@CptLucky8 If I had to guess, the full-screen force was a move to avoid input issues similar to what we see in DCS when users accidentally click outside of the desktop application. It’s making testing really frustrating.
What setting is for TAA%? I’m using SteamVR and am trying to set FS2020 TAA to 60%…
Captain, have you arrived at a final conclusion regarding WMR vs SteanVR Open XR implementation on the G2?
I am wondering the same. Also what is the relationship between render resolution in OpenXR vs in the game settings?
I have the following setup:
32gb 3800 CL16
HP Reverb G2
I’m running on MS store version with open XR developer tools installed.
Given your recommendations above are for index and steam, what is it I should be trying in terms of ballpark? Motion reprojection always on, custom render scale = 200% or 100%? (That is the bit that confuses me as I’m new to VR). In the sim I run 100TAA.
Currently I find it difficult to read instruments in the cockpit with custom render scale auto and 100TAA, I can’t seem to find what it is that’s causing this as DCS is running crystal clear. Thanks!
I tried openxr at 50 today and ended up with a weird fish eye lens view that wouldn’t go until
I shut the pc down and rebooted.
Since then I’ve had horrible performance issues. Tried new and old drivers and getting bad performance, I wish I’d left it as it was now.
(Note to self):
I don’t think it makes a lot of difference between openxr res and game render res, just find the max res it’ll run at 30fps and leave it.
I’m not sure, pdawg17, if you’re asking the same question I have or not: I don’t understand this terminology of referring to “60 TAA” or “TAA 60%”.
In the game graphics menu there is a render scaling slider and, separately, there is an anti-aliasing setting.
The render scaling slider, it seems to me, is the same as an external super sampling setting. This setting, combined with any external SS setting, determines the render resolution.
The anti-aliasing setting, FXAA, DLAA, TAA (temporal anti-aliasing), on the other hand, sets the anti-aliasing mode and seems independent of the render scaling. For instance, I could have the game set to 110% render scaling and FXAA anti-aliasing. That wouldn’t be “110% FXAA”, it would be 110% render scaling and AA set to FXAA.
So, that’s my confusion. Why, in this thread, are there references to “60 TAA” or “TAA at 60%”? There is no slider specifically for setting a TAA “percentage”, is there?
I’m curious, so you have the dropped frames without a glass cockpit? I’m working with latest drivers trying to isolate the stutter.
The spikes on my machine happen with the 172 g1000 and to vary some with the glass cockpit refresh rate, but in a steam gauge airplane they don’t seem to happen.
Hey there. Welcome to tech blogging. Please note that most of us don’t know how to adjust TAA in game so would be good if you could mention that, same with what EFIS means? Lastly, reflections make a huge difference, surprised you didn’t mention that. Otherwise thanks for the amazing work and help!
When he says 60 TAA he is just referring to render scale set to 60 and TAA on. With the current recommendation (for the G2 anyway) when he says 110 TAA and 50 render, he means in-game graphics render slider to 110, TAA on, and openxr developer tools custom render scale for the reverb g2 set to 50.
Wow everything here is so interesting!
But i just would like to know why with a 3800X-2080Ti-32GBRam, at minimum settings i have -150% headroom, and so jittering-juddering-stuttering that makes you puke after 1 minute?
PC specs: Ryzen 5600X/32GB/RTX3070/G2
Just to let you know all, I’ve lost 2 days reading and fighting for VR performance only to found out that switching from OpenXR dev.tools to SteamVR environment gives me WAY BETTER results!
I have no idea why, but with WMR OpenXR and motion reprojection ON there’s no way I could get a stable 30 fps (stutters, glitches, nightmare!). Now with SteamVR env. and the recommended Nvidia CPL settings (thanks CptLucky8!!!), I’m getting stable 35-45fps in almost all situation (low, medium settings, ultra clouds) in FULL RES (2236x2192)!!! It’s a night and day experience!
If you’re strugling with your G2 just give it a try!
I can still loose a frame here and there but man, I’m finally happy!
Merry Christmas everyone!
Could you post your settigns in MFS2020 and SteamVR?
I’m also struggling with 3070 and G2, still unhappy, very inconsistent performance, barely at 30FPS, lots of glitches, stutters
The problem for me with steamvr is that I can’t maintain a stable fps, 45 is too high to keep at reprojection and the stutters make me sick. For now i will play with openxr and reprojection enable for stable 30fps, it’s still not perfect, and i do have some stutters but it’s beareable for me.
When I’m writing “Motion Smoothing” it is anything which is related to “doubling the frames” or similar. In WMR it is the Motion Reprojection drop down setting.
Thank you for your kind words! Please note game render resolution affects legibility up close especially EFIS. VR super sampling affects legibility in the distance. For sharper EFIS with the Index, use TAA 110 or 120 first, then adjust SteamVR SS to what’s confortable fps wise for you. As for 40s territory, although you must enter VR with the Index at 90Hz in order to reduce CTDs when starting VR mode, once there, you can change to 80Hz. I used to do this in order to unlock the simulator from trying hard displaying at 30fps when it could do better and have it locked to 40fps instead!
This is most likely the case although when in Window mode it is constraining the mouse into the window boundaries unless you ALT+TAB or press the WIN key. Rather, I suspect they’re using a different DX11 present mode which is better suited for VR and/or 3D (I don’t remember which exactly but this was an issue raised a few weeks/month ago about how FS2020 was not using the correct one for full screen borderless windows).
Yes I’m almost done comparing many different NVidia drivers, TAA/SS ratios, settings, all best suited for a base line. My approach is really to find this baseline which is working really good as-is in most situations/aircraft/place for a given hardware (my test system which is in the higher range CPU wise and mid range GPU wise and seems to be actually quite mainstream in the simulation world). From the base line, people with better hardware can raise some settings with a better knowledge of what is affecting the most the experience.
I’ll post this soon but I can already say I’ve found I believe an ideal sweet spot with fluid flights and legible EFIS over LAX region in the default seated A320 position (using the center post red/white orbs for alignment) with some room to increase EFIS legibility further over less dense regions.
No I didn’t because I didn’t spent much time trying SeamVR OXR with the G2. I just quickly tried out yesterday because there was a new SteamVR beta and it seems quite good or actually nearly similar to my WMR tests I did just before, but there was a big problem with the HMD tracking causing frames displaying in a backward timeline fashion (like when playing some AVI files which are wrongly encoded with a simple video player making frames to jump back and forth).
I’ll be posting my suggestions for the G2 + WMR soon, in the meantime, with the video card you have, I’d suggest you try this instead for crystal clear legibility: TAA 120 + OXR 50 + Motion Rep. Disabled. Let me know about this one…
Please look at the tables comments I’ve posted with different combinations for the Index. It does make a difference because both scales are not affecting the same pixels the same way, let alone EFIS drawing resolution solely depends on FS2020 Render Scale, not OXR/SteamVR Display Res.
It is just a shorter way to tell both FS2020 Render Scaling value and Antialiasing setting in a convenient 3 letters label. Mind you, besides TAA mode, all others are causing shimmering and would require MSAA or similar to compensate. Therefore it is easier for me to just say:
- “TAA 60” is FS2020 Render Scaling 60% with TAA mode
- “SS 150” is OpenXR Render Scale or SteamVR Super Sampling 150%
I never tried to compare without any glass cockpit. I’m trying to find a base line working best with the most demanding cockpits and areas (I’m usually flying a certain pattern I’ve build up to know over time in the KSNA region - you can see me flying outlawed crazy pilot there when I do), in order to make sure I won’t have any unpleasant drop when flying in a variety of places and aircraft. I usually validate it is working good there, then I take a fly in the Hawaii islands to enjoy how much it is actually working great!
Thank you! I’ve answered this “TAA” question in more details above now. I usually prefer helping people learning how to fish, but here is EFIS definition just a qwant.com search away
I don’t find reflection making a huge difference actually, at least at LOW, but I suggest leaving them OFF to get started with. If your hardware can deal with it, fine. Otherwise, you might prefer raising other settings more important to the pilot than reflections, such as clouds from MED to HIGH.
Thank you for your kind words! It is often neglected but very important indeed, and I personally find most people doing YT videos about “the best settings” don’t know much either. When you’re accustomed to 3D rendering APIs and the evolution of the hardware over more than 25 years (some remember their first Voodoo cards?), it doesn’t take much looking into the internet and properly reading the NVidia materials to start refining these settings to something actually working better.
As a matter of fact, I’d suggest you review your NVidia settings for your other games too, because these are working great for me with SW Squadron, HLA, and basically all others I’m using.
I’ll also update the post because I was suggesting VR Pre-Rendered Frames “1 most likely” but I’ve been trying out different values and I can confirm it is “1 and only 1”. Any other value is stuttering much more on different NVidia drivers I’ve tried lately for the G2 WMR recommendations I’m preparing.
YOU, SIR - ARE A GENIUS!!!
Entered settings for Geforce 1080, ditto recommendations for MSFS 2020 VR graphics - INCREDIBLE IMPROVEMENT! All cockpit numbers visible - smooth graphics. Yes - 3 miles ahead you could see the scenery populate itself - but minor inconvenience when the flying experience is SOOO improved.
SETUP: Intel core i7 9700K / 32 GB mem DDR4 / GTX 1080Ti w 11 GB / Rift S
Are you kidding? Realy? Smooth? can you show video/screen on FPS monitor?
I9-9900k/32/2080ti/Rift-S. Jiggling/Stuttering/Low FPS here. Nothing is help
Thank you for your kind words! I’m glad this is helping! I believe with the 1080TI you might want to try pushing Terrain LOD to 100% and see how it goes.
Besides, after reading so much reports over the years about video cards and perf in various simulators, I strongly believe any “TI” variant are more capable of super sampling than the non TI variants. It is as if every time the non-TI cards have more performance problems when shrinking a higher res rendered image to a lower res displayed image. I also believe they are better suited for MSAA kind of AA and this might be because of the extra CU or even maybe because of dedicated silicon.
I might be delusional but this seems like a pattern to me be it a 1080TI, a 2080TI and anything TI. If this is the case, I suspect you should be able to raise reflections for example without much fps hit.
I’m using a Rift S, i7 9700F, 48GB RAM, RTX 3090.
I generally have good results, especially in the smaller planes at smaller airports, like I like to fly. The cockpit text can’t be made super clear in the Rift S, but I’m used to that.
My problem (or confusion, I guess) is that I decided to set my OpenXR Runtime to Steam as a test today rather than Oculus. I set the Steam SS to 100% initially and was planning on playing with it vs the game render scaling to see what gave me the best visuals and smoothness.
But I couldn’t even do the test because as soon as I switched to Steam moving my head in the cockpit causes a weird fish-eye like distortion. For instance, looking down at the center console in the A320 brings it closer to me. Looking up moves it away. Moving my head toward the co-pilot seat moves the distance of the far window.
I have built up a lot of tolerance to VR motion sickness in playing sim racing games and FS2020. But this fish-eye distortion when moving my head with the Rift S and Steam driver almost instantly made me queasy. It’s lingering even as I type this.
I obviously need to switch back to Oculus in the Open XR registry entry, even though it seemed that the Steam driver may have given me enough FPS boos to get ASW working at 40fps. I’m just confused how ANYBODY can get this to work with the Steam driver, at least with the Rift S. Wonder what’s different in my setup. I’m still queasy, just thinking about it.
You disable motion reprojection? So with 30 FPS, you will just have these 30 frames per second? Shouldn’t this be nausea inducing? No frames inserted in between by reprojection?
I really appreciate your efforts of finding the best settings. For a VR novice like me it would be extremely helpful if you could illustrate the settings with screenshots.The VR terminology is a bit complex, so sometimes I don’t know which setting is in OpenXR Developer Tools or in SteamVR or in some other place. For example it would never come to my mind to disable the reprojection (I will try it, if this will reduce the strange artifacting within the propeller ring).