? Is there much custom hardware in the Quest 2 or is it similar to a phone in a different form factor?

How much of the accuracy of the sensor fusion algorithm comes from the inherent accuracy/calibration of the underlying sensors?

Has anyone tried bringing up a stock android distribution on the Quest 2?

John Carmack does several hour-long sessions at every annual Oculus Connect event (now called Facebook Connect). [1] [2]

He has often spoken about deeply optimizing virtual reality by peeling back all the software layers. From his annoyance at the Android's process scheduling on the GearVR to writing custom microcode for the Qualcomm chip's Digital Signal Processor to try and optimize playing PC games (over Oculus Link streaming) by encoding and transmitting video per scan line (instead of per frame).

He said that up until a few months before the original Quest launched they weren't even sure whether they could get inside-out optical tracking working to sufficient quality, but that they just continued to iterate on the algorithms until it worked well.

It's clear that most of Oculus' innovation are deep in the stack, seemingly pushed forward by the legendary John Carmack. Now that Carmack is merely the "consulting CTO" rather than working on these problems full-time there's real questions about whether Oculus' deep stack innovations will continue.

Oculus Quest is effectively a smartphone in many key ways, but it has several monochrome wide-angle cameras for tracking. Also unlike smartphones it has active cooling fans for improved thermals. It seems there's a million little things than come together to make Oculus Quest standalone VR way better than any phone-holder based virtual reality platform that I've used.

As far as an open-source custom firmware for Oculus Quest goes, I am unaware of any project that has yet cracked the bootloader to run unsigned code and then booted stock Android.

[1] https://www.youtube.com/watch?v=PMIDaomx0GA

[2] https://www.youtube.com/watch?v=sXmY26pOE-Y

Inside-out tracking works perfectly well in the Windows Mixed Reality headsets too. It's not insurmountable.

So it takes a $500+ billion company to crack this nut.

Perhaps, but there are several of those now. Apple, for example.

Transmitting video per scan line sounds very much like slices to me. And those are not particular new.

Google and HTC both independently failed to match Oculus. I don’t think an open source implementation at this stage would work at all.

If not now then when? Given the top-notch specifications, Facebook appear to be selling at least for the 64GB Quest 2 at a significant loss. The advantage of an open-source custom firmware is it will leverage all the work that the Oculus hardware engineers have done, and the subsidized price.

Or do you believe standalone VR headsets are doomed to being closed forever?

It does suprise me that Google effectively gave up investment in VR and effectively handed the entire industry to Facebook. Though Microsoft's efforts with Windows Mixed Reality headsets haven't made a splash at all (except maybe the HP Reverb G2).

> Or do you believe standalone VR headsets are doomed to being closed forever?

I think they're doomed to being closed until quite a few patents expire.

Maybe I was too far into some denial. I retract some of my words. A rooted Quest 2 is going to be sizzling device in some ways, I admit.

How hard did google really try, though?

They convinced Lenovo to make the first standalone 6-DOF headset and a 180⁰ camera for building content to go with it.

> The original Oculus Quest is already really good, and was constantly sold out

I accidentally bought one until I realized the mistake and bought a rift S a day after that. Want to buy it? It is as good as new.

Thought about selling it on ebay but got a bad conscience to sell it to some kid, even for a good price.

Jokes aside I think not needing to connect a PC has to be the future, but I would expect the device to be able to run a conventional OS. Currently I think it will stay a dream however.

Otherwise VR devices have amazing capabilities already. A few more pixels would be good, but you will quickly need a lot of processing power.

> Jokes aside I think not needing to connect a PC has to be the future, but I would expect the device to be able to run a conventional OS.

Why, given the opposite has proven true in the mobile space?

I wish they did, but consumers clearly do not value this.

Putting aside the software for a second...

I've thought a little on this as well, from time to time. Hardware-wise, I don't see how the Quest 1/Lenovo Mirage Solo/Vive Focus were anything much more than Qualcomm's reference design headset (https://developer.qualcomm.com/hardware/snapdragon-835-vr-de...). Different strap designs, minor variations on display modules, removing ancillary stuff like the cellular radio and GPS, but otherwise the same CPU, same GPU, largely the same cameras.

With the Quest 2 running the Snapdragon XR2, a lot of the basic features of 6-DOF VR, with controllers/passthrough/voice control/etc, could be running straight on chip. (https://www.thundercomm.com/app_en/product/1584417287281447)

Qualcomm provides an SDK for VR on their hardware that touts all the things Facebook has always claimed were their own big innovations, like asynchronous time warp and lens corrections (https://developer.qualcomm.com/software/snapdragon-vr-sdk)

I can't find the link right now, but Qualcomm/Thundercomm are also providing a base Android image specifically for use in VR. Which is probably why you can build and install regular ol' Android APKs on all of these headsets and they map the controller to mouse actions.

My point is, I am starting to doubt Facebook has all that much "special sauce". I think it's mostly just marketing. I think Qualcomm has done the hard work here and is happy to let other companies take the credit because Qualcomm is about chip licenses, not consumer electronics. Other than tuning parameters for camera positions, there should be no reason why the OS image for any one of these headsets to run on any other one.

I think it's doable. I think you could probably root the firmware and get an open source OS running on any one of these headsets. And it'd also breath new life into systems whose creators basically got bored with not making a billion dollars overnight.

I would tend to agree with you, but is's John Carmack leading the software group for it and I just cannot believe that he hasn't done a bunch of optimizations and other work.

It would just negate a reputation that he has worked his ass off his entire life to build if he hasn't done valuable targeted work on the visuals of the device.

But he did most of that work on Gear VR (Snapdragon 805 when it first came out for the Galaxy Note 4) and Go (Snapdragon 821), long before Qualcomm started sinking so much stuff into the SOC (Snapdragon 835). My understanding with his involvement on Quest 1 was that he was primarily concerned with the PC streaming. His role since November is "Consulting CTO", which I basically take to mean "on call for questions as needed". Was he even invovled with Quest 2 beyond stuff ported from Quest 1?

All I'm saying is that there is a path through which everything Carmack said could be true, and also the latest hardware doesn't really use what he built.

Yes your information definitely creates a reasonable doubt that he would have had the time and job description to do serious head-down graphical-stack type developments.

But I have a return volley for you...who on the Facebook/Oculas graphics programming team, knowing that jcarmack was your boss, wouldn't go full nuts on trying to create some slick tech to impress him with? I'm sure FB has the pick of the crop when it comes to hiring talent, and at least of few of those guys would have to be his fanboys and come up with very useful optimizations and enhancements to add to the Quest graphical stack, right?

Facebook only has the pick of engineers they can convince to move to California and work for Mark Zuckerberg. I know it's hard to believe, but there are a lot of us who have turned them down on multiple occasions.

The competency of FANG engineers in general is highly overrated. They are humans. They don't have access to some kind of forbidden knowledge. The idea that these companies are uniquely competent such that competitors don't even have a chance is largely a marketing story.

Interested! I’m really passionate about VR/AR and would love to work on something in the space. Currently an undergrad. How can I contact you?

I already have a decent amount of experience with Linux internals, but my next project is far less ambitious in scope: simply reverse-engineering the Android-based firmware on a Huawei router and creating a drop-in replacement.

There is a lot of foundational hacking to be done around the Oculus standalone headsets. I don't believe the headsets have yet been cracked to unlock the bootloader to launch unsigned firmware.

It's probably worth moving discussion over to a forum and Wiki dedicated to such reverse engineering efforts (such as XDA developers) [1]

I do think that now is the time for such a project, and that crowdfunding has changed the economics to make this possible in a way that it never has been before.

[1] https://forum.xda-developers.com/mobile-vr/oculus-quest

Ping me at my HN username at gmail with your resume and I can forward it to oculus recruiters.

I am into FreeBSD device drivers, but I would love to help out! Ping me.