NVIDIA 5000 series and DLSS4 for VR

[Qcumber]

57 minutes ago, Hippo said:

I see that this appears to be the general consensus on here, but how are people coming to this conclusion?  I would say the opposite: I find that it is the GPU that maxes out when I try (tried - I sold my PC a few months ago) to hold 72 fps with my Quest Pro / 4090 / 13900k, high VR rendering resolution and in-game graphical complexity.  I would have chosen to upgrade the GPU rather than the CPU, and going 3080Ti -> 4090 was a substantial improvement.  I do tend to run single player missions with not much going on, and test that way, so maybe that's it?

Unfortunately it doesn't seem that the 4090 -> 5090 uplift in rasterisation performance is going to be particularly substantial.  I did manage to sell my 4090 for around 80% of what I paid for it.  If a reasonable price can be obtained for the 4090 then the (supposed) improvements in upscaling quality and performance could make it worthwhile.  I'm looking forward to seeing the results of real world DCS VR testing.

I presume you have not been using eye tracking with your QP? With QVFR enabled your CPU will be the bottleneck. 

[Hippo]

2 hours ago, Qcumber said:

I presume you have not been using eye tracking with your QP? With QVFR enabled your CPU will be the bottleneck. 

You presume incorrectly sir/ madam: I do use QVFR and will not consider a headset without eye-tracking; if you're interested to see how I tested and my settings, you might consider clicking on the link in my previous post.

Edited January 12 by Hippo

[Qcumber]

31 minutes ago, Hippo said:

You presume incorrectly sir/ madam: I do use QVFR and will not consider a headset without eye-tracking; if you're interested to see how I tested and my settings, you might consider clicking on the link in my previous post.

Sorry. "Presumption is the mother of all F*** ups". Do you see any performance gain with your CPU/GPU combination? 

[Hippo]

3 hours ago, Qcumber said:

Sorry. "Presumption is the mother of all F*** ups". Do you see any performance gain with your CPU/GPU combination? 

No worries, only kidding.

1.  I sold the PC in my sig. a few months ago.  2.  My testing was limited to missions like that uploaded in the linked post.  3.  I attempt to hold 72 fps (no ASW) and try to get just below 100% gpu util with a demanding mission such as (2), so that I can have plenty of headroom for more normal loads.

Bearing (1), (2), (3) in mind:

3080Ti, i7 8700k -> i9 13900k - no benefit whatsoever that I could discern.

13900k, 3080Ti -> 4090 - major benefit.  There was no way I could hold 72 fps in the mission with those settings (see linked post) until I got the 4090, the gpu util would always hit 100% and ASW would kick in.

I found that if I wanted 72 fps on a mission like that in the link, with those settings, I needed a 4090.  I don't know how people with, say, a Crystal Light (or eventually Super!) or similar higher res headsets can run without reprojection or lowering rendering resolution from optimal.  I was running QP at 5408 x 2736, PCS is 12840 x 7412 at 100%.

If, as is very likely, I'm missing something, I'd be very grateful to be informed by the knowledgeable folks on here.

Edited January 12 by Hippo

[nikoel]

Okay, I’ll bite.

This is why it’s called CPU and GPU pairing

Within DCS, it’s not difficult to create scenarios where one is CPU-limited or GPU-limited, even with the same hardware. It depends on the scene being rendered. For instance, if we were to use a carrier deck population template, you would have seen a massive performance uplift by upgrading from the 8700K to your 13K. However, in scenarios where you are GPU-bottlenecked like a 1:1 knife fight, getting a faster processor only results in very minor gains. It all depends on what is causing the bottleneck

For people with identical hardware, the experience can vary greatly depending on what they are doing within DCS and this is why opinions vary so much

There are tools like QuadViews that can shift some of the load from the GPU to the CPU. This is one of the reasons why it’s incorrect to say it “gives you more performance.” It provides more performance in GPU-limited scenarios but reduces performance (or performance headroom) in CPU-limited scenarios. Or Increases CPU frametimes and decreases GPU frametimes

Now, regarding your question about where you’re going wrong: this brings me to the topic of resolution. Pimax’s new headsets are indeed high resolution, but it’s unwise nuts not to run them with foveated rendering as it brings the GPU to it's knees

This means the total number of pixels being rendered is far lower than you might think, and the uplift in resolution of the headset is much greater than the actual number of pixels the GPU has to render. This is solely because of QuadViews. It allows you to significantly reduce the number of peripheral pixels without noticing a difference. It's not small either, depending on the end user's taste, the difference is over 70% fewer pixels rendered for roughly the same picture quality.

No difference because our eyes perceive sharpness and focus within roughly a 3° cone, so the periphery pixels matter far less than the ones inside that 3 degree arc. You can test this for yourself with the Quest Pro in your signature. Eye tracking allows you to lower GPU overhead at the cost of slightly higher CPU frametimes. I run mine on a 4090 with all settings dialed to the moon with the use of QuadViews, something I could not do without it. It's comes at a cost when I am on a fully populated super carrier deck and this is where performance can suffer

So because of my CPU bias created by QuadViews, compared to someone who is running QP at 5.5kx2.8k, I am running a lower overall resolution, a sharper higher resolution image within the zone that I can see, at higher overall settings at a cost of some stuttering where I am severely cpu limited (usually in that said super carrier in campaigns)

Finally, youre right to doubt the precision of frametime tools offered. if you’re looking for a proper utility to measure frametimes, check out Fred’s OpenXRFrameTools here
https://github.com/fredemmott/XRFrameTools

 

Edited January 13 by nikoel

[Youtch]

5 hours ago, nikoel said:

This means the total number of pixels being rendered is far lower than you might think, and the uplift in resolution of the headset is much greater than the actual number of pixels the GPU has to render. This is solely because of QuadViews. It allows you to significantly reduce the number of peripheral pixels without noticing a difference. It's not small either, depending on the end user's taste, the difference is over 70% fewer pixels rendered for roughly the same picture quality.

No difference because our eyes perceive sharpness and focus within roughly a 3° cone, so the periphery pixels matter far less than the ones inside that 3 degree arc.

What is the difference in pixel rendered between quadview for eye tracking with dynamic foveated rendering vs fixed foveated rendering for no eye tracking headset?

[Mr_sukebe]

51 minutes ago, Youtch said:

What is the difference in pixel rendered between quadview for eye tracking with dynamic foveated rendering vs fixed foveated rendering for no eye tracking headset?

 

As I understand it, both fixed and eye tracked have the options of varying the size of the fully rendered area (let's assume the middle) and the decrease in the number of pixels being rendered outside of that central area.

So if the size of the middle area is the same and the decrease in pixels outside of that middle area are the same, from a performance difference, there should be nothing in it.

The key point is that eye tracking ensures that if you move the focus of your eyes to a different point with the VR display, that the centre of that area fully rendered follows it.  With fixed, look at a corner, and it'll be a blurry mess.  With eye tracking it won't.

The implications being that with fixed, you're likely to have a large "middle" area at full resolution.  With eye tracked, you can run using a smaller central area and still see full resolution wherever your eyes look.

In short, eye tracked, when configured correctly "should" result in not only a better performance, but also a uniform visual experience.

 

 

 

Edited January 13 by Mr_sukebe

[Hippo]

6 hours ago, nikoel said:

So because of my CPU bias created by QuadViews, compared to someone who is running QP at 5.5kx2.8k, I am running a lower overall resolution, a sharper higher resolution image within the zone that I can see, at higher overall settings at a cost of some stuttering where I am severely cpu limited (usually in that said super carrier in campaigns)

Thank you for your post, all very interesting and sensible.  Wrt to the above, if I'm understanding you correctly, you're setting your QP / PCL at a lower res in its software, but then running at a higher resolution (at least as high as the headset's recommended for barrel distortion correction) in the foveated area?  If so, I didn't think to try that, and didn't even know if it was possible.  I'm currently running a lower spec PC and will give it a go - thanks.

If you can spare the time, it's always interesting to know hw spec + settings + fps obtained for the types of mission you normally run.

[Youtch]

Thanks for your reply.

This somehow confirms my initial impression, although i have a hard time to find data which could quantify the actual difference between Dynamic and Fixed foveated rendering in terms of number of pixels rendered, or effective reduction achieved compared to full resolution for each solution.

I was thinking indeed that you could potentially achieve much higher reduction with DFR than with FFR, because with DFR you can focus on the cone only and leave all the rest with very low resolution, and not realize it, while in FFR the resolution outside the cone still need to be somehow higher to not be shockingly blurry.

Hence, if DFR with eye tracking can achieve 70% reduction of pixel (as per a post mentionned above), what can you reasonnably expect for reduction with FFR to obtain a similar impression of clarity?

[nikoel]

A quick edit since people are treating 70% as a given. This is an estimation—an educated guess—based on the advertised resolution of the Pimax with DLAA. It also comes down to the preference of the end user. With lower-resolution headsets, such as the Quest Pro, the percentage will be relatively higher (even though the end resolution is lower) due to the headset starting with a lower resolution native

There’s also an elephant in the room named MSAA, which, granted, produces a sharper image but will likely require a higher resolution, as the shimmer is on another level and highly distracting. That said, I know many people prefer it over other options. Here’s a screenshot of the settings I was using. They’re probably not the most optimal, but with a 4090, being in the ballpark is good enough

null

 

[Mr_sukebe]

1 hour ago, Youtch said:

Thanks for your reply.

This somehow confirms my initial impression, although i have a hard time to find data which could quantify the actual difference between Dynamic and Fixed foveated rendering in terms of number of pixels rendered, or effective reduction achieved compared to full resolution for each solution.

I was thinking indeed that you could potentially achieve much higher reduction with DFR than with FFR, because with DFR you can focus on the cone only and leave all the rest with very low resolution, and not realize it, while in FFR the resolution outside the cone still need to be somehow higher to not be shockingly blurry.

Hence, if DFR with eye tracking can achieve 70% reduction of pixel (as per a post mentionned above), what can you reasonnably expect for reduction with FFR to obtain a similar impression of clarity?

 

The amount of reduction is always a question of personal expectations and what is considered acceptable.  Some might be happy with a really small central area in FFR, whilst others not.

As I understand it, quad views by default has the central area set to the x1 for the number of pixels processed, and then an option of say x1/8th (which can be varied) for the outside area.  One way of looking at it, is that with the savings on the low resolution area, you could deliberately oversample the central area, to for example x1.5 and still have used less processing power than with QV disabled.

 

[Qcumber]

2 hours ago, Youtch said:

This somehow confirms my initial impression, although i have a hard time to find data which could quantify the actual difference between Dynamic and Fixed foveated rendering in terms of number of pixels rendered, or effective reduction achieved compared to full resolution for each solution

With QVFR you can control the centre and periphery. The default settings are 0.5x0.5 at 1.1 times base resolution and 0.5 for the periphery. This means that 75% of the headset image is at 0.5 resolution. The only difference between fixed and dynamic is that eye tracking allows you to move the centre. As a result you can get away with a smaller centre size. I have mine set to 0.3x0.3 so 91% of my screen is at resolution 0.5. It gives me a massive performance boost. 

[gonvise]

Quest pro & 4090 & 7800x3D here... What do you think is best?

1. Super Sampling 2.0 via OTT and Quad views center resolution 1.0 or...

2. No Super Sampling (1.0=default=original resolution) and Quad views center resolution 2.0

[Youtch]

1 hour ago, Qcumber said:

With QVFR you can control the centre and periphery. The default settings are 0.5x0.5 at 1.1 times base resolution and 0.5 for the periphery. This means that 75% of the headset image is at 0.5 resolution. The only difference between fixed and dynamic is that eye tracking allows you to move the centre. As a result you can get away with a smaller centre size. I have mine set to 0.3x0.3 so 91% of my screen is at resolution 0.5. It gives me a massive performance boost. 

Thanks for your reply, very helpfull.

Good to learn the possibility to live with a smaller center size with eye tracking dynamic foveated rendering compared to fixed.

-> Hence the difference here would be 97% at 0.5 resolution vs 75% at 0.5 resolution.

But I am quite surprised to hear that both dynamic and fixed accommodate the same 0.5 resolution for the peripheral area. I was assuming (maybe wrongly) that you could live with lower resolution on the periphery with dynamic eye tracking, since you would not really realise the resolution of the object out of your focus because out of your eye cone, while with fixed looking at the sides would result in a very blurry impression hence requiring to dial up the resolution of the periphery.

[gonvise]

1 hour ago, gonvise said:

Quest pro & 4090 & 7800x3D here... What do you think is best?

1. Super Sampling 2.0 via OTT and Quad views center resolution 1.0 or...

2. No Super Sampling (1.0=default=original resolution) and Quad views center resolution 2.0

I answer myself...

1 Far better, 2 much worse center of QuadFR view quality and very noticeable area outside the foveated.

Similar performance in both.

[EightyDuce]

21 minutes ago, Youtch said:

Thanks for your reply, very helpfull.

Good to learn the possibility to live with a smaller center size with eye tracking dynamic foveated rendering compared to fixed.

-> Hence the difference here would be 97% at 0.5 resolution vs 75% at 0.5 resolution.

But I am quite surprised to hear that both dynamic and fixed accommodate the same 0.5 resolution for the peripheral area. I was assuming (maybe wrongly) that you could live with lower resolution on the periphery with dynamic eye tracking, since you would not really realise the resolution of the object out of your focus because out of your eye cone, while with fixed looking at the sides would result in a very blurry impression hence requiring to dial up the resolution of the periphery.

The problem, if you can call it that, with DFR is the precision of eye tracking that the headset is able to achieve. It is not exactly 1:1 precision, so you need a bit if wiggle room. 

[Qcumber]

1 hour ago, Youtch said:

Thanks for your reply, very helpfull.

Good to learn the possibility to live with a smaller center size with eye tracking dynamic foveated rendering compared to fixed.

-> Hence the difference here would be 97% at 0.5 resolution vs 75% at 0.5 resolution.

But I am quite surprised to hear that both dynamic and fixed accommodate the same 0.5 resolution for the peripheral area. I was assuming (maybe wrongly) that you could live with lower resolution on the periphery with dynamic eye tracking, since you would not really realise the resolution of the object out of your focus because out of your eye cone, while with fixed looking at the sides would result in a very blurry impression hence requiring to dial up the resolution of the periphery.

You can easily change all the settings in the config file or using Quad Views Companion. There is a limit with peripheral resolution as DCS does not always cope well with textures and shadows at lower resolutions. 

[Youtch]

2 hours ago, Qcumber said:

You can easily change all the settings in the config file or using Quad Views Companion

I would if I would have a headset, unfortunately my G2 is finished, and I am in a market for a new headset and also a new GPU (coming from a 3080).

Hence, I am questionning myself if I should buy a new headset without or with eye tracking, based on added performance and clarity, or if it is better to go with fixed foveated headset and invest into a more powerfull GPU.

I was initially under the impression that it might be better to invest more money into a headset with eye tracking to use Dynamic Foveated rendering, than investing into a 4090 instead of a 4080 for instance, or a 5070 Ti instead of a 5090.

But since I cannot experiment it by myself, I am putting this impression out there to be challenged by people who might have played with both Dynamic and Fixed Foveated rendering, and be able to assess the performance gain from one to another, as well as the overall impression of clarity.

 

 

[Tensorial_Architect]

(Own or have owned, G2, Aero, Crystal, Crystal Light, Quest 3, and Quest Pro)

You will get more bang for your buck Youtch by upgrading to a Quest 3 and enabling the Fixed Foveated mode from MBucchia's Quad Views code. I think I read somewhere that you hate Meta, ... well, ... so do I. But, unless you want to risk the beta project that is the Crystal Light or Crystal from Pimax, ... the Quest is your best answer. You will get better results than upgrading to a very high-end GPU. (To drive an Aero or Quest Pro to very high visuals, you will need both the expensive headset and a high-end GPU). I really enjoy my Quest Pro but I use a 4090 to drive it (and the Quad Views code).

Edited January 13 by Tensorial_Architect

[Aapje]

@Tensorial_Architect

I suspect that you used the Crystal & Crystal Light early on. From what I've seen on the forums, they've improved the software a lot since then. The main issue with the Crystal Light now seems to be the poor QA, so a relatively high percentage of shipped headsets have hardware issues.

@Youtch

My suggestion would be to get a Crystal Light with a 5080. You can probably get the 5080 within a few month, especially if you join one of the discords that watches for supply at retailers.

And you might have to send back the Crystal Light anyway, if it has issues. And otherwise you can start using it with the 3080 at first.

[Tensorial_Architect]

Aapje, I got my Crystal Light back in December. It has both software and hardware issues. I am not trying to be mean, it is just not a mature product ready for release. My Crystal also has both hardware and software issues but less on the hardware side than the Crystal Light. 

[Aapje]

Recently I've seen people who were surprised by how well the software worked, so then I concluded that they improved relative to their reputation.

But I don't have first hand experience with the Crystal Light, but I also don't know how high your standards are.

[Blackhawk163]

1 hour ago, Aapje said:

Recently I've seen people who were surprised by how well the software worked, so then I concluded that they improved relative to their reputation.

But I don't have first hand experience with the Crystal Light, but I also don't know how high your standards are.

Pimax play is just a launcher for me at this point. Other than that my unit works great. What software issues are there for the PCL?

[nikoel]

Pimax conducts its beta testing on paying customers by releasing products that look good on paper but are unfinished, buggy, misadvertised, over-promised, and incomplete. These products are often delivered well after the promised date when pre-orders were taken, yet somehow still feel rushed, with promises to “fix it down the line.” Once the products finally reach customers, Pimax promptly announces another new product, reallocates resources to it, and abandons the previous one—often leaving promised features and inclusions unfulfilled. All of this is wrapped in a support network that ranges from indifferent at best to downright illegal (in some countries) at worst—only for the same cycle to repeat again and again

There’s a guy running around claiming he’s joined Pimax to fix this, but he clearly hasn’t realised he’s not the first. It doesn’t take much digging to see it thats it's likely just all talk

Nothing screams “confidence in the brand and the products we make” quite like an extended one-year warranty for—checks notes—$250 AUD.

Well, nothing except the realisation that even at that price, it’s now “out of stock.” How terrible must their average quality control be if they’re losing money on a warranty that is legally mandated in some countries?

https://pimax.com/products/pimax-crystal-1-year-extended-warranty

If you’re planning to share this with your local shady used car salesman, just remind them to consult a healthcare professional if it lasts more than four hours.

It all is not helped because the one of the two VR content creators who I actually trusted (Wolta) has sadly deleted his youtube account. The rest of the landscape are paid promoters, shills and salesmen who are trying to peddle whatever this weeks sponsor has given them and are too scared to say anything bad about the product in case they do not receive the sponsorship next time or it affects their affiliate sales. Or my favorite throwing tantrums because a headset company did not send them a headset for a review. 11 year old children, all of them - I digress

P.S. The company reminds me of another startup in the UK called Hill’s Helicopters. They’re a little behind Pimax, but the grift is eerily similar

Edited January 15 by nikoel

[HoBGoBLiNzx3]

Thread has gone off topic. Any viable information on the topic should definitely be posted. Esp from the devs and their implementation of DLSS in DCS. Interestingly enough the upgrade to DLAA alone should be help. It would be really interesting to see if multiframe gen can be forced in the nvidia app in VR.