Eye tracking module and Foveated Rendering


#101

My friend had an Apple IIe which was monochrome. I told him back then that the IBM clone I was using (cga) had 4 colors and there was more potential from it. He harped on about painstakingly about his monochrome mac on the quality, ease of use and the fancy interface and this clicky thing called a mouse… I didn’t care what he said because… I had 4 colors!!

I remember saying that having color was better…and that all the other stuff could be replicated on the pc eventually…then came ega, adlib…286, mouse…windows,386sx…etc… I kind of lost touch with that friend, but as far as I recall he upgraded to another apple iie or something…and nothing had really changed there…All I remember was apple lost a lot of ground. In the 90s, the mac was the worst piece of crap I had to administrate ever… Obviously they have found their feet again due to innovation…


#102

Sorry you are mistaken. All Apple II’s from the Apple II thru the Apple IIe. all supported 8 bit color to 32bit color in the GS. The text was in fact mono but the graphics supported 8bit in lo-res and 16 bit Hi-RES. The Macintosh line did not have color at all for several iterations. I worked at Apple in Addison, TX. while they were manufacturing the Apple II line. I was their Test Line Support Tech. The Lisa which was the predecessor of the Macintosh was monochrome as I recall. I still own apple II’s Having access to all of the line until they stopped making it. see https://en.wikipedia.org/wiki/Apple_II_series.
They could have dominated the market had Jobs not opted for a closed system. Much Like the Oculus Rift is a closed system. It too will fail to achieve dominate status.


#103

Ofcourse you cant compare pimax to VIVE or RIFT.

Rift=backed by facebook and funds+expertise
VIVE= HTC also got enough money+expertise

Pimax is a standalone company as far i know. And for a little company to bring more innovation then GIANTS with HUGE MONEY RESERVE…tells a lot about the dark side of that industry…

Imagine this little company Pimax would up the quality and go beyond BIG companies…now that would be crazy…

Pimax will learn and up their game i hoepf or them otherwise they will make bad name for their company worldwide, and it very hard to undo a bad image/reputation…

You see the problem with RIft and oculus is they wanna sell as much HMD as possible, so they keep the specs low so everyone can join the party…It almost like those big Game companies, who dont upgrade their graphic engine. they just use the same for years. And then you got CRYTEK, they say fuck that shit, we gonna make your computer sweat…And as your pc is sweating, NVIDIA and AMD start to get moving and put out new cards that can run that shit decent so they can draw more people and attention to their product.

Innovation is brought by those who put the bar higher and don’t care about $$$. innovation is stopped by people who care more about $$$ then raising the bar.

And VR is still in my eyes a high end item. Just like computers one day were, the same is with VR now.

It is expensive, you need a demn good pc, and it should be like this…Maybe we will see faster development in graphics cards…because at the moment, the game engine are not pushing the 3d card market as they should…VR is the new player thats gonna push NVIDIA and AMD to make better cards.

when volta is out for example, 4k60fps is gonna be easy to archieve…
But with VR we might need a next generation to get ready for 8k90fps.

Because you cannot wait for monitors to go 8k…most people still have 1080p screens or maybe 1440p max.
Why would people need 8k screen monitor…well the story changes with VR next step maybe 12 or 16k.

I predict the VR world is gonna be tied very close with the 3d card market in the future. the question is who is gonna push who?

Eye rendering imho is only good for 4k native and higher resolution VR but once you get 3d cards that can power it easy, i believe eye rendering is useless…so its usefull when 3d cards are running behind.
Sure it might stress your gpu less…but the point is when you can run it without eye rendering…you can run it …


#104

All we know today about 3D VR rendering is thanks to Valve and Oculus (forget Facebook and HTC, the former only brought investments and the latter was manufacturer). Valve and Oculus did quite a lot of research and prototyping to figure out all things we take now for granted.

For example the magical 90 FPS is something Carmack was researching 5 years ago. The optimization of signal path, drivers optimization, lenses, and the whole compositor was developed by Oculus and Valve.

The lighthouses from Valve are pretty smart idea too. And they are even improving it with v2 now. The ergonomics of the controllers is another story. All those technologies defined and formed the industry.

What Pimax did so far is only improvement on current tech (bigger displays and FOV are not really a breakthrough). I do not want to marginalize their effort, but we should pay tribute where it is due and for me it is at Valve and Oculus. Pimax is yet to prove itself.


#105

Both the Vive and Rift are very high quality products compared to the Pimax 4k. However they indeed lack innovation, I agree that they seem to try to milk out existing products rather than bringing cutting edge products like the Pimax 8k. Of course the market for such products is rather small so from a business perspective I can understand it. But I personally always choose innovation over quality, that’s why I bought the Pimax 4k in the first place.


#106

I think this is much rather a strategical question, not so much one of lack of ambition or milking profits.

The companies busy with increasing VR adaptation have to find the balance between
a) pushing every new stage of innovation as soon as it comes available, but risk that the demand on the rig to fuel such high-end HMD will be pretty much the same as it was when Rift & Vive released, i.e. the latest flaggship GPUs etc., or

b) go with a slower pace, meaning that as time goes by even middle-class rigs will have sufficient power and the HMDs themselves can be offered at a cheaper price because they do not need latest & greatest components during the later stages of their sales cycle.

If you go for a), the advantage is that you will reach a level of technology faster which makes the VR case ever more compelling; however, you will not have grown the VR user base much as it will be the same geeks with some grow due to the increasing attractivity of the swiftly evolving VR tech.

If you go for b), you will grow the VR user base much more rapidly, now common people in emerging markets come into the range of potential customers for the VR experience, and owners who just invested substantial amounts do not feel too bad because their expensive new toy is not made look old just a few months later.

I fully get the approach which sees b) to be the safer route to VR mass-adoption, I just believe that we are still at a stage where we still need 1-2 generations until we have the base level which satisfies the needs of the mass market - first and foremost sufficient resolution to use VR for desktop-like environments, i.e. the ability to read text in real life font size. So I feel that Oculus & Valve/HTC should push more until such standard has been achieved, even if this means that they may disgruntle some of the customers who are not prepared/able to afford repeated substantial investments in the fast-paced releases of improved versions.


#107

After reading this whole thread i also would like to share my 2c :wink: :slight_smile:

My background: I was following this forum and youtube reviews of the 4k for more than a year, after having bought VRX1 (around the Millennium), io-glasses (VGA, 800x600 in 2004 or so), then playing in 3d-vision with beamers for years, then buying DK1, AntVR, DK2, Vive, CV1, happy about improvements but knowing that that SDE has to go! But reading about the downsides i did decide not to buy a 4k. The (mainly software-) improvements did finally let me buy a used 4k for cheap. Unfortunately i bought a unit with wrong lenses/lensholder combo, but Pimax gave me the right combo free! Even so i bought the 4k used. The SDE was gone, but the dim backlight and the ghosting (not smearing - that was okay) was terrible. Then i removed the shutter-glasses and was much happier having a almost perfect (traded the ghosting for little smearing) picture. Still having strained eyes when using, i did a IPD-mod in moving the lenses about 4-5mm apart (my measured IPD is 69,5). Now i look into the position tracking issue and i will probably order a vive puck for positional tracking.

Pimax is a smal company (comparably very limited resources) which surprised us all (as a chinese company) in delivering constant improvements and not just try to only maximise revenue to a point where their next product really looks good! Like all what we wanted from the big two players in VR, CV2, Vive2 already right in our hands. And they did not want to give us that for whatever commercial (Steam) or legal (oculus-zenimax) reason.
So, yes, again no native 4k, but by looking at the 4k HMD it should be clear that 2x 4k native makes no sense at all right now. Almost no better picture than 1440p upscaled (i trust Pimax on that one) with about half the requirement of gfx-power. And i use a beefy system with 1080ti and i beg for even more fps and cry for every setting i have to reduce for reaching my >45fps goal.
And this brings me to the foveated rendering.
Yes, foveated rendering is not ready right now to be 100% accurate, but i highly doubt that it has to be. Our visual system accepts almost 50% „fake“ frames with ASW and nobody sees it. Also in playing with z-buffer3d only few people see the distortion around close objects moving, and if you watch a normal DVD and look just around a foreground object like a person moving, you suddenly see a halo of distortion around that object.
FOVE would not have produced (and sold) a HMD if the hardware wouln‘t have been good enough. The software base was not ready to adopt, so the time was not ready because of the only little performance improvement because of the narrow FOV and the not so perfect tracking.

BUT i definately think that NOW is the right time and Pimax has the power to change that. Because with a 200FOV you suddenly do not have a 75%-100% performance gain (100FOV), but THREE to SIX times as much (my personal estimate looking at the additional screen percentage). And this makes a BIG difference! Because finally it is worth (or you even need) to use the software tools already here for the needed improved performace.

So, looking in the witch-glass-orb in front of me, the other companies will have to match the 200FOV and this will bring developers in the boat to use Nvidia VRWorks and LiquidVR tricks to finally make foveated rendering an obligation for fluid VR. This is why we need VorpX to jump in, because hardware without a software base means nothing. And the more people see how amazing even 2D games could look, the more people will consider to buy. Yes, all VR is still a niche, but if u can play all the old games you have in VorpX-3d-VR, more people will buy and enjoy that. Maybe even replay old titles like i do.

It is time for a big change and Pimax is right now a big mover. In forcing software to adopt to hardware. So please bring that module to us when it is ready. Every sold pimax with this module will change VR for the better.
For me wireless is a far bigger leap and will not happen before eyetracking with foveated rendering is used within high end hmds, trickling down to mainstream headsets. I think that the second is a requirement for the first and will make it happen that with the bandwith of a tp-cast we can enjoy our 8k wirelessly.

I backed the 8kx and a 8k only to show my support to pimax and honoring their passion for development. I think that if foveated rendering is pushed, every future high resolution headset is earlier in our hands!

Thank you for reading!

Edited for better reading


#108

Probably, I was about 8 or 9 yo at the time…Ah had a look at the wiki,…All I know is that i was something like 1984 and it was only 1 color, perhaps it was just his monitor or an option…I don’t know… I was vague on his exact specs coz I didnt spend long on it…and was…well a kid.


#109

I was in my mid-late 20’s at that time. Yea, I’m an old guy. You are right about the early macs being a nightmare for admins.


#110

@matthew.xu @PimaxVR {edited formatting)

Have you all seen AdHawk’s eye tracking solution? It look’s seriously impressive. It’s said to be much faster, cheaper, smaller, and more accurate.

Here’s a recent tom’s hardware article about it. http://www.tomshardware.com/news/intel-invests-4.6m-adhawk-eye-tracking,35737.html

The article says "HMD makers can request a device and an evaluation kit by contacting the company directly at http://www.adhawkmicrosystems.com/contact "

in the article they say it can be $10 per eye (depending on situation)

Eye tracking is very promising. If it was implemented well, it could make pimax 8kX widely available to many people with much less powerful computers due to performance savings of foveated rendering (ideally).


#111

Hi @Taco1,
Thanks for your reminding. We already know this. We are confirming relevant information, thank you.


#112

Thanks @Taco1, I have sent mail to them, looking forward to hear from them. Thanks again.


#113

Eye tracking will require USB, is there a USB port on the inside of the HMD? Or is there a hole to route the USB cable to the outside?


#114

Yes…the HMD has 2 usb ports…


#115

Yes, but the eye tracker sits in the interior of the HMD next to the lens. My question is, how will the eye tracker reach the USB ports on the exterior of the HMD?


#116

I don’t have a clue. I hope Pimax have an idea…


#117

Hello
i have a question
can the Eye Tracking module be installed on the Prescription Frame?
@bacon @matthew.xu @PimaxVR


#118

Yes sure, because they are separate modules.


#119

Really ? simultaneously? @matthew.xu


#120

Yes, as far as I know.