Credit where credit is due: https://github.com/relativty/Relativty
He basically built (and advanced) Relativity's open source headset.
Credit where credit is due: https://github.com/relativty/Relativty
He basically built (and advanced) Relativity's open source headset.
It's always cool to see such projects, and I was discussing about the topic just yesterday.
... but please, don't spend $150 on that. This is limited has it has :
so it's basically a more open but not standalone version of the Oculus Go. For context the Go is from 2018 and back then was $200 while being standalone. Note also that the Go is rootable, cf https://developers.meta.com/horizon/blog/unlocking-oculus-go/
There are also other DIY VR HMDs, e.g. https://github.com/relativty/Relativty from 2020 which similar limitations. In fact this made so much buzz back then the founder managed to ride the hype and make (sadly) a VC funded startup. I say sadly because the initial project was all open VR and openness but once the money was locked-in... well I let you check.
So... again this is VERY cool to build but please do not consider this anything but a way to learn. If you do want to play with VR with a limited budget consider instead an accountless (meaning no Meta involved) second hand Meta 2 then Alvr or WiVRn (cf https://lvra.gitlab.io/ for more).
This is so damn cool I want to throw one together just to tinker with it.
This has been the best year for VR as a whole.
This has been the best year for VR as a whole.
am a fan and vr dev... but this hasn't been the greatest year personally. what's made it a highlight for you? perhaps another perspective will help me!
Linux support has made some decent strides, particularly with the open source monado / wivrn + xrizer stack (steamvr / alvr still have their issues but are getting better).
I'm quite curious to see how quest 3 works with godot & openxr, apparently they have a build you can run on android inside the quest 3 and do vr-dev in vr.
Indeed, I made a video about that but unfortunately in French https://video.benetou.fr/w/g111LSRGfj6HmNxa91au3q
but unfortunately in French
nay, merci boucoup, my 9th grade french will make due! or the auto translate which is always fun!
Feel free to ask questions but you can also seek to https://video.benetou.fr/w/g111LSRGfj6HmNxa91au3q?start=35m50s directly which is the part with the live, in HMD, edition.
As another user said Linux compatibility has been huge the past couple of years.
And PSVR2 has had some amazing games out that have completely changed the way I experience VR. And they also opened up their headset to PC via steam with an external dongle.
Mainly the accessibility has been at an all time high. But the games that have come out lately too are amazing. I can even do cross play multiplayer with many different headsets now which is also super cool.
Well, for one, some YouTubers made a cheap, open source VR headset. That was pretty neat.
I am wondering about the software side. Is it just a glorified monitor for your face, or does it have some kind of internal control/OS like a Quest or other commercial headset?
It seems to only have an Arduino as processor, so presumably the former i.e. you need a PC to plug this into. It seems to have highly simplified tracking compared to something like a Quest - a single Inertial Measurement Unit. All enough for sim racing which is this guy's jam, but I wonder how well the tracking stays calibrated.
The critical thing with these is response time. If it's even slightly too high (I think 20-30ms is easily too high), some/many people get very motion sick. Getting that time down as low as needed is also not trivial.
With it only being 60 Hz on the controller itself, that's basically impossible to hit. That's 16.6 ms already. Then the processing, sending to the PC, and the PC reacting has a budget of just a few ms? Yea, not happening.
I'm assuming he's really not sensitive to this. As it's open source now the people who are sensitive can improve it. That's the beauty of open source after all.
Normal video games can make me motion sick so I can only imagine how bad this might be! Maybe in the future it'll get better!
I bet I could use it. I can play the VirtualBoy in a moving car. If there's any kind of way to tie it into MechWarrior, I want to build one.
I’m wondering if I wanted to fly a jet, would I need another sensor for 3 dimensional tracking
Looks like it has an inertial sensor to control the view, here's there relevant bit from the article:
The parts he purchased online include two displays that max out at 2880×1440p and 90 Hz refresh rate, two lenses, an IMU (Inertial Measurement Unit) sensor, and an Arduino microcontroller board. The other parts of the headset were 3D-printed.
...
The headset features individually adjustable IPD, interchangeable faceplates, and head tracking. But it does have downsides — at full resolution, it only runs at 60 Hz. Also, it only has three degrees of freedom (3DoF), which means it tracks looking up and down, left and right, and tilting the head left and right. He explained he didn't bother equipping it with 6DoF as 3DoF was enough for his sim racing.
It relies on the Relativity VR driver so theoretically similar software support, namely https://github.com/relativty/Relativty#14-software-setup but mind you beside being 5 years old this is 3DoF, no controller or hand tracking, whereas the Quest is 6DoF, namely you can move around, not just turn your head.
Oh FFS stop designing things I can and want to build
Huh, this is kinda prefect for me too, only the refresh rate might be an issue.
I wonder if I could buy cheap no-name headset & fit in that arduino chippy.
Because you pay with data. Quest is the first device, where meta has full control over the Operation System, and even beyond.
You are forced to have a meta account to use it, IRCC
You did but not anymore :
So yes, by default you are paying with data. In fact IMHO if possible one should not rely on Meta hardware. That being said if you get e.g. a 2nd hand Quest 2 or 3 and use it without an account then you might be providing little to no money to Meta and no data. It's not trivial but it's feasible. Arguably it's even easy for somebody who seriously consider such an endeavor of assembling their own HMD.
PS: Meta has access to the whole device but... they are not owning the OS itself, it's still an Android device. The OS is very much driven by Google. In fact it's quite interesting to consider that Meta failed to develop their own OS and that Google is shipping soon AndroidXR.
I wonder how this compares to Mañolo's HadesVR/PersephoneVR setup... I think they have 6DoF and use lights embedded in golf balls to act as PS Move sensors
Overall is that even a deal over a used headset? Even a fully featured non-stripped down one? Like given what features his headset does have, it's comparable to some pretty old headsets... and it likely does even those bare minimum features more poorly than an older used headset would. Not to mention comfort.
Like a 10 year old Rift CV1 has almost as much resolution at 90hz/fps instead of 60. And while it's lenses would be relatively terrible now, they were pretty much the best option of their day, and likely still better than whatever this dude sourced. Not to mention their motion to photon was around 12 ms. The absolute best result this guy can hope for is 16.6ms, and that's only if everything else in the pipeline is faster than the screens refresh rate... maybe it is... but I wouldn't bet on it personally.
I'm sure it was a fun project though.
No, it's not a deal. It's only about learning IMHO and that's already be great.
In fact it's such a "bad" deal that I replied a bit in depth. My goal isn't to bring the project down, heck I'm even envious about being able to do all that, but rather that people curious about VR don't get confused. This is nearly a decade outdated and there are better open solutions, e.g. Lynx or even Meta Quest 2 without account IMHO.
all 60 comments