VR Tips Worth Knowing…add them here

This will only help if you have a 30-series GPU… Unfortunately for me!

1 Like

Also, cockpit movement/scanning is very smooth, like TrackIR, except for some movement/ head trembling associated with Windows 11 currently.

I saw your post showing a 1fps gain from 60 FPS to 61fps in 2d mode some time ago, do you believe the benefit is significantly greater now? (I haven’t yet flashed this to my GPU and mobo due to conflicting reports)

I have a Q2.
I don’t notice anything (it’s probably my sight).
And now I am testing 0.7 0.7.

I don’t have a link handy, but it’s on this forum, I’m sure if you search you can find it. You’ll need a free Photoshop DDX plugin from Nvidia, or some other similar tool. Normal trick is to just increase prop opacity, but my trick is make a solid disk of the same opacity instead of the prop texture.

I tried this and it gave me more stutters instead of less… Actually is’t a horrible stuttery mess. Total vomit comet.

I really feel for those of you with RTX 30** that just can’t seem to get satisfactory performance in MSFS. It seems to have been an issue with MSFS since before VR. I wish I knew. perhaps it’s something to do with the architecture differences, but a 30FPS, smooth experience should be a walk in the park for those high-end cards.
I’m still hoping MSFS will significantly improve on RTX30** before I upgrade from my GTX1080ti.
OXR 100 RS 70:

Those are strange settings - clouds at low look horrible. No grass and brush at all. But Ultra textures… And Cessna 152 on somewhere sparse…

This is really puzzling to me as well. I’ve been running 6800XT and am getting stutter free 35fpls using 90OXR/ 100Render Scale with Reverb G2. Could this be mainly a driver issue?

I’m interested in flying, reading instruments etc. Place is Singapore WSSS - so hardly sparse.

I’ve found with my i9-10900 OCd to 5.2Ghz stable/RTX 3090 (non-OCd) and 128GB DRAM running at 2926 Mhz and Reverb G2 I get great (super smooth) performance with OXR set at 70 and 100 Render scale. I set Motion Reprojection to Always On. I turn HAGS off which gives me a slightly smoother ride. I set most settings to High and medium with an odd ultra. Clouds are always at High. I don’t lock my frames in 3d Nvidia settings, and I use virtual reality prerendered frames setting at 2. I fly the Hawk almost exclusively (I’m a former RW USAF fighter pilot (2300hrs) (T-37/38/F-15Es)) at low level and operate in the 450-500KIAS regime. I rarely get a few mico-stutters and it’s usually left 90 degrees and it’s not even noticeable. Interestingly, I almost never get micro stutters right 90 degrees.

I attempted to crank/OC both the GPU and DRAM only to see more stuttering and less stability if that makes sense.

Two things may change the playing field for VR: 1) DirectX12, and 2) A new version of OXR which is apparently coming out soon (see Kronos).

Thanks for all the tips. I learn something everyday.

Yes, most likely. 3xxx cards should get a huge advantage, but people on 2080 often get better and smoother results. It just doesn’t make sense that an overclocked 3080 can’t get me a smooth flight over Manhattan at 70OXR/100TAA MR on mostly Medium settings (high clouds). And now I’ve exhausted all the tricks in the book.

  • Resizable bar
  • Resizable bar profile tweaks
  • Dev tool tested in all kinds of settings, in all combinations, preview on and off, custom render scale at different values.
  • Process Lasso to get everything to last 4 cores, and get MSFS off cores 0 and 1. It’s free to use all of my other 18 logical cores (I have 24 of them on 12 physical cores!)
  • Limited frames (tested 24 to 34 - no significant difference, bu limiting is better than not)
  • Killed all unnecessary processes and software
  • Killed the Precision X1 software
  • Tested all kinds of in-game settings, different OXR/TAA combinations (200/50 is a real vomit comet)
  • Overclocked GPU, CPU and RAM (and tested on stock to make sure OC doesn’t perform worse) - on a liquid cooled CPU with a triple fan radiator, and 5-fan case that can almost take off by itself.
  • New AMD Chipset Drivers (first time I tried them - they killed FPS to about steady 5FPS, but this time - it worked OK, but no improvement to FPS either)
  • Rolling cache on a RAM drive - I have 64Gb of fast RAM, I can spare 16Gb for the cache as MSFS isn’t using even 32Gb.

I did get my main test PA28 around LOWI with Ultra/High clouds with no stutter. My best result was with a Process Lasso, frame limiting (24 or 28 or 32 - doesn’t make much difference), OXR75/ TAA100. Almost no stutters at all, I could fly 30 seconds without a single one, and then probably loading scenery would do a small one, but no regular mucrostuttering at all. That’s much better now than before this topic.

But then I loaded Toronto in my CYKZ airport and my Toronto pack - and had to adjust clouds down to High and OXR to 70. And still got some stuttering, although it was bearable, byut barely. And then I loaded KLGA in New York and tried to fly over Manhattan. And that killed it - stutterfest again. Horrible. I give up. I don’t think there’s anything more I can do… 3080 should do better than that on 5900X. But MSFS never fails to bring it on its knees… We’ll have to wait until Asobo optimizes VR better. Till then - flying in the Alps or New Zealand on a GA aircraft is the best I can hope for without vomit-inducing stutters…

Reducing resolution below 70/100 is ruining the experience, I won’t enjoy it with the blurry gauges and fuzzy aliased scenery…

Copy that Roman. I haven’t operated in/near large metro areas/airfields except for KATL in the TBM. When I role medium/high & slow, I typically turn reprojection off but stick with the other settings. I get smooth performance. Other than that, I stick to low fly in the UK (I was stationed at RAF Lakenheath in the 492FS) and am pretty familiar with the low fly areas and weapons ranges there. Occasionally, I’ll low fly in the Nellis Range complex outside Vegas–super fun/brings lots of memories back.

Tip: If you fly the 310 Radial for about 13-14DME from RAF Marham (Tacan channel 24) you will find yourself at Holbeach Weapons Range. Locate the “donut” which is a small donut shaped circular island (looks like a drain in the Wash (the bay). From there fly roughly a 315 heading and you will see the strafe pits and to the right and long of the strafe pits you will see the bombing circle.(315/19DME from Marham). It’s a great place to do practice gun and bombing runs in the Hawk. We used it all the time in the F-15E with 20mm rounds and practice BDU-33 ■■■■■ bombs. Cool to see it in MSFS!

3 Likes

One of the main issues with comparing all the user feedback in terms of performance is the flight conditions used. Statements like “I’m only getting 25…30 FPS while you are getting 40+” are, not being offensive, nonsense. Depending on whether you’re flying low over photogrammetric cities, dense forests or 2k up in the air over plains will easily make 10…15 FPS difference. Different airplanes and weather conditions as well as lighting conditions (time of day, season) add to that. That’s why I’m doing my benching using the Flight Recorder addon with a pre-recorded flight using the same aircraft, same year, same date, same time of the day, same weather.

I suggest that if MS implement a replay feature as proposed for SU7 (?) which also allows exporting/saving those flights, we should record ONE 5…10min bench flight that will then be shared here in the forums and used as a standard for comparison.

I’ll start a corresponding thread and will try to include my current bench flight using the Flight Recorder addon.

Here’s the thread: Standard Bench Flight for Performance Comparison

1 Like

To compare fps with different setups and settings why not use one of the discovery flights? Berlin for example as its demanding and could be counted as a worst case scenario and you can judge the measurement time easily until you reach the tv tower. When I make adjustments this is what I use so that all the sims variables are constant each time giving a like for like comparison.

Because there’ll be significant variance in altitude, heading, speed and other parameters for different users. And as FPS are very sensitive to such variations, it is best to fix as many variables as possible. In my opinion this is the only way to have a solid basis for comparison.

2 Likes

A discovery flight has the least amount of variables, at least if you are comparing between two users. I’ve seen a few times people come out with claims about 45fps everything ultra at 100/100, but what traffic do they have enabled? What weather (has a huge impact)? What altitude? With a discovery flight you just keep the heading, altitude and speed which will be roughly the same for everyone and the best way to compare between users. It’s also very good for individual comparisons when changing settings.

You’ve just written it yourself. Someone is also getting roughly the same performance as someone else when there is 3…4 FPS difference. But for most of us these are a significant 10% of total FPS. Please see my new thread for details.

My intention is to provide a helpful tool for all of us who are trying to achieve the best possible MSFS experience, and also to diagnose for possible issues or performance regressions.

However, no one is forcing you to follow my suggestion.

Yeah I’ve read your post but I will stick to discovery flights, far simpler and accessible as not everyone wants to use software such as CapFrameX.

My suggestion is for for people to quickly and casually compare fps in a pretty constrained environment. The flight is already set up so you just do the flight, keep pretty much level and straight, say what you got from the dev mode fps counter. If you think a more in depth and detailed way is appropriate then that is your prerogative, personally I don’t think this deserves that amount of my time.

These are different approaches in their own right. Mine wasn’t intended for casual FPS checks but rather for reliable diagnosis. In the same way, watching devmode FPS counter won’t be any helpful for diagnosing low/min FPS either.

To follow up, these are specs and settings I discussed earlier in this thread"
COMPUTER:

Gigabyte Aorus Elite Z490
I7-10700
32 gb ram @3600
Evga RTX 3070 Ultra XC3
Hyperthreading - Enabled
Resizeable Bar - Enabled
Monitor: Sceptre 35" 3440x1440 100Hz
VR: HP Reverb G2

OS:

Windows 11 Pro 21H2 Build 22000.318

WINDOWS:

Game Mode - Off
HAGS - On
VRR - On

W11 MIXED REALITY: SEE SCREENSHOT

OpenXR DEVELOPER TOOL SETTINGS:

Lates Preview OpenXR Runtime - OFF
Custom Render Scaling - 150
Motion Reporojection - Disabled

NVIDIA CPL: SEE SCREENSHOT(S)

MSFS GRAPHICS - PC: SEE SCREENSHOT(S)

	VR:	SEE SCREENSHOT(S)

NOTE: Mostly about steady 33 FPS according to EVGA Precision X1 [similar to MSI Afterburner]. Cockpit resolution is quite good. Terrain resolution is good, but not nearly as sharp as monitor at 1440. In-cockpit scanning is smooth with some head jitter apparently associated with Windows 11 at the moment, but no stutter, black screens or frame drops.

PS: Sorry if screenshots may be out of order.