dburne Posted December 1, 2019 Posted December 1, 2019 It definitely works, no doubt of that. And it's been a while since I did my comparing with custom res vs oculus pd, but from my guesstimate is 208% in Steam is approximately 1.5x in PD and 178% is in the neighborhood of 1.3. Now that can't really compare with the Index all that well, I have never even seen one since Valve still refuses to sell hardware in the EEA. And well my opinion and disappointment of the 5k+ is hardly any secret. I also suspect desired PD is highly individual in the S since we can't adjust lenses in it, it varies between each anyone how each eye hits the lenses. So far I have all only done my testing in the F18, I'll give the viper ago we well but so far I'm not having to much of an issue with performance and pd at 1.3. Yeah I find PD of 1.3 about the sweet spot for me in quality/performance in DCS. Don B EVGA Z390 Dark MB | i9 9900k CPU @ 5.1 GHz | Gigabyte 4090 OC | 64 GB Corsair Vengeance 3200 MHz CL16 | Corsair H150i Pro Cooler |Virpil CM3 Stick w/ Alpha Prime Grip 200mm ext| Virpil CM3 Throttle | VPC Rotor TCS Base w/ Alpha-L Grip| Point Control V2|Varjo Aero|
VirusAM Posted December 1, 2019 Posted December 1, 2019 I did some more A/B testing and indeed SSAA doesn't work at all. So I'll admit to that one, ;) Also been testing a little more but there it's as if the anisotropic setting doesn't kick in from the profile inspector, past 2 am here do getting to tired for vr now so will test more later. So far the shimmering is far to bad for me with a lod bias of less than -0.375 and it's still a far cry in fidelity as what I had with supersampling of 1.3, I'm testing this with a PD setting of 1x And regardless of of running with 1x PD, I drop to 40 when looking over a city scape or any other some what busy place. So I don't really see much point in taking a drop in pq when I won't get any big performance increase when it matters. And run PD at 1.3. I am having similar experiences. I tried these recommended settings, but i think something is wrong....I disabled AF and MSAA in dcs and with a PD of 1.2 (oculus tray tool) but then i saw a bad visul...a lot of aliasing and shimmering. Then resetted nvidia profiler to default settings and put back AF to x16 and msaa x2 in DCS and a PD of 1.4 and i gained back a lot of the lost visual claritiy. Probably i am doing something wrong and the nvidia profiler settings won’t kick in so with no msaa and lower PD i lose the quality....someone know what could be? R7-5800X3D 64GB RTX-4090 LG-38GN950 N/A Realsimulator FFSB MKII Ultra, VKB Stecs Max, Winwing F-16EX Throttle, Winwing Orion (Skywalker) Pedals, Razer Tartarus V2 SpeedMaster Flight Seat, JetSeat
Bob_Bushman Posted December 1, 2019 Posted December 1, 2019 (edited) I am having similar experiences. I tried these recommended settings, but i think something is wrong....I disabled AF and MSAA in dcs and with a PD of 1.2 (oculus tray tool) but then i saw a bad visul...a lot of aliasing and shimmering. Then resetted nvidia profiler to default settings and put back AF to x16 and msaa x2 in DCS and a PD of 1.4 and i gained back a lot of the lost visual claritiy. Probably i am doing something wrong and the nvidia profiler settings won’t kick in so with no msaa and lower PD i lose the quality....someone know what could be? Yeah, I'm not sure what could be the cause if some settings doesn't kick in or not. I have not only been restarting DCS but I have also been restarting the oculus service between testing applied changes. So short of a fresh reboot between applying each setting I'm at a loss. Hmm, checking drivers, seesms I'm on 441.20, and current are 441.41 Updated drivers, Testing some more, and I will have to conclude for now: The only thing the inspector does for me, is to override what usual tweaks Nvidia/Oculus is done from before, making things generally look like garbage in and out of DCS. Anisotropic filtering from the inspector never kicks in, but forces it to off for the ingame setting, even when set to allow application control. AA settting, regardless of mode, does nothing. LOD BIAS, no, can't really see that does anything at all, at least nothing I would consider beneficial. Pretty much anything done with the inspector leaves DCS looking as if it's running very low texture detail, for example, any detail texture on the cat's, skid marked section on the runway, all gone. and all textures in the cockpit generally blurrier with more aliasing artefacts. Even when upping PD back up, which removes any performance increase from this. I'm back to defaults, and PD 1.3, anything suggested, and bunch of other settings suggested in this thread was a huge drop in image quality with hardly any performance gain whatsoever. Edited December 1, 2019 by Bob_Bushman i7 8700k @ 4.7, 32GB 2900Mhz, 1080ti, CV1 Virpil MT-50\Delta, MFG Crosswind, Warthog Throttle, Virptil Mongoost-50 throttle.
Sniper175 Posted December 1, 2019 Posted December 1, 2019 Weird, nvidia inspector works magic for me. Textures look amazing with 16x forced in inspector and high quality texture filtering. Latest drivers for me also I7-8700 @5GHZ, 32GB 3000MHZ RAM, 1080TI, Rift S, ODYSSEY +. SSD DRIVES, WIN10
VirusAM Posted December 1, 2019 Posted December 1, 2019 Yeah, I'm not sure what could be the cause if some settings doesn't kick in or not. I have not only been restarting DCS but I have also been restarting the oculus service between testing applied changes. So short of a fresh reboot between applying each setting I'm at a loss. Hmm, checking drivers, seesms I'm on 441.20, and current are 441.41 Updated drivers, Testing some more, and I will have to conclude for now: The only thing the inspector does for me, is to override what usual tweaks Nvidia/Oculus is done from before, making things generally look like garbage in and out of DCS. Anisotropic filtering from the inspector never kicks in, but forces it to off for the ingame setting, even when set to allow application control. AA settting, regardless of mode, does nothing. LOD BIAS, no, can't really see that does anything at all, at least nothing I would consider beneficial. Pretty much anything done with the inspector leaves DCS looking as if it's running very low texture detail, for example, any detail texture on the cat's, skid marked section on the runway, all gone. and all textures in the cockpit generally blurrier with more aliasing artefacts. Even when upping PD back up, which removes any performance increase from this. I'm back to defaults, and PD 1.3, anything suggested, and bunch of other settings suggested in this thread was a huge drop in image quality with hardly any performance gain whatsoever. Exactly the same results i see unfortunately...maybe we are doing something wrong in forcing inspector’s settings R7-5800X3D 64GB RTX-4090 LG-38GN950 N/A Realsimulator FFSB MKII Ultra, VKB Stecs Max, Winwing F-16EX Throttle, Winwing Orion (Skywalker) Pedals, Razer Tartarus V2 SpeedMaster Flight Seat, JetSeat
Bob_Bushman Posted December 1, 2019 Posted December 1, 2019 Have a hard time figuring what that would be. As mentioned I completely close DCS and reboot the rift-S between changes. And I'm running the inspector with elevated privileges. i7 8700k @ 4.7, 32GB 2900Mhz, 1080ti, CV1 Virpil MT-50\Delta, MFG Crosswind, Warthog Throttle, Virptil Mongoost-50 throttle.
Chapa Posted December 1, 2019 Posted December 1, 2019 (edited) Guys, here are some maths to explain why your PD/Super sampling settings looks amazing. First, PD or should we say Super sampling is be definition an antialiasing technique. As far as I know, it is the pendant of SSAA DCS setting but for VR. It is not totally exact to say that but it's an easy to understand vulgarisation: When you set SSAAx2 on a monitor, your GPU will render an image wit the double the monitor resolution and then downscale it to the monitor resolution with the benefit to have reduce aliasing. So, for example, if you set SSAAx4 with a 1080p monitor, it will render a 4k image and then downscale it to the native resolution of the monitor 1080p. Doubling the resolution imply to multiple by 4 the number of pixels, so the SSAAx4 and not SSAAx2. All that to say that PD is the same idea. ----- The Rift S has a native resolution of 1280x1440 per eyes, so a total resolution of 2560x1400 for a total of 3'686'400 pixels. A 4k monitor has a resolution of 3840x2160, so a total of 8'294'400 pixel Here are now the maths: [TABLE]PD Resolution per eye Number of pixels per eye Total number of pixels Multiplicator native resolution/ Equivalent SSAAx Multiplicator 4k resolution 0,8 1328x1424 1 891 072 3 782 144 1,03 0,46 0,9 1488X1600 2 380 800 4 761 600 1,29 0,57 1,0 1648x1776 2 926 848 5 853 696 1,59 0,71 1,1 1824x1952 3 560 448 7 120 896 1,93 0,86 1,2 1984x2128 4 221 952 8 443 904 2,29 1,02 1,3 2144x2304 4 939 776 9 879 552 2,68 1,19 1,4 2320x2480 5 753 600 11 507 200 3,12 1,39 1,5 2480X2656 6 586 880 13 173 760 3,57 1,59 1,6 2640x2848 7 518 720 15 037 440 4,08 1,81 [/TABLE] The resolutions per eyes have been obtained by using the HUD provided by the Oculus Debug Tool and the PD corresponds to the DCS PD setting. In reality, the true HMD PD is sometimes for example 0.91 instead of 0.9 as indicated by the DCS cursor... As you can see, a PD of 0.8 is about near the native resolution of the Rift S. Interestingly, a PD of 1.2 is the equivalent of SSAAx2 and PD of 1.6 is equivalent of SSAAx4. PD has for effect to improve clarity and reduce aliasing and shimmering. If you play at PD1.3+ and are happy with that, there is really no advantage to run the settings proposed in this post. This post is for people that want good clarity but can't afford so high PD for some reasons: want to play always at 80Hz, GPU not strong enough, HMD with high native resolution, etc... With the proposed settings: Negative LOD Bias, it is possible to have the same clarity at PD 1.0/1.1 than 1.4/1.5 but of course without the nice antialiasing and shimerring provided by the so strong but costly PD settings Edited December 1, 2019 by Chapa i7 12700k, RTX 3080ti, 64GB RAM, M.2 512Go, Quest 3
Chapa Posted December 1, 2019 Posted December 1, 2019 @VirusAM: In the texture filtering LOD Bias (DX) setting you have probably set a positive value... you must set a NEGATIVE value. i7 12700k, RTX 3080ti, 64GB RAM, M.2 512Go, Quest 3
VirusAM Posted December 1, 2019 Posted December 1, 2019 Nvidia Sharpening and VR @VirusAM: In the texture filtering LOD Bias (DX) setting you have probably set a positive value... you must set a NEGATIVE value. Yes...i used a negative value, but i had to use -0.625 as the advertised -0.6250 is not available in my settings R7-5800X3D 64GB RTX-4090 LG-38GN950 N/A Realsimulator FFSB MKII Ultra, VKB Stecs Max, Winwing F-16EX Throttle, Winwing Orion (Skywalker) Pedals, Razer Tartarus V2 SpeedMaster Flight Seat, JetSeat
Wicked.- Posted December 2, 2019 Posted December 2, 2019 Yes...i used a negative value, but i had to use -0.625 as the advertised -0.6250 is not available in my settings That small of an amount won’t make a difference. So I guess it’s hit or miss on the Oculus S. My friends using the Reverb had to make some change to a WMR file, then use SteamVR stable version and WMR beta to get good results. As of tonight I went back to DCS anisotropic filtering at 16x. Like Bob_Bushman I wasn’t seeing any results and objects like trees in the distance were blurry. I set it back to 16x in DCS and it’s cleaned up again with no frame rate hit. So back to the original topic of this post, negative LOD has most definitely sharpened the graphics in the cockpit and made three dimensional objects like knobs stand out for me. It induces shimmering which so far is best handled with setting AA transparency to AA MODE REPLAY ALL. MSAA is better at getting rid of jaggies for sure but at a high frame rate hit and a reduction in spotting targets at distance. These results are on a Valve Index which only uses SteamVR. I’m still encouraged that I can actually use something in the Nvidia driver besides prefer maximum performance and FXAA with 2x MSAA. Hopefully we VR users will have more options soon. [sIGPIC][/sIGPIC] Intel i9 9900K @ 5.1Ghz HT Disabled, Asus RoG Strix z390E Gaming, 64GB G.Skill Trident Z 3200, Asus RoG Strix RTX2080Ti OC @ 1.9Ghz, 1TB Samsung Evo 970Pro M.2 TM Warthog, CH Pro Pedals, Saitek Pro Rudder Pedals, Samsung 49" Curved Gaming Monitor, Samsung 50" 4KUHD TV, Acer 27" Touch Panel, CV1, Pimax 5K+, Valve Index, FSSB3 Lighting, F-16SGRH, 3 TM Cougar's and a Saitek X36 that I can't bring myself to part with.
Bob_Bushman Posted December 2, 2019 Posted December 2, 2019 @chapa I know full well what Supersampling is, and what it does. and those numbers are actually a little bit off. To account for loss of detail in the lens distortion they render a fair bit higher than the native panel resolution, in essence The CV 1 has a render target of 2688*1600. These numbers account for both panels. The S has not one but two, in the Oculus software for the device there is a semi buried between "Performance" or "quality", this affect the default render target. Performance: sets a default render target of 3008*1616, this is surprisingly close to the original rift resolution, but the lenses on the S has a much lower distortion than the CV1. The second Quality option has a render target of 3296*1776. How this setting is applied by default also varies, but I think the main trigger was having a GPU considered\identified by the Oculus software as equal or greater than a 1070. As for Pixeldensity vs steamvr custom resolution, PD of 1.3 is about the same as a SteamVR custom resolution of 170%, As for the inspector settings, yea the more I test these, they less they do, or make things worse, and yes I'm using negative LOD values, and even at native resolution, no SS I might hit 80fps slightly quicker, but I still spacewarp around cites <5k', and even on some airfields I was dropped as low as 20fps sitting on tarmac. With my original settings of 1.3 I would stay a little more in spacewarp, but I would NEVER go below it either, so, so far these settings has been nothing but a drop in both quality and performance. i7 8700k @ 4.7, 32GB 2900Mhz, 1080ti, CV1 Virpil MT-50\Delta, MFG Crosswind, Warthog Throttle, Virptil Mongoost-50 throttle.
Chapa Posted December 2, 2019 Posted December 2, 2019 @Bob: My post was not dedicated for you specifically but for people reading this post as simple information about what is PD and to rebound about the famous SSAA settings that as no impact with HMDs. You state my numbers are false but they aren't and the render resolution per eyes are the one displayed by the HUD provided by Oculus themselves and accessible from the Occulus debug tool. For people, not knowing these tools (accessible from Oculus Debug Tool and OTT also): https://developer.oculus.com/documentation/pcsdk/latest/concepts/dg-hud/ As you state, the Rift Sa has two settings: Performance and Quality. Behind the scene, Performance setting is in fact a PD of 0.9 or 0.91. It correspond to a resolution per eye of 1488X1600, meaning a total resolution of 3296x1600 (it is the same number than you say, your number are not totally right but same magnitude doesn't care). The quality settings corresponds to a PD of 1.0, a resolution of 1648x1776 per eye, so a total resolution of 3296x1776 (we have the same numbers again). The columns in the table are simply the number of pixels per eyes, the total number of pixels (so 2 times the pixels per eyes and some ratio). My maths looks rights and we agree on the values. The goal of the table is just to have numbers and real figures about what the GPU have to digest to really display the true native pixels of the Rift and some comparison with a 4K display for the range of usual PD values. i7 12700k, RTX 3080ti, 64GB RAM, M.2 512Go, Quest 3
Recommended Posts