My 3090 Settings for my G2 (2.7.11)

[speed-of-heat]

7 minutes ago, zildac said:

Just curious, where people are running FSR (1.2 in my case) are you running steam SS at 100% when running these benchmarks?

For the purpose of testing i run at 80% in CRM to provide a consistent set of numbers, in practice given the frame times i run at 100% operationally 

[dureiken]

1 minute ago, speed-of-heat said:

For the purpose of testing i run at 80% in CRM to provide a consistent set of numbers, in practice given the frame times i run at 100% operationally 

yep I understood that, my question was more : why your 100% isn't 2160x2160 as for the G1 ?

 

[speed-of-heat]

12 minutes ago, dureiken said:

@speed-of-heat

I don't understand resolution in your guide.

 Custom Resolution Modifier for DCS to 80% 2836x2768

With G2, native resolution isn't 2160*2 = 4320 and 2160 high ?

I have this for G1 :

Render resolution per eye: 2200x2160(by SteamVR settings, Max.) (HMD driver recommended: 2201x2158)

Thanks

No.

1) resolution in steam VR is always per eye rather than the combined panorama of both eyes 

2) one of the main differences between G1 and G2 is that the G2 operates at 150% of it physical resolution to accommodate barrel distortion, which gives the setting 3168 x 3096 at 100% in Steam's Resolution per Eye for the G2

4 minutes ago, dureiken said:

yep I understood that, my question was more : why your 100% isn't 2160x2160 as for the G1 ?

 

you will see in my post i was answering @zildac's question not yours , which i answered above here

Edited October 27, 2021 by speed-of-heat

[dureiken]

5 minutes ago, speed-of-heat said:

No.

1) resolution in steam VR is always per eye rather than the combined panorama of both eyes 

2) one of the main differences between G1 and G2 is that the G2 operates at 150% of it physical resolution to accommodate barrel distortion, which gives the setting 3168 x 3096 at 100% in Steam's Resolution per Eye for the G2

you will see in my post i was answering @zildac's question not yours , which i answered above here

 

oh ok sorry. So when you usually play at 100% of CRM, it's like if I was at 150% ?

 

[speed-of-heat]

yes

[dureiken]

ok so all my comparaison with you are false, I won't be anymore at 72 fps lol

[zildac]

14 minutes ago, speed-of-heat said:

For the purpose of testing i run at 80% in CRM to provide a consistent set of numbers, in practice given the frame times i run at 100% operationally 

Thanks @speed-of-heat I'm going to try some benchmark runs just out of curiosity.

[speed-of-heat]

11 minutes ago, dureiken said:

ok so all my comparaison with you are false, I won't be anymore at 72 fps lol

Yes it came as a bit of a shock to everyone who thought we were going to get better or similar frame times with the G2.  But I am pretty clear about my pixels per eye in my guide.

 

9 minutes ago, zildac said:

Thanks @speed-of-heat I'm going to try some benchmark runs just out of curiosity.

I look forward to the results m8

Edited October 27, 2021 by speed-of-heat

[ShaunX]

52 minutes ago, speed-of-heat said:

For the purpose of testing i run at 80% in CRM to provide a consistent set of numbers, in practice given the frame times i run at 100% operationally 

80% for me too. I fly at 80% too

[zildac]

So very quickly ran through 3 tests. I also basically chopped the first and last 20 or so rows off the resultant CSV as there was some spikes in there as I guess is to be expected (no idea if there is a more scientific way of doing this). Results, in a somewhat rough format, were as follows:

       FPS                 GPU FT              CPU FT              GPU Util%

      80.91873169    10.97153954    7.537679868     91.87444289 - 2MSAA HIGH Clouds (My "normal" settings")

      71.03609763    12.597221        7.681767387      91.62711366 - 4MSAA HIGH Clouds

      81.08613467    10.94766127    7.434528399    91.58943716 - 2MSAA ULTRA Clouds

Other notable settings:

Water - HIGH

Shadows - MEDIUM

Terrain Obj Shadows - FLAT

Mods in use

@speed-of-heat SOH CW Shaders (2.7.7.14727)
Mustang 256 Shaders (Fixed for 2.77 - 0.96B)
Shadows MOD (as listed in @speed-of-heat OP - "Shadows Reduced Frame Rate Impact"
Bartheks Caucasus Redone 2021
Taz Trees - Medium - Low
Taz Caucasus runways
OpenFSR 1.2 (Steam SS set to 80% for all tests)

 

One of the most interesting things for me is the significant increase in GPU FT with 4xMSAA but without an increase in corresponding GPU usage, I''m guessing this must just be the result of averaging as that doens't seem logical, although it is obviously common sense and we all know 4xMSAA has a big impact. Also, Ultra clouds actually seemed more performant! That said there didn't appear to be any real clouds in the track so perhaps, so I wasn't expecting any real negative impact.

Going to try a comparison with sharper eye enabled this evening as I do still use it.

I can see this becoming somewhat of a compulsion

 

Edited October 27, 2021 by zildac

[speed-of-heat]

Its always best not to over analyse this stuff... you can end up vanishing up your own ar$e if you are not careful... good enough is good enough 

[Burt]

Absolutely ! Well said and so true. Don't play 2 games play 1.

Edited October 27, 2021 by Burt

[zildac]

2 minutes ago, speed-of-heat said:

Its always best not to over analyse this stuff... you can end up vanishing up your own ar$e if you are not careful... good enough is good enough 

Agreed

[ShaunX]

1 hour ago, speed-of-heat said:

Its always best not to over analyse this stuff... you can end up vanishing up your own ar$e if you are not careful... good enough is good enough 

LOL. Wise words, though maybe too late for me this week !

Edited October 27, 2021 by ShaunX

[Digitalvole]

So you guys have fixed the clear water shader mod for 2.7.7? If so, what fantastic news, (good birthday present!) thank you, thank you, thank you!

Honestly, where would we be with out some members of this community? Sterling effort. 

[dureiken]

Hi, can guys here launch the SOH276 track test with these settings, without mods, no FSR, no shaders, to compare ?

Here are my results with SS100, and SS200 (same resolution that your G2à for a reverb G1 and 6900XT.

SS100%

fpsVR Report:
App: DCS HMD: HP Reverb VR Headset VR1000-2xxx0 (90.000 Hz, IPD 63.6)
GPU: AMD Radeon RX 6900 XT (27.20.22025.1006, Tavg 57.8, Tmax 62) CPU: Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz (Tavg 50.5, Tmax 62)
Delivered fps: 56.95  Duration: 1.5min. Headset was active: 100%
GPU Frametimes:
Median: 12.9 ms
99th percentile: 14.9 ms
99.9th percentile: 15.5 ms
frametime <11.1ms(vsync): 11%
CPU frametime:
Median: 9.3 ms
99th percentile: 12 ms
99.9th percentile: 12.9 ms
frametime <11.1ms(vsync): 96.7%
Reprojection Ratio: 48.7% (for Index/Vive/VivePro headsets only)
Dropped frames: 1 or 0.0% (for Index/Vive/VivePro headsets only)
Max. SteamVR SS: 100%
Render resolution per eye: 2200x2160(by SteamVR settings, Max.) (HMD driver recommended: 2201x2158)

SS200%

fpsVR Report:
App: DCS HMD: HP Reverb VR Headset VR1000-2xxx0 (90.000 Hz, IPD 63.6)
GPU: AMD Radeon RX 6900 XT (27.20.22025.1006, Tavg 58.6, Tmax 64) CPU: Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz (Tavg 48.8, Tmax 66)
Delivered fps: 41.57  Duration: 1.5min. Headset was active: 100%
GPU Frametimes:
Median: 19.7 ms
99th percentile: 22.9 ms
99.9th percentile: 23.5 ms
frametime <11.1ms(vsync): 0.1%
CPU frametime:
Median: 9.8 ms
99th percentile: 12.3 ms
99.9th percentile: 13.2 ms
frametime <11.1ms(vsync): 92.4%
Reprojection Ratio: 59.4% (for Index/Vive/VivePro headsets only)
Dropped frames: 2 or 0.0% (for Index/Vive/VivePro headsets only)
Max. SteamVR SS: 199%
Render resolution per eye: 3104x3044(by SteamVR settings, Max.) (HMD driver recommended: 2201x2158)

Thanks

 

[dureiken]

Hi,

is there a way to remove tree shadow and keep buildings ?

thanks

[speed-of-heat]

no not that i know off,  you could do something similar, which was keep terrain shadows but turn object shadows off but that, options was removed with 2.7.7 when they encrypted the metashaders directory 

[Defbond]

@speed-of-heat I have followed your settings as closely as possible but at around 80% SteamVR SS on my G2 I can't spot any contacts unless I am zoomed in, best case is around 5-10km. When I set the resolution to 20% I can spot targets from many many miles away, sometimes as far as 50km, but ofc the game looks like Minecraft. I usually play on cold war servers so spotting is extremely important especially since we don't use radars in order to avoid being detected. Any thoughts on how to have a nice resolution but also decent spotting? It's too ugly to play at 20% but I also want to be competitive. 

[speed-of-heat]

not sure what to say…

Edited October 30, 2021 by speed-of-heat

[Defbond]

12 hours ago, speed-of-heat said:

not sure what to say…

I know you think I sound crazy but I tried it today and now I can spot fine even with 80% SS,  I don't know what I did and frankly I don't care I'm just glad it's working so I can enjoy the game. Just out of curiosity could you estimate at what distance you can spot targets in your G2 without tags?

[Weebilly1211]

Thank you SOH for this topic, your OP and all the testing has greatly improved my VR experience (without MODS). Im currently running the lastest DCS version 2.7.7. Can i ask what MODS you are currently running operationally as I dont want to install a mods that might mess everything up.

 

Thank you again

[speed-of-heat]

1 hour ago, Defbond said:

I know you think I sound crazy but I tried it today and now I can spot fine even with 80% SS,  I don't know what I did and frankly I don't care I'm just glad it's working so I can enjoy the game. Just out of curiosity could you estimate at what distance you can spot targets in your G2 without tags?

glad its now working fine for you.

1 hour ago, Weebilly1211 said:

Thank you SOH for this topic, your OP and all the testing has greatly improved my VR experience (without MODS). Im currently running the lastest DCS version 2.7.7. Can i ask what MODS you are currently running operationally as I dont want to install a mods that might mess everything up.

 

Thank you again

minor changes really, but pretty much as as listed in my OP i tend to use the "latest" version of everything so I'm on FSR 1.3  which changed this morning and i havent had a chance to update it yet for example

[1Wolf]

@speed-of-heat

Thanks SOOOO much for going to all the work, and effort of all of that testing, sharing your methodology, sharing your results, and above-all-else...keeping this thread alive and keeping them updated.

Only a few months ago I built a new system so that I could run in VR, picked up a Reverb G2, and got back into DCS again when I'd been away a while.  You're thread helped me immensely.  For me at least, the G2 and DCS were a bit tricky to get set up to run together correctly and a bit difficult to understand.  It was confusing as hell for me.  Once I was up and running I used the methodology and tools you shared to create my own benchmarks so that I could find what worked for me and my system.

Thank you again for all the hard work.

Question regarding FSR:  I don't tend to run many mods but the performance that people are claiming of FSR seemed worth the effort for me to give it a try.  I'm a novice at understanding the way things are rendered in VR.  So after reading the GitHub entry for it online and the readme information,  I just wanted to see if I'm understanding this correctly.  So [b]without[/b] FSR you normally run SteamVR CRM for DCS at 80%.  [b]With[/b] FSR you run Steam VR CRM for DCS at 100% but have the openvr_mod.cfg set to a RenderScale of 0.8.  Isn't that pretty much the same thing as running SteamVR CRM for DCS at 80% and not running FSR?  Is it that FSR does the same thing that SteamVR would be doing...only better or more efficiently?  What would happen if you set SteamVR for DCS to 100% and the RenderScale to 1.0?

I tried using your openvr_mod.cfg settings exactly...setting RenderScale to 0.8, Sharpness to 0, and Radius to 0.75.  I didn't run a scientific test yet as I was just trying it out but the performance does seem better.  However, things do look a little more blurry.  I'm testing this in the Viper and the instruments look just a little bit more blurry, looking at vehicle targets with the targeting pod camera look just a little more blurry.  The terrain details in the Nevada desert look just a little more smooth and blurry.  Don't get me wrong, nothing is "too blurry", its very subtle.  I just noticed that things seem a little more clear without the FSR.

Is that due to the "Sharpness" setting in openvr_mod.cfg?  Most everyone here complained of shimmering or shining with the Sharpness above 0 so I was afraid to set it higher.  What might you recommend to try?  Or maybe the slight blurryness is just a side effect of FSR and perhaps that mod just isn't for me.

I'm sure you might say just to experiment and try things out but I happen to be one of those people that I start to go back and forth trying things and pretty soon, no matter what I do, they all look the same and I end up going down that rabbit hole of tweaking and never stopping.  So I just thought I'd get your opinions and advice.

Thank you again!

[DucS2r]

We are running SOH ragged these days because he is so helpful, but I’ll throw in my two cents while you wait for his answer. I noticed a slight degradation in clarity with FSR as well, with the same settings. However I thought it was worth it given the improvement in performance. Unlike SOH, I also use Reshade using the settings SOH tried. I find it adds enough sharpening that I am happy with the combined result. I also run Kegety’s shaders.