Hi res, hi FPS VR focus in year 2021

My disappointment was large enough that I feel I need to wait for a bigger leap. I bought the rift-s fully aware of its limitation and with the goal of passing it on, I just wanted to try it to see if I wanted to spend the big bucks on a pimax 5k or 8k and computer upgrades needed to actually run that. I didnt :).

So now Ill wait for eye tracking to enable foveated rendering and dynamic focus, hopefully in combination with higher resolutions and wider fields of view. I have little doubt it will come, as it already mostly exists in the professional market, just need a consumer price tag and proper software support. I can wait a few years.

/OT

So now Ill wait for eye tracking to enable foveated rendering and dynamic focus, hopefully in combination with higher resolutions and wider fields of view. I have little doubt it will come, as it already mostly exists in the professional market, just need a consumer price tag and proper software support. I can wait a few years.

More people waiting: I knew it! :smartass:

There are various and separate technologies in use that will conform the perfect headset, but it is still a technology in development in honor of truth.

It is very subjective. As stated my VR is ready for DCS moment happened with the release of the Reverb & Index. Either IMHO were good enough visually to properly enjoy DCS. Of course there will be those who happily use the original Rift and others for whom VR isn’t ready yet. Hard to advise other than relate own experience

Just about any other flight sim ever? I only fly one other, which is a soaring simulator. To give you an idea:

 

 

another one with sea and trees:

 

https://www.youtube.com/watch?v=rb2m5BGwOt0

 

No, its not the greatest looking sim ever, its developed by one full time and one part time dev, and their focus is on flight modelling and atmospheric/thermal/slope wind/wave modelling more than eye candy, but it look ok, viewing distance is in the same ballpark as DCS and you can play it in VR at >90 FPS with a laptop with a gpu like a 1050.. Without VR you can play it on literally anything you could possibly buy the past 5 if not 10 years, including laptops with integrated graphics. I get like 300 FPS in VR on my system.

 

And btw, you can also play it online with 64 players on a server without crazy lag or warping.

That's not even funny. You cannot compare these.

How about a more real world test to where ED wants to go with the sim.

How about X-Sim with mid to high pc specs and high settings (ortho4xp). New Vulkan beta (yes Vulkan) 23 to 30 fps. This is also with better weather (less fps etc) which ED will also be doing.

Remember for the reverb 2160 x 2160 per eye!!

3 screen example is only

1920x1080 x 3 => 5760x1080 resolution = 6,220,800 pixels generated

My disappointment was large enough that I feel I need to wait for a bigger leap. I bought the rift-s fully aware of its limitation and with the goal of passing it on, I just wanted to try it to see if I wanted to spend the big bucks on a pimax 5k or 8k and computer upgrades needed to actually run that. I didnt .

 

So now Ill wait for eye tracking to enable foveated rendering and dynamic focus, hopefully in combination with higher resolutions and wider fields of view. I have little doubt it will come, as it already mostly exists in the professional market, just need a consumer price tag and proper software support. I can wait a few years.

 

/OT

eye tracking to enable foveated rendering and dynamic focus

Exactly what I've been saying, it's not just on ED to make VR move forward and be more practical on a half decent VR pc. As seen in the X-sim video.

Just get this.

http://www.varjo.com/

That's not even funny. You cannot compare these.

But you can and I do. Its not about comparing the sims, its about comparing overhead of graphics and terrain engines. I get vastly better frame rates in condor while rendering a large chunk of the Alps, with terrain, shadows, 100K trees and half decent clouds than I get in DCS rendering nothing but flat water and blue sky. If it takes DCS almost 5ms to render an "empty" frame, how do you expect it to ever manage to render a "full" frame in VR above a city in 11ms?

How about X-Sim

Now that is harder to compare. First of all I dont have it, but what Ive seen it renders so many more buildings, far more detailed trees, crisper ground textures and it has volumetric fog and clouds. And best I can tell, it can still manage playable framerates in VR.

Has anyone try this sets of tricks?:

It may seems some reduce critical aspects for improving FPS on VR and others may improve VR on windows and the DCS executable.

But maybe all of them are placebo.

I´m reading this thread in case it clarifies something: https://forums.eagle.ru/showthread.php?t=215373

When they fix it?

You mean make it more optimized where possible.

DCS was never designed around vr like the small vr games in the vr store, no or little ai, missile logic, weapon logic, ballistic logic etc.

Even the other sim with vr have little going on compared to dcs.

It will always be a bit of a balance for vr, ED will do there best and the vr tech needs to improve also to allow fo very large dynamic worlds.

Dcs is by far the largest most dynamic and straining vr game out there, so it will most likely use and eat up the best tech you can throw at it, now and over the coming years as vr displays improve even more. Hopefully the hardware can take some of the load off in the future and not just rely on dev's constantly chasing there tail. Things like sli would make sense for vr etc. Run two 2080ti's one per eye. That sounds expensive lol.

Sorry mate, but your comment has no base, you don't even know that VR doesnt support SLI/Crossfire... Do your research before spreading "fake myths".

Just get this.

http://www.varjo.com/

Its only 5000 euro. And still only offers 90 degree FoV and no dramatic higher resolution, and Im not even sure foveated rendering is already supported, let alone dynamic focus. Oh and no PC on the planet will actually run that with DCS.

Its coming, but we arent there yet.

Sorry mate, but your comment has no base, you don't even know that VR doesnt support SLI/Crossfire... Do your research before spreading "fake myths".

I wrote four words on google to find (I'm cheating, I knew it exists) that SLI works with VR already (Pro line, not consummer) - you are fighting fake "fake myths".

From Nvidia website/Pny website :

"...even VR-SLI to assign a board to each eye, allow NVIDIA Quadro RTX to offer VR solutions unattainable elsewhere."

"Accelerate virtual reality renderings with VR SLI"

"...even VR-SLI to assign a board to each eye, allow NVIDIA Quadro RTX to offer VR solutions unattainable elsewhere.""

Kewl. So to complement that 5000 euro VR goggle, all we need is 2 $5500 quadro RTXs and assuming DCS even supports that, if we are lucky we may get 45 FPS.

Or, you can wait a few years, and spend a similar amount of money to fly an actual mig 29 almost to the edge of space. Or do simulated air to air combat in two actual L39s.

No screendoor effect there.

Not a difficult choice for me.

I've not even put an eye on the "price" section...

Its only 5000 euro. And still only offers 90 degree FoV and no dramatic higher resolution, and Im not even sure foveated rendering is already supported, let alone dynamic focus. Oh and no PC on the planet will actually run that with DCS.

 

Its coming, but we arent there yet.

Well: Varjo have cooperated with DCS very recently:

And there is the Xtal thing too (maybe the best mix of Pimax with Project VROne):

And maybe Varjo doesn´t work without a Quadro (I don´t know, but it is probable).

You will confront problems as a user with all those headsets. Non of then are prepared for the mainstream user.

Let´s see if Samsung can bring something good to the scene but, for the moment: Hp Reverb G2 are our safest bet. (but a 45 FPS one).

There was still a reason why they have and spent mega dollars developing sli for VR use right. Good clear military and civi training etc.

So you can have max performance with x hardware that can and does utilize 2 x Quadro RTX 8000, 48G. At 10K AUD, it's cheap compared to the full simulators they build etc.

Consumers need to wait for better hardware to help offload.

"eye tracking to enable foveated rendering and dynamic focus"

I'm also sure ED will do everything possible to help.

The way I look at it is trying to run 3 screens at 1440p @ 90 fps etc

No gpu can do that in a sim, even with the latest sim engines and our high end consumer cards.

Consumers need to wait for better hardware to help offload.

Nope. ED need to make their engine (far) more efficient. It cant even run smoothly at sustained 90+ FPS for most people on on flat panels. If they dont fix that, nothing can fix it for them. My future ideal VR goggles will most definitely require more, not less performance. Foveated rendering will allow the system requirements to be closer to current VR goggles, rather than just multiplying the requirements many times over, but its still gonna be harder. because you simply need a bazillion pixels to get "retina" resolution and decent FoV. Im not going to wait years to get eye tracking only to have that further degrade the image quality outside my focus compared to todays VR sets!

The way I look at it is trying to run 3 screens at 1440p @ 90 fps etc

No gpu can do that in a sim,

I have 3x 1080p and my humble 1070 can get >200 FPS in a flight sim, i feel confident Id get 90 with 3x1440p. Just not in DCS.

Tried this condor thingy, but the planes lack plenty of power and the HUD is kind of minimalistic too.

You guys wanting the penultimate reslolution and framerate. Just buy a real F-18 and get a drivers license ;-)

Tried this condor thingy, but the planes lack plenty of power and the HUD is kind of minimalistic too.

You must have bought the wrong one. The EB29 has a whopping 64HP

Thanks, I will add that one to my collection :-)

Nope. ED need to make their engine (far) more efficient. It cant even run smoothly at sustained 90+ FPS for most people on on flat panels. If they dont fix that, nothing can fix it for them. My future ideal VR goggles will most definitely require more, not less performance. Foveated rendering will allow the system requirements to be closer to current VR goggles, rather than just multiplying the requirements many times over, but its still gonna be harder. because you simply need a bazillion pixels to get "retina" resolution and decent FoV. Im not going to wait years to get eye tracking only to have that further degrade the image quality outside my focus compared to todays VR sets!.

As said on the begining, I calculate 10x increase in FPS (from were we are now) for 90-100 FPS on a dual 4K panel.

Imposible with brute force.

I only want to know if ED is working on that or waiting for that brute force.

As said on the begining, I calculate 10x increase in FPS (from were we are now) for 90-100 FPS on a dual 4K panel.

Thats not how it works. There is an overhead (and a fairly significant one in DCS I think) that is independent of your resolution. Go ahead and measure frame rates at 640x480 or whatever lowest resolution you can, its not going to give you a bazillion FPS. Then and scale it up to 4K and draw a line through the numbers and it will give you a rough idea. I think you may find the loss due to increased resolution is rather low, and the overhead quite significant.

Multiple projection angles should also not decrease speed linearly. In DCS it does seem to do that, or close to it, and running multiple monitors is really painful. In my other sim, going from 1 to 3 screens makes relatively little difference, despite tripling the number of pixels and camera angles, FPS is maybe 70% compared to a single screen, not 33% you might expect or the 50% or so Im guessing it is in DCS. .

Of course. The bottleneck doesn´t have to be on the pixel side of the equation. You know more than the average, that´s for sure. Thanks.

In fact, I think improving the core engine you could increase dramatically the oportunity to improve the FPS with Vulkan doing his thing, but without that part, maybe Vulkan can´t stretch his legs.

That´s what I meant: you go anywere without planing the improvements with a year or more in time.

But, surely, it have to be hard to stop developing other things and seems like you are not doing anything.

Imagine you change the core of the game and make it future proof: you surely have to rework almost all the modules to fit the thing or they allready would have done that before.

That legacy engine core will be hard to beat.

Maybe they have to look that problem from another side and emulate it so they can keep the same stuff but on another envelope that can multiproces the thing. I don´t know.

What I know is that the new damage model is taking much more time than expected and surely is delaying other developments and resources.

In fact, my complain is with the lack of knowledge on the developing arena.

At least, they gave us a table on the development efforts at the end of 2019. It´s only that they could update it more frequently.

In fact: the VR implementation of the game must to be good if the profesional headsets works well inside the game. (look at min. 9.30:

Well, people smarter than me (and that is not so difficult) is working on the problem of feeding those 4K and hier resolution VR panels on the future.

It seems upscalling resolutions is getting better: https://www.roadtovr.com/facebook-develops-ai-supersampling-boost-rendering-performance-high-resolution-vr-headsets/

The DLSS 2.0 on Nvidia works well allready upscalling 2x and, with more time, it seems the games could use foveated rendering and 16x supersampling in real time and transform a 960×540 rendered image on a 4K one with not much effort.

Obviously Oculus is working hard on that becouse they have in mind to continue on the success of the standalone Quest, and the need to cut corners toward 4K panels was there.

Amazing times guys!.....

Yes: they need to fix the VR problem because, I think, it will be the preferred game mode for many from now to the future.

Won´t be great to climb up the ladder to your cockpit from the supercarrier desk?

Even work on the clicking mechanics with the controllers or hand tracking.

There are many things to improve.

I said that: I won´t play the game untill a proper VR fix but, in honor of truth, almost every flight sim have problems in VR. Maybe it´s the type of game that suffer the most. I don´t know why but the problem is there.

Let´s see if Star wars squadron or MSFS 2020 suffer the same VR problem or not.