AMD 6900xt tuning and settings for VR in dcs. My optimal recipe.

[TED]

14 hours ago, TZeer said:

@ppokit

I have been using OTT up until now. But I guess I can use the official debug tool now.

I did some tweaking, and turned off ASW as you mentioned. Much better!

Running the track now I get a minimum of 60-63 FPS as I am coming around the hill close to ground inverted over the forrest.

Sitting on the carrier deck (training mission F/A-18) I have 72 FPS. Looks promising.

Even now, trying ASW makes to much artifacts.

yo @TED, With a G2, are you able to lock your frames at 45 FPS and get smooth performance? Like no jitter or stutters?

 

im running motion vector (auto - not always on) and motion smoothing (enable not forced) so that locks it pretty much at 45. I’m not totally convinced how much the Radeon software fps control works but I just leave it alone now on 45.

 

A few other observations from the last couple days. OpenFsr makes a huge difference for me. I’m running 0.9 sampling and 0.8 sharpening.

Force dx11 is critical in wmr.

clutter and grass detail in dcs has a much bigger effect than I expected. I now leave it at zero and barely notice a difference visually but big performance effect. If you have not tried adjusting this, set it to zero and see if it helps. 

Make sure supersampling is off in svr.

Nevada map is one of the less taxing but flying low over Vegas is a good test. With my settings now of 2x msaa and 4x antisopotric, 130% svr res, 100% per app setting it’s all very nice. 
 

im using 90hz on the headset. I tried 60 and while it did increase some performance the flicker was just too much for me for a minimal increase in performance. In fact in some situations I had more stutter with 60hz so it’s a no go for me. 

Edited February 11, 2022 by TED

[Picchio]

Hi, I've been reading through both this thread and the one created @nikoel for suggestions and comparisons.

I wanted to comment on one of your suggestions: by capping the framerate through the Radeon Software I assume you used the Chill option and set your min/max framerate to 45. While it does effectively work in limiting the framerate, in my system this option also has the detrimental effect of producing very high (20-ish) frametime values.

For comparison, with Chill set to off, my framerate is free to fluctuate between 40 and 90 fps but my frametime values are very low (5-16 ms, using the same settings as @nikoel).

So you might want to check your frametime values as well, perhaps I'm missing something else.

As far as I know, we're unable to lock the G2's framerate at 45 fps without activating motion smoothing... meaning that we're forced to deal with its artifacts. Am I correct?

Edited February 28, 2022 by Picchio

[DeltaMike]

^Consider getting a utility that separates out GPU time eg fps4VR.  

Since "chill" doesn't affect that, locking your frames at 45 (which is possible in adrenaline, in a couple of ways) may prevent dropped frames but it introduces "judder," which is a stroboscopic effect that occurs when you pan the camera. Noticable looking out to the side when you're flying low, or looking at the horizon when you do a barrel roll.   

So these are your options

1. Lock your frames at 45 and deal with judder.

2. Use motion vector and live with those artifacts

3. Let your fps run whatever it wants and deal with synch issues

4. Run DCS on your secret Los Alamos supercomputer at 90hz and enjoy!

All depends on what drives you the least crazy. Game is playable no matter which one you choose

[WipeUout]

Anyone tried the new adrenaline 22.2.3 driver? Apparently some FPS gain in 2D games.

[Picchio]

10 hours ago, DeltaMike said:

 

Thank you for your response. Well, yes... I was monitoring the performance through fpsVR and I clearly noticed the "judder" as the result of the increased frametimes induced by the driver (Chill) locking the maximum framerate... I just wanted to understand if you managed to lock it in a way that didn't have this side-effect.

I'm asking because with the Index you're able to throttle the framerate with a fixed value that doesn't have the issue above and I wondered if you had found a way to get a similar result without resorting to forced motion smoothing - and that really is a fantastic feature, if you ask me!

Well... I'm happy with what I have   Thank you anyway!

[DeltaMike]

1 hour ago, Picchio said:

 I just wanted to understand if you managed to lock it in a way that didn't have this side-effect.

 

No.  "Judder" as we are using the term here is not dependent on frame processing time, the same effect is seen in film. It has to do with angular velocity of objects in the environment vs FPS.  It's a stroboscopic effect.  

 

It's not really an issue when you're just sitting in SVR Home.  It's a big issue when you're flying at 500kts/500ft, or doing a snap roll in the A4E; and so we want FPS as high as we can get it, either by investing in hardware, lowering settings, or using synthetic frames ("ASW" or "motion vector")

Edited March 1, 2022 by DeltaMike

[TED]

Just to confirm, no I don’t use “chill” to lock fps. I wouldn’t recommend it for most cases in dcs. Never tried it actually. But I attempted to lock it with the adrenaline setting that attempts to target and caps fps. I’m not entirely sure it works to be honest but has no negative affects. 
I use motion vector and force dx11 and suffer very few if any artifacts. Certainly nothing that I notice. I’ve upped the power on the gpu though and min freq as well as overclock To 2650. 
I got rid of fpsvr as I actually found this to be inducing judder . Im getting i g good smooth performance even looking out the side in a harrier at low level. On some mp servers I do need to drop the resolution a bit in svr below 100%, but others it manages just fine. 
I’ve found in general it’s better to leave most of the adrenaline settings alone except overclock, power and maybe fan timings. 

[DeltaMike]

^agree completely. 

Did you get a chance to try the 3080 in MP?

[TED]

18 hours ago, DeltaMike said:

^agree completely. 

Did you get a chance to try the 3080 in MP?

Yep. Lots. I still have it actually but its packed away and will sell it when I get time. 

The 6900 was just way more stable. Some pros and cons and 3080 is a great card too. Bit less tuning required. But when 6900 is tuned on balance for me it's just more powerful and more reliable. Especially when it gets crowded on the server. 

Edited March 3, 2022 by TED

[DeltaMike]

Interesting question (in my mind at least) what makes a GPU good for DCS.  Always seemed like it was TMU count. Maybe VRAM does make a difference.  By that logic, it would be interesting to see how the 6900XT stacks up against the 3080ti in DCS.  

AMD definitely has its quirks.  Weirdly, AMD CPU's work like a champ right out of the box.  At this time I don't think it's possible to make the GPU work without fiddling with Adrenaline settings.  Which has its own idiosyncratic logic.

I really like the Ryzen series.   I never intentionally set out to buy an AMD GPU but it always seems to work out that way.  Like, literally always. 13 and counting.  That seems to apply to most of the people posting 6900xt results. "Somehow wound up with this thing and by gosh ... it works!"  

Some of it is the G2 itself. What is that thing supposed to do, anyways?  What is its niche?  How do we define "good" in this context? 

I think the ideal headset would have the FOV of a Pimax, and the central resolution of a G2.  Coupled with a GPU that has enough ass to run it at native resolution, AND apply MSAA.  That HMD may exist (eg, Pimax 8K) but I don't think the GPU exists. One fall back position would be, "G2 at native resolution."  Which has its limitations to be sure, but at least it is possible to run it that way without having to spend too much money. (And imo not running it any other way raises the question, why did one buy a G2 to begin with)

[TED]

I had a bit of time on my hands this week and decided to try some blue sky thinking and a few different approaches to the whole tuning. There has been som excellent contributions to this thread and I cant thank everyone enough for all the participation and great inputs. 

Following a very interesting post by @winghunter about combinations of MSAA and SSAA, I thought that this sounded like it made sense so gave it a go. 

I agree also with @Al-Azraq on the subject of openFSR. I have used openFSR for a while and indeed it does seem to give some performance gains in FPS, the image for me just is not that great and its an extra layer of processing. I find also the image is better at 80% resolution without it than 100% with FSR, so i removed it. 

The great posts by @nikoelhas also had me thinking. Ive been slavishly using MV and motion smoothing and whenever i turned them off I found it unacceptable jerky, but could not quite understand why others were getting good results and I was not. 

So today I threw out a bit of my previous routine and started over. I started by removing OpenFSR. Clearing out all the shaders, fxo and metashaders. I dropped SVR res to 70%, MV and Motion Smoothing off, DX11 force off. MSAA x4 and SSAA x1.5, AF off.

I downloaded and installed the latest AMD driver and software update.

I set up a mission from Batumi in the A10C. 

To my utter astonishment it was vastly better than I had been experiencing before. I was already very pleased with the performance, but like all of us enjoy trying to squeeze out more and tinkering.

The first and most noticeable effect was that with the combination of MSAA and SSAA the antialiasing is much better and the straight edges of wires, posts and buildings as well as any terrain outlines were much much smoother. Big plus.

My biggest surprise was that having now removed all motion smoothing I was expecting it to be very stuttery. I taxied out and took off with a solid 60fps and almost no stuttering at all. I flew a few profiles through the mountains to the east, along the valleys, over trees, over the mountains and over the city. Low level at around 300 ft for most of it. Even looking out the side was smooth. It was actually as smooth as it had been with motion vector and smoothing on, except the fps was higher, usually between 50 and 60.  Not quite believing my results I closed and restarted DCS a couple of times with the exact same results. And now of course absolutely no artifacts from the MV. 

To be honest I dont really know exactly why this is working so much better than I expected and I had tried most of these things at some point in the past but in isolation. It could be that the latest patch in DCS has improved some of the shaders and engine, as well as the latest AMD driver improving things. Most likely its a combination of several factors. 

After a few runs I could see that there was easily enough headroom to increase the SVR resolution to 80%, which gave an even better visual image with a bit of an FPS cost at around 5 fps giving be a steady around 45 to 55 at all times, of course a good bit higher when at altitude. It never dropped below 45. 

Edited March 5, 2022 by TED

[DeltaMike]

Agree re: shader mods.  We should probably steer new users away.  Partially to keep their heads from exploding, and partially because something (?WMR ?AMD) is already doing a halfway decent job of scaling.  

MSAA looks great, especially on the NTTR map, but keep in mind, if you undersample and use MSAA it cuts your spotting distance in half.

As an alternative, consider 100/130/1.0, MSAA off, steam advanced supersampling on.  Doesn't eliminate the shimmering but I think it helps a little, and you should get higher frame rates besides.

AF maxed out either way 

ETA:

RE: SSAA.  Another knob to twiddle.  That makes four by my count.  Way I look at it, there are two numbers.  A) How many pixels are rendered B) How many pixels are displayed. As long as A>=B, which I think gives you the best central vision, I'm a happy camper.  Which net of everything, you are doing, you're rendering 105% no?  Not sure there's any magic to be had, twiddling one knob as opposed to another.  Not much, anyway. 

Found out I can run 100/100/1.0 with MSAAx2 if I turn shadows off.  Kinda surprised me.  Am an even happier camper now.  Shadows are pretty but they flog my GPU and don't contribute to game play.  Terrain shadows are buggy anyway. Prob for the best, if I ever get back into MP. 

Remain unconvinced forcing DX11, or not, does anything

Edited March 5, 2022 by DeltaMike

[TED]

 Should just add that all these settings are a bit map and aircraft dependent too and may need a bit of time tuning. My approach for now is msaa x2, SSAA x1.5, most dcs settings high or maxed except detail on medium. Mv and motion smooth off.

from there I only adjust the resolution depending on the map. Last one I did in caucuses and a10c I was getting up to 80 fps at times. More usually 60-70 so I bumped msaa to x4. Looked fantastic. This was with svr res at 85%.

basically any tuning is small adjustments of these. A heavy mission with lots of ai will probably need to drop to msaa x2. 
 

I agree on shadows. Although I like to have the cockpit lighting effects and shadows in the cockpit. Terrain shadows - meh. Flat is fine. 

 

[nikoel]

On 3/5/2022 at 10:23 PM, DeltaMike said:

Agree re: shader mods.  We should probably steer new users away.  Partially to keep their heads from exploding, and partially because something (?WMR ?AMD) is already doing a halfway decent job of scaling.  

MSAA looks great, especially on the NTTR map, but keep in mind, if you undersample and use MSAA it cuts your spotting distance in half.

As an alternative, consider 100/130/1.0, MSAA off, steam advanced supersampling on.  Doesn't eliminate the shimmering but I think it helps a little, and you should get higher frame rates besides.

AF maxed out either way 

ETA:

RE: SSAA.  Another knob to twiddle.  That makes four by my count.  Way I look at it, there are two numbers.  A) How many pixels are rendered B) How many pixels are displayed. As long as A>=B, which I think gives you the best central vision, I'm a happy camper.  Which net of everything, you are doing, you're rendering 105% no?  Not sure there's any magic to be had, twiddling one knob as opposed to another.  Not much, anyway. 

Found out I can run 100/100/1.0 with MSAAx2 if I turn shadows off.  Kinda surprised me.  Am an even happier camper now.  Shadows are pretty but they flog my GPU and don't contribute to game play.  Terrain shadows are buggy anyway. Prob for the best, if I ever get back into MP. 

Remain unconvinced forcing DX11, or not, does anything

 

This is the way. Although I trade off MSAA for the [Low] Shadows

I learned to embrace the shimmers and a few jaggered edges here or there - or at very least accept them for the sublime clarity and performance

Next step is to get yourself a Viggen and enjoy the signature Mach speeds 10meters off the deck whilst doing barrel rolls as you contour your ikea cluster meat-bomblets into an orphanage - all in perfect smoothness

Edited March 6, 2022 by nikoel

[TED]

@nikoelcan u just quickly confirm - u r running 100/130/1.0 for svr res, per app and pd.

Then no msaa or af. 

And you find the image quality OK in terms of sharpness and detail? 

Edited March 6, 2022 by TED

[edmuss]

I'm on 100/120/1.0 fsr 0.75 and I still get some aliasing (particularly A10C UFC buttons) but it is minimal. On Caucasus I can turn on MSAA and generally have sub 14ms frametimes. Can run Syria with MSAA enabled but smoothing is generally constantly engaged.

I've not tested to see if 120 with 0.8 is less aliased than with 120/0.75 with MSAA. Despite the performance hit, the MSAA almost completely kills the distant shimmer in conjunction with the supersampled fsr.

edit: just carried out some rough testing, fsr 0.75 is still my sweetspot and 100/120/1.0 with MSAA is about the same performance as 100/150/1.0 without MSAA.  Running steamvr@150% gives really good clarity on the hud/mfd text and whilst there is still some aliasing it's a good deal sharper than running MSAA.  General flying around Caucasus angels10 on the A10C easy instant action gives about 13-14ms.

Edited March 6, 2022 by edmuss

[TED]

1 hour ago, edmuss said:

I'm on 100/120/1.0 fsr 0.75 and I still get some aliasing (particularly A10C UFC buttons) but it is minimal. On Caucasus I can turn on MSAA and generally have sub 14ms frametimes. Can run Syria with MSAA enabled but smoothing is generally constantly engaged.

I've not tested to see if 120 with 0.8 is less aliased than with 120/0.75 with MSAA. Despite the performance hit, the MSAA almost completely kills the distant shimmer in conjunction with the supersampled fsr.

edit: just carried out some rough testing, fsr 0.75 is still my sweetspot and 100/120/1.0 with MSAA is about the same performance as 100/150/1.0 without MSAA.  Running steamvr@150% gives really good clarity on the hud/mfd text and whilst there is still some aliasing it's a good deal sharper than running MSAA.  General flying around Caucasus angels10 on the A10C easy instant action gives about 13-14ms.

 

Thanks. I just tried some of that too but for now I’m done with openFSR as it just deteriorates the image too much for me. 85% svr res without fsr is considerably better than running 120% svr with 0.77 fsr imo.

I was testing a bit with no fsr and 85% svr res, msaa and SSAA on. Generally a more agreeable performance and image for my flying around generally in a10 at  3-400’ caucus, Syria and Nevada. Interestingly I’m able to add mv and motion smoothing if needed with a slight softening of the image but I don’t get any artifacts now. 

what on earth are you doing at angels 10 in an a10c? .

it’s all very personal I guess. My preference is for smooth performance at low level with the best image possible. Basically no shimmers and no stutters and for that I’m prepared to sacrifice a little sharpness and detail. 

[edmuss]

1 hour ago, TED said:

what on earth are you doing at angels 10 in an a10c? .

Admiring the scenery

Oh to have a halo GPU! My midrange one is holding up ok but does need to be propped up a bit with the vrperfkit.

Rooftops over Beirut gives me 20-23ms which is more than enough for 60Hz. This is generally my heaviest GPU workload baseline so if my tuning keeps it ok there then it's generally great elsewhere.

[TED]

36 minutes ago, edmuss said:

Admiring the scenery

Oh to have a halo GPU! My midrange one is holding up ok but does need to be propped up a bit with the vrperfkit.

Rooftops over Beirut gives me 20-23ms which is more than enough for 60Hz. This is generally my heaviest GPU workload baseline so if my tuning keeps it ok there then it's generally great elsewhere.

Have u updated to latest amd driver yet? In my case I noticed quite a bit of real gains. Make sure u clear out your fxo and metashaders in between adjustments btw too. 
I am just testing now with no fsr at 100/125/1.0. Surprisingly good results and at that resolution don’t need msaa at all. 

[edmuss]

I hope not, I've got a 3070

I'm down on horsepower and vram capacity compared to the 6900xt and the AMD specific tuning is obviously null and void but the headset, shader mod, upscaling and DCS tweaking are all pertinent.

I find that fsr doesn't significantly affect image quality unless I get below 0.7 at which point it gets muddy. If I had spare horsepower (a 3080/90 6800/900xt) then I would probably be running 150% with MSAA turned on to get the clarity and smoothness.

[TED]

1 hour ago, edmuss said:

I hope not, I've got a 3070

I'm down on horsepower and vram capacity compared to the 6900xt and the AMD specific tuning is obviously null and void but the headset, shader mod, upscaling and DCS tweaking are all pertinent.

I find that fsr doesn't significantly affect image quality unless I get below 0.7 at which point it gets muddy. If I had spare horsepower (a 3080/90 6800/900xt) then I would probably be running 150% with MSAA turned on to get the clarity and smoothness.

Ahhh my bad. I assumed as you were posting on this thread you had a 6900.

2 totally different animals. Tuning the nvidia and amd require quite different approaches. I still have both with the 3080 still in my study. 
With either there is definitely some merit in maxing the resolution then there is very little requirement for msaa. I’m playing with that at the moment and feel it gives a better result. No fsr. No msaa. No af but max resolution. Unfortunately everything you add on top softens the image, especially msaa. For me right now on the g2 with resolution at 100/125 the shimmering is minimal and actually no worse than 90% with msaa. 
So several different routes to the results depending on what your priorities are. 
I am extremely happy with it all right now and dcs just looks goddam awesome in the g2 now I u derstand all the tweaking a bit better. This is the best I’ve seen it. I suspect the latest OB patch has improved some things too and the drivers keep adding small improvements so everything is going in a very positive direction. The last month all of my performance and image quality has improved significantly. 

[edmuss]

Indeed the last OB gave me some good increases in vr performance. I fully understand what you mean about the softening of the image though, oddly enough even at 120% with Nvidia I still get aliasing around the HUD frame and UFC. At 150% it's pretty minimal and oddly the performance hit for supersampling doesn't seem that high.

I'll have a play with your thoughts on disabling fsr and see how it handles it, I think there's about 3ms on the table there for me though, which is a huge amount. I'll keep the vrperfkit installed because the foveated rendering does make a big impact and with the edge blurriness if the G2 Fresnel lenses is not visible in the headset.

edit: so my guess of 3ms was closer to 4-5ms performance uplift due to the vrperfkit fsr. 120% and fsr off looked worse for me but about 1ms slower than 150% and fsr 0.75; it's still playable but just on the threshold of engaging smoothing so it jumps about a bit and artifacts more as it switches.

Edited March 7, 2022 by edmuss

[nikoel]

21 hours ago, TED said:

@nikoelcan u just quickly confirm - u r running 100/130/1.0 for svr res, per app and pd.

Then no msaa or af. 

And you find the image quality OK in terms of sharpness and detail? 

 

I will give you my settings but keep in mind that I am using FSR and I strongly believe that it's significantly better than steam's upscaling. I am curious did you by any chance do what someone else in my thread did and accidentally used CAS instead of FSR? CAS has a very crappy upscaler - either way FSR should give at the very least the same performance as Steams upscaler. I do not use any sharpening filter, not even in FSR settings

I seem to settle somewhere around 110-130% with FSR at 0.7-0.8 range. This is where the sweet spot is, somewhere. I am also CPU limited right now and have my sights lasered on the 5800X3D so I haven't bothered tuning for my CPU too much - wasted time

Keep in mind this means that DCS engine is still rendering 77%-104% 

I am not running the MSAA, AA or any of the tick boxes. I have the shadows on low and flat. I am using the technique above basically what @edmussand @DeltaMikeare doing with some variations - but the overall idea is the same, I am antialiasing by rendering a super-resolution. I have written in detail of why it works for me, and it's shortcomings

The technique above will give you the sharpest, most detailed image. It does this because antialiasing blurs the image by design (a simplified truth) this is why I am able to read the map on the F/A-18's HSI page (It's still crap though, no way around it). I have to accept the shimmer and some jagged edges here or there, but I have no doubt that I have the clearest image. DeltaMike has definitely less shimmer but a slightly softer image and one with no shadows

I will screenshot when I get a chance but from memory as I tweaked them a bit more and I am very happy with how things are looking for how limited the DCS egine is - (show me an other game where after rendering over 7k combined res you're still CPU limited)

PD 1.0

SS VR 110-130% Depending on how I feel on the day (eg lowering the setting on Syria/F-14, bringing them up for F/A-18 & Balkans)

FSR - 0.7 - 0.8 with never any sharpening 

I am also using the Simplex shader - gave me 2ms and no noticeable loss of visuals as was recommended to me ¯\_(ツ)_/¯ 

FPS-VR GPU Timing is around 8-12ms
CPU is not great 11-15ms. Such is life 

 

[EDIT - Screenshot Attached]

Spoiler

# Upscaling: render the game at a lower resolution (thus saving performance),
# then upscale the image to the target resolution to regain some of the lost
# visual fidelity.
upscaling:
  # enable (true) or disable (false) upscaling
  enabled: true
  # method to use for upscaling. Available options (all of them work on all GPUs):
  # - fsr (AMD FidelityFX Super Resolution)
  # - nis (NVIDIA Image Scaling)
  # - cas (AMD FidelityFX Contrast Adaptive Sharpening)
  method: fsr
  # control how much the render resolution is lowered. The renderScale factor is
  # applied to both width and height. So if renderScale is set to 0.5 and you
  # have a resolution of 2000x2000 configured in SteamVR, the resulting render
  # resolution is 1000x1000.
  # NOTE: this is different from how render scale works in SteamVR! A SteamVR
  # render scale of 0.5 would be equivalent to renderScale 0.707 in this mod!
  renderScale: 0.77
  # configure how much the image is sharpened during upscaling.
  # This parameter works differently for each of the upscaling methods, so you
  # will have to tweak it after you have chosen your preferred upscaling method.
  sharpness: 0.0
  # Performance optimization: only apply the (more expensive) upscaling method
  # to an inner area of the rendered image and use cheaper bilinear sampling on
  # the rest of the image. The radius parameter determines how large the area
  # with the more expensive upscaling is. Upscaling happens within a circle
  # centered at the projection centre of the eyes. You can use debugMode (below)
  # to visualize the size of the circle.
  # Note: to disable this optimization entirely, choose an arbitrary high value
  # (e.g. 100) for the radius.
  radius: 100
  # when enables, applies a MIP bias to texture sampling in the game. This will
  # make the game treat texture lookups as if it were rendering at the higher
  # target resolution, which can improve image quality a little bit. However,
  # it can also cause render artifacts in rare circumstances. So if you experience
  # issues, you may want to turn this off.
  applyMipBias: false

# Fixed foveated rendering: continue rendering the center of the image at full
# resolution, but drop the resolution when going to the edges of the image.
# There are four rings whose radii you can configure below. The inner ring/circle
# is the area that's rendered at full resolution and reaches from the center to innerRadius.
# The second ring reaches from innerRadius to midRadius and is rendered at half resolution.
# The third ring reaches from midRadius to outerRadius and is rendered at 1/4th resolution.
# The final fourth ring reaches from outerRadius to the edges of the image and is rendered
# at 1/16th resolution.
# Fixed foveated rendering is achieved with Variable Rate Shading. This technique is only
# available on NVIDIA RTX and GTX 16xx cards.
fixedFoveated:
  # enable (true) or disable (false) fixed foveated rendering
  enabled: false
  # configure the end of the inner circle, which is the area that will be rendered at full resolution
  innerRadius: 0.6
  # configure the end of the second ring, which will be rendered at half resolution
  midRadius: 0.8
  # configure the end of the third ring, which will be rendered at 1/4th resolution
  outerRadius: 1.0
  # the remainder of the image will be rendered at 1/16th resolution

# Enabling debugMode will visualize the radius to which upscaling is applied (see above).
# It will also output additional log messages and regularly report how much GPU frame time
# the post-processing costs.
debugMode: false

# Hotkeys allow you to modify certain settings of the mod on the fly, which is useful
# for direct comparsions inside the headset. Note that any changes you make via hotkeys
# are not currently persisted in the config file and will reset to the values in the
# config file when you next launch the game.
hotkeys:
  # enable or disable hotkeys; if they cause conflicts with ingame hotkeys, you can either
  # configure them to different keys or just turn them off
  enabled: false
  # toggles debugMode
  toggleDebugMode: ["ctrl", "f1"]
  # cycle through the available upscaling methods
  cycleUpscalingMethod: ["ctrl", "f2"]
  # increase the upscaling circle's radius (see above) by 0.05
  increaseUpscalingRadius: ["ctrl", "f3"]
  # decrease the upscaling circle's radius (see above) by 0.05
  decreaseUpscalingRadius: ["ctrl", "f4"]
  # increase the upscaling sharpness (see above) by 0.05
  increaseUpscalingSharpness: ["ctrl", "f5"]
  # decrease the upscaling sharpness (see above) by 0.05
  decreaseUpscalingSharpness: ["ctrl", "f6"]
  # toggle the application of MIP bias (see above)
  toggleUpscalingApplyMipBias: ["ctrl", "f7"]
  # take a screen grab of the final (post-processed, upscaled) image.
  # The screen grab is stored as a dds file next to the DLL.
  captureOutput: ["ctrl", "f8"]
  # toggle fixed foveated rendering
  toggleFixedFoveated: ["alt", "f1"]

 

 

Edited March 7, 2022 by nikoel

[edmuss]

Only to add to @nikoelabove, I'm also with all sharpening off. All textures high, cockpit shadows high, AF 16x, terrain shadows flat, visibility/water/clouds on medium. Depth of field and the block of three in the bottom corner are off and the tree/building sliders are all maxed out.

Also using kegetys shader mod which is faster than the simplex but doesn't pass IC for multiplayer.

[WipeUout]

15 hours ago, TED said:

The last month all of my performance and image quality has improved significantly. 

Same here, I used to have to turn MSAA off as soon as things were getting busy but now, I can keep it.  I also find myself not using reprojection as often as before, my rig is keeping up with 80+ FPS even in MP most of the time.  The pimax 5kS has a minimum 90hz refresh rate and if FPS goes below 80, then it becomes very stuttery.

What I see now is my CPU being the bottleneck.  I am considering switching to a Rysen 5900X or i7-12700KF.  If only I knew when the multi-core engine will be available, that would help in deciding which one to pick.  I have the impression that Rysen will still be a better choice for the 6900XT in a multi-core engine compared to an Intel i7 setup.

Edited March 7, 2022 by WipeUout
typos and clarifications