Hacker News new | past | comments | ask | show | jobs | submit login
Simula One vs. Quest 2 vs. Valve Index (simulavr.com)
50 points by gurjeet on Dec 27, 2021 | hide | past | favorite | 72 comments



I am curious to see where this goes. VR for productivity has a lot of uncaptured upside. Putting a whole Linux OS on it seems to be a great idea.

I am surprised that Apple or Google are not in the VR space as one company invented the iPhone and the other spearheaded Android. Both are relevant forms of experiences that could help in building a VR headset suited for productivity.


Apple is very much working on a product in this space but they, as usual, aren’t showing anything until they’re ready to.

One of my friends recently took a position at Apple to specifically work on this hardware and I get the impression they’re hoping to ship sooner rather than later.


They have so much potential to release something amazing.


Co-founder of Simula here. Agreed. In many ways it makes zero sense that our micro team of 3 people with extremely limited resources is making the first (at least Linux) VR Computer (VRC). This is something that Big Tech companies should be doing in a well-functioning economy, but for some reason aren't, due to some combination of bureaucratic inertia and risk-aversion.

Every year Apple is one year away from releasing its supposed AR headset. Google basically abandoned Daydream (which was gaming/entertainment focused anyway). Microsoft had a great early Hololens prototype, but doesn't seem to be pushing consumer facing AR/VR at all (instead selling contracts to the military??).

Meta is investing heavily in VR (which is cool), but is unfortunately jamming it through the awkward framing of "make everything in VR a social network", which we're not fans of: https://simulavr.com/blog/we-dont-want-the-metaverse/

In any event, we're really excited to help bring a Linux VRC to market that is focused primarily on office work, programmer productivity, and general purpose computing. =]


Apple has (supposedly) been working on VR/AR tech for quite some time, but the release date keeps getting pushed back. Latest reports suggest a 2022 launch for their VR headset.

VR is in a very good place, but it still has a lot of jank. I imagine Apple is working through as much of that as possible before releasing anything.


Google has made a headset with Lenovo, but cancelled it: https://www.lenovo.com/gb/en/smart-devices/virtual-reality/l...


Maybe it's a silly question, but what happened with the attempts of using a smartphone instead of a whole new hardware? Isn't it easier to just create a helmet with some extra sensors and connect it to the smartphone?


It's not a silly question. Based on the experiences that were made with GearVR, it turned out that strapping a phone to your headset before starting to play increases the barrier and makes it significantly less convenient. And, convenience is king.

Furthermore, there are a lot of different smartphones out there with different performance attributes. This makes is much harder to make a compelling VR experience, because you trend towards the lowest common denominator. Another point is that while yes, a smartphone has almost the same components as a VR headset, you can optimize the components and space much better if you are using it for a specific use case (which is as a VR headset) and not a general purpose phone. Finally, it's much easier to keep your VR headset charged and ready to go compared to relying on your phone to jump into VR.


I just answered this question on another article yesterday. Here was my comment:

Oculus already tried the phone route and abandoned it because of the inherent limitations of not controlling the whole hardware stack. Let's look at first mobile Oculus headset, the Oculus GO which was binary compatible with content made for GearVR;

- Performed better than GearVR with a far slower SoC than the Galaxy handsets because the thermal solution could be designed around the headset.

- Cost $200 vs the $1000 for a Galaxy handset + GearVR

- Was more comfortable with higher weight because of better distribution across the headset instead of being concentrated in the front where the phone attaches.

- The display was sharper with a lower resolution than GearVR, because you design the panel differently for headset than a handset


>Performed better

Yes, performance would be tied to how good your phone's SoC is.

>Cost

The idea is that everyone already has a phone. You don't need to include it in the price.

>More comfortable

You can put the phone on the back as a counter balance or just in your pocket.

>Worse displays

You don't need to use the phone as your display.


You already need a compute unit in the headset to do the sensor fusion / tracking. The chip in my smartphone can't handle the number of cameras needed. I guess is that the cost of upgrading that chip to one that can support playing the VR content is worth it.

I'm not sure what part of having a separate smartphone as a compute unit makes it any easier. It might even be harder to do.


Interesting. Being a portable headset, I assumed the Quest 2 would be rather low-resolution. That is not the case after all. I wonder how often the games it runs actually run at the full resolution.


I’m not entirely sure how many things truly render at the native res vs up-scaling but the display resolution is important regardless of any upscaling that takes place to reduce/remove the screen door effect.

Graphically the native-quest stuff is relatively basic to run well on the hardware but for the kinds of things I enjoy most it doesn’t really matter/bother me. I use Virtual Desktops for PCVR on the quest 2 and in some cases I think the fancier graphics are worse. I prefer best saber natively on the quest because it has way less effects that reduce readability.

They are also using/pitching ASW as a method to effectively double frame rate so they can up the fidelity and maintain a high frame rate. I haven’t played anything that does this yet as far as I know but I’m a little skeptical on this technique.


I'm tempted to give it a try. I haven't used my Rift Touch in years because of how damned inconvenient it is to get running and how difficult using it with glasses can be.


I have snap-in prescription lenses for my Quest 2, but before I got them I just wore it over my glasses and that was tolerably OK. I would still recommend getting the ($70) snap-in lenses, though.

It is pretty convenient to get going, at least for onboard apps - I haven't tried the PCVR support. You pretty much just put it on and you're off and running.


Do you actually need to wear glasses? I guess maybe it depends on your prescription.

To me it seems for those who are nearsighted, wouldn't need glasses as the screen is very close to your face.


Physically yes, optically no. I don't know exactly how the hardware does it, but objects at a given virtual distance are at that focal distance, too. (Without the illusion of depth, there'd be no immersion...) If that's further than your uncorrected vision can focus, then you need correction to prevent those objects looking blurry, just as you would in real life.

It depends somewhat on what you're doing whether this poses a significant problem, I suppose, but I have found the experience more enjoyable for being able to see it without having to fuss with fitting my glasses under the headset.


If you connect it to a PC with a decent GPU..all the time :).

I got one for the boy a year ago (last christmas present) and while he uses it a lot I use it more connected to my PC for elite dangerous etc.

The extra resolution over the Rift S I had/have is well worth it (text is readable on ship displays on the Quest 2, not on the Rift S) - for the price they are an excellent bit of kit.


Keep in mind that display panel resolution and "game" resolution are not very similar in VR headsets. The lenses in most headsets only show a roughly circular area of the display, and the pixels in the center of the display end up much larger than the pixels at the edges due to lens distortion.

You often need 2-4x more panel resolution than the effective resolution to the viewer, just to deal with those factors (though this does vary wildly with lens and panel design).


Looks like there's more competition in the portable headset sphere now, this [0] micro-oled VR headset is already on kickstarter.

With that kind of resolution (5120x2560) and lightness (200g), I'd be VERY interested in trying it as a virtual desktop with multiple screens.

https://www.kickstarter.com/projects/969477673/arpara-worlds...


Sadly at this point I consider anything that isn't already in the hands of consumers to be vaporware. There is little to no guarantee that funding Kickstarters results in tangible products. Typically, they just lie. It's even comical, as highlighted by YouTube series "Kickstarter crap" by idubbz.

I wish we had a better system wherein the facilitator of funding, Kickstarter, would require an action plan, goals, dates, and skeptically reviewed the progress and required the claims to be met. Along with limiting initial funding and splitting it into funding rounds based on delivery of the goal plans and proof of tangibility. Or whatever.


I had the same reaction, however, there are a lot of review-videos of the production units on their Kickstarter and it looks like it's already in mass-production.

I'm not going to back this one though, I'd rather wait for Valve to release a pancake-lens oled headset with proper audio (this one's audio is crap, according to reviews).


I understand this viewpoint, but some counter points:

1. We originally tried raising money for this from VCs/angels, and got turned down by everyone. We also tried partnering with other headset manufacturers to help us put Simula on their headsets, and also got turned down/ignored by...everyone. So Kickstarter has allowed us to go direct to consumers to help fund it, which is a pretty cool use of the platform.

2. We're actually going to ship our headsets, leaving an extremely clear paper trail of weekly updates behind us for the duration of the whole project.

I know, I know..you can't actually know this for sure just by me claiming it. But basically: we have been working on Simula for years, putting in tons and tons of sweat to get the stable Linux VR compositor we have now. Also: our software is already testable right now on older headsets like the Valve Index and HTC Vive Pro (https://github.com/SimulaVR/Simula). We've staked our reputations and life's work into this, and are not going to let it go to waste by failing to ship our headsets.

With that said, our upcoming Kickstarter is only the first step in this process. We already have plans for a second headset iteration after this one, which we will start working on after these ship. If you wait for that iteration, you'll have the proof of us shipping the first one, and won't have to wait as long for us to bootstrap the manufacturing process on top of a pre-order campaign =]


I've seen and worked with a fare share of projects that base the mass production in China in the past 20+ years, this is what you can expect:

- Sourcing of electronic components vary from day-to-day currently, expect quotes to go from 1x to 5x the actual price.

- Plastic moulding is expensive [0], you need injection moulds for every part of the headset and each one of these moulds range from $2K to $20K per part depending entirely on the factory, place of manufacture and the business of choice. Also, use a factory that have frequently do new moulds, because the differences in pressure and precision is something that needs to be dialed in, doing your own moulds and then ship them to China is risky due to differences in implementation methods.

- If you do assembly in China, expect every single part to be copied and sold on the Chinese market and to wholesalers that ship out of China, probably even before your kickstarter backers have gotten their initial units. This is even more true for the low-cost factories, which cover their invested costs by selling your IP to larger factories.

- Shipping out of China for packages over 3KG is currently very expensive, especially during Corona where shipping by water can take +6 months at the very least. Shipping by Air is even more expensive because prices have exploded 3-5x depending on carrier, but you can ship out in a couple of weeks (depending on volume of goods). Also take into consideration distribution in the US and the EU, both vary greatly and stock can be outright stolen (happened twice for a client that did a Kickstarter with distribution in US).

[0] https://www.plastopialtd.com/pricing-guide/


- We plan to do the MP in Germany or the US, so most doesn't apply. Especially for a first product I do not want to have to deal with a language/culture barrier on top of everything else. - Yep, we have the mold costs factored into our BOM estimates. - Re; distribution, we plan to ship out of our own warehouses directly. DHL Express prices are fairly reasonable if you have a large volume.


I hope my message was deterring, it would be awesome if you succeeded, but it would also be horrible if the effort was wasted and you became bankrupt at the same time. Your plan sounds very doable, especially if you source within EU countries and assemble in-house and use your own warehouse. Best of luck :)


Sorry, it should be "I hope my message WASN'T deterring", auto-correct :-D


> The FOV is 95° at a pupil distance of about 8mm and 90° at a pupil distance of 12mm.

Meh, pass. The Index is at 130 degrees and even that looks like you're looking through a pair of toilet paper rolls.

Why do they bother going with such high resolution if they don't expand the FoV? It's just gonna chug more with your average mid range graphics card.


>The Index is at 130 degrees and even that looks like you're looking through a pair of toilet paper rolls.

Not even close. The binocular overlap is >0 degrees. It's more like wearing ski goggles.

>Why do they bother going with such high resolution if they don't expand the FoV?

Because resolution in VR is a big problem since it's being stretched over your whole view. This means that in order to display a virtual 1080p monitor you are going to need a lot more resolution than 1080p to be able to do that. As you starting adding more degrees of FoV you start needing to render more and more pixels due to the distortion.


I exaggerate, but it's still nowhere near the usual 210 natural FoV. That's just how it feels after a while in these first and second gen headsets.

> As you starting adding more degrees of FoV you start needing to render more and more pixels due to the distortion

Exactly, and since they've added no FoV (the original 6 year old oculus had an 80-90 deg FoV, albeit with a shitty 1200x1200 resolution) the extra density seems a bit gratuitous.


Can’t you make your lens so that you have more pixels in the center of vision, and less on the outer edges?


That's exactly what we do to achieve our 36 PPD figure.


Isn't that a problem since you can still move your eyes?


It's accounted for. Eye rotation is not very severe; 86% of saccades are within 15 degrees.


Worth to note is the display is only 70Hz at 2560p.

The "95% light path amplification" for their pancake lenses is also either misleading or an impossibly high transmission figure for pancake optics. Maybe I just misinterpreted this though.


Remember that oculus started out on Kickstarter, too. They were sold to Facebook shortly after and are now a privacy nightmare. What a waste of crowd-sourced funds to fill the pockets of a greedy founder.

I won't support any crowdfunding projects after that.


Cool stuff (even though my post might come across as critical), but I will wait for high resolution + high Hz. For example, on the Simula One the Hz is low (90 Hz) but the resolution is high. On the Valve Index, the opposite: the resolution is low, but the Hz is 144 Hz. I won't use/comment on Quest 2, as Facebook requirement is a dealbreaker for me, so the entire product is irrelevant in a comparison. I totally get that such a product is going to be expensive, and that its gonna take a while till it exists, but if I am going to shell out several thousands of EUR for a device like this, I want it to be really good (and breaking my privacy I am not going to fall for regardless of how good the price/performance might be. Not with AR/VR. Honestly, it scares me shitless to see Meta going that path).


I always use my Index in 90 Hz because 144 Hz is hard to drive. You have to generate a frame for both eyes in under 7ms.


I've tried SimulaVR and it just crashes on me, constantly. RTX3090, Valve Index, high end CPU, latest drivers. What's the point of desktop productivity software that just blows up all the time?

Perhaps instead of spending time developing and bringing to market new hardware, fix your software first, there are plenty of hardware vendors already in the market but nobody that I'm aware of is taking Desktop Productivity in VR seriously (yet).


What crashes are you getting? If you post an issue in our Github Issues or Discord, we can help out.


I suggest posting about the issue either in GitHub or their me Discord, the dev team is very helpful and communicative.


What are you guys doing for text entry? Obviously a keyboard is tough to beat, and I think using this device at a desk is probably going to be a popular use case, but it can be tricky. Will it be able to automatically track the position of your keyboard? Could some solution such as putting tracking stickers on your keyboard be helpful? What about those of us with unusually shaped keyboards?


M/KB for 2D windows with AR passthrough should handle it I think. I'll think about some solution for people that want AR passthrough turned off.


I worry about comfort and heat. As this is intended to be used for work it will need to be comfortable to wear for hours end. To-date there hasn't been a headset that meets that criteria. The Go/Quest/Quest 2 are especially bad in this regard; despite well designed thermal solutions that radiate the heat out through the front of the headset you often find yourself reduced to a dripping pile of sweat.


We're taking both comfort and heat management into account for Simula's design (note also: the compute pack driving our headset is detachable). We are planning on writing a blog post about this at some point in the coming weeks.

With that said, we will definitely look back in 5-10 years in horror at how large and bulky our VR headsets were. Keeping weight down will be a critical factor in driving up "all-day" usage from knowledge workers.


Curious comparison, there is the pico 3 and the vive focus 3 standalone vr glasses which seem to be closer to this type of glasses.

I also wonder how they are cooling and powering this cpu. The vive focus 3 has a noisy fan and heats up pretty quickly. This ones seems much more powerful (without deep research) and still planned to be in the goggles.


We're looking to mount the compute unit at the back of the head, which gives both more space for heat dissipation and allows for a balanced weight.


Oh wow, that IPD range actually includes me!


Nice! One thing we forgot to mention is that the IPD adjustment is per-eye, which is helpful on faces which have non-symmetric IPD measurements (like mine..).


Cow or Orangutan?


I'm at 77 on the nose. Well, a bit above my nose buy you understand.


I see there is no final weight yet. With an i7 you need decent cooling and larger batteries, which adds more weight. This thing looks like it has the potential to be really cool, but if it's so heavy that it causes neck pain, it's not gonna be super useful.


We don't have final numbers for our weight, but will make an announcement when we do. (Note our compute pack is detachable though, which can help mitigate weight.)

Agreed RE neck pain though. Our goal is to make a VR Computer (VRC) so good that it completely replaces laptops/PCs, but we can't do that if the headset is too heavy.


That's great to hear! I look forward to more news about your product. I'm super excited about this!


I just want to install Simula OS on the quest 2. Quest is the only affordable option for me.


The Quest doesn't support Linux, so we had to make our own headset.


Yeah it was wishful thinking. I hope oculus adds support.


Beside games and maybe working with 3D animation and CAD software, what kind of uses do you have for your VR headset, if you have one?

(I can see a number of uses for an AR headset, but it's an entirely different beast.)


I wanted to use my Index for virtual desktop, even watching movies, but resolution is low, really low.

I know these things will get smaller and have higher resolutions, no cables, etc. But I am really not looking forward to having an antenna strapped to my head all day long.

It's obvious that many young people don't have a problem with that, since bt headphones seem incredibly popular and well on their way to replace the cords, but I am not like that


It's a bit weird to me that they're going with intel, is the new i7 better than ryzen processors if you count in energy efficienty (and thermal output)?


Really simple reason: Intel talked to us, AMD went "lol here's a bunch of embedded boards, go fuck yourself."

Intel also has the NUC compute element solution which significantly simplifies the EE and allows for a seamless upgrade to e.g. Alder Lake.


What's EE?


Electrical engineering. Getting a working PCB design for a modern CPU, especially in tight space constraints, is quite challenging.

It's probably something we will do in the v2 (the NUC solution has its drawbacks in terms of exposed IO and such), but I'm glad we don't have to do a full design immediately; only the carrier board that routes the NUC compute element to whereever we need it.


For a device that's apparently going to be going on sale in days, it's odd how many of the specs and features are "TBD"


Stuff like weight/dimensions are pointless until we have the integration finished, and we don't yet.

We're approaching the limit of what we can do with a 3 person team, and we need to place orders for our components so the insane lead times at least result in a product that's soon enough. We can't do that without both funding and an expectation of how many units we'll ship.

Hand tracking is TBD because ultraleap still hasn't V5 Linux support, so we might need to pivot to a different solution. We'll definitely have some form.

Eye tracking is the one feature we're still undecided on, but right now we're tending towards yes. It allows us to have automatic IPD adjustment which is a requirement for our class.

SLAM for this kind of headset is just plain difficult, so we need to both evaluate different solutions (see previous blog post) and hire more people if we end up with a custom solution. Backup plan is to just use a module but it's less preferable.


> Eye tracking is the one feature we're still undecided on, but right now we're tending towards yes. It allows us to have automatic IPD adjustment which is a requirement for our class.

You can leverage front-facing cameras for IPD adjustment – look at the headset before wearing it, boom, your IPD is correctly set.

Why is SLAM difficult for this kind of headset?


Outside and dark environments, primarily. A stereo RGB camera approach isn't satisfactory when the conditions are less than ideal. I've also talked with people working in the space and they mentioned that a custom-integrated SLAM (i.e. not a prepackaged all-in-one module) is a big time-sink of corner cases if you want high-quality tracking, which is why "just use a module" is our fallback if we run into a complexity explosion.

Front-facing cameras, hmm. I'm not sure I like it from a usability POV, but it's something to consider. I think going the extra mile and including internal cameras for a seamless UX is better, though.


It's market for "office work" all over, does it mean that it cannot be used for games? I don't understand...


The headset has no speakers, microphone, eye tracking, or face tracking. These are must haves for the next headset I get.


Where are these high resolution displays from?

I'd imagine Facebook or any other big company would've been able to get their hands on them.

Or are they waiting for the next generation of headsets to use them?

By specs, those seem great. Resolution is one of the most important things in VR goggles.


The displays are a semicustom Sharp part.


Quest 2 costs $300, this one is 2k+.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: