Split shadows off didn't work that well for me visually. Shadows are better with them on and I don't see any artifacts in WMR. Little trick on boosting sharpness with LOD BIAS. Use NVidia Inspector and two highlighted values. The more negative LOD Bias, the more aggressive the sharpening (and artifacts).
Did anyone notice nvidia MFAA to make any visual difference in R3E + VR? Also, this seems promising (20xx only), but I had no chance of testing it yet: https://www.reddit.com/r/oculus/comments/eljrr6/vrs_variable_rate_shading_how_to_enable_it_on/
I've tried, but didn't get it to work. Others wanted to try, but never heard again. I think, it hasn't worked there.
Technically MFAA should work on any title supporting MSAA. It works with AC in VR. The easiest way to test is to set lower MSAA level (x2) so jaggies are more obvious and run with and without MFAA.
Anyone have settings for a Valve Index that I can use? I have a 9900KS and 2080Ti with DDR4-4000 32gb.
The settings in the OP are pretty much the optimal settings to go for, any extra headroom I would put into supersampling
I was about to write a frustrated post about how VR performance has been significantly worse since the december patch, but then I noticed that I bumped SS up a little bit at the time and that is the root cause of the decline. What's interesting is that my GPU has enough headroom for higher resolution, but it seems it has a more significant impact on the CPU. Is this observation right?
The "details" are in this thread, but the short answer is: Yes, CPU is the bottleneck, r3e uses a single thread to render stuff.
Yes VR users like us are limited by CPU/RAM performance, as you already saw in Dx9 CPU Benchmark thread, despite using same processor, your slower RAM (latency especially) kit puts you roughly 10% behind my score, so you are limited by your RAM kit imo. Note that even with a faster kit we are limited too but with more performance nonetheless. https://forum.sector3studios.com/index.php?threads/directx-9-cpu-benchmark-thread.13473/ I've now switched to WMR HP Reverb so I no longer need the Open Composite to bypass Steam VR and use only Oculus. I've raced in the Reverb at native resolution and it worked absolutely fine at 90 fps with these settings Toggle Mirrors............On Track Level of detail.....Medium Car LOD Distance setting..Low Particle detail...........Low Tiremarks.................On Specular..................Off Car reflection quality....Low Track animations..........On Shadows...................Lowest Car shadows...............On Contact shadows...........32 Shadow split..............On Multi-Sampling............4xAA FXAA......................Off Bloom.....................Off Depth of field............Off Motion blur quality.......Low Lightshafts...............Off Lens effects..............Off Track texture quality.....High Car texture quality.......High Rear view mirror quality..Low Shader quality............Medium Corner markers............Off Visible cars..............20 Opponents cockpits........Off So you should be able to use the same settings with a bit of super sampling to reach similar resolution at 90fps (2160x2160 per eye in Steam VR) with a faster RAM kit.
How many cars does that cater for out of interest? I run almost identical settings, and I find that somewhere between 17-20 cars* is when I'll start dropping to 45fps. *Depending on class, and amount of model variation.
Thanks, I'm using similar settings (slightly lower actually) and my system can maintain 90 FPS with 130% SS in general, but 150% is the sweet spot in terms of visual clarity and it is just above the CPU limit, which is a shame, because my GPU is producing something like 2-4 ms frametimes according to fpsVR (so there's obviously headroom there). I set 'Visible cars' to 18, this way the game runs pretty stable apart from very crowded spaces (starts, pitlane, or massive pile-ups), GT3 is the best class to test because of the various 3D models.
If it's just above the CPU limit (less than 10%) faster RAM could probably get you there as this is what seperate our systems in terms of performance. BTW I haven't found 3600c14 to be really more performant than 3200c14 with infinity fabric (FCLK) set to 1766. So don't go crazy if you replace your RAM kit.
Woah, 3600c14 RAM's are not even available in my country (to be more precise, the one available kit costs a fortune, we have crazy VAT here). And I'm not convinced the difference between 3200c16 and 3600c16-17 will be significant enough to buy new RAM just because of this. But thanks!
Actually I'm using a 3200c14 kit OC to 3666c14 using 1usmus Ryzen DRAM Timing calculator, and DX9 benchmark thread proved the gains to be real https://www.techpowerup.com/download/ryzen-dram-calculator/
Hello Guys, I have just upgraded to a RTX 2080 S from a GTX 1070 , I changed Raceroom video settings to High but unfortunately the FPS are stuck at 40 FPS whilst the headroom performance is 100%. What can I do to get 80 fps starting from an "high" profile?
Hi, unfortunately it does not work like that. Take the settings posted above by Sebastien, and go from there. If you want to get more of an understanding of what to do and what not read this thread.
even if I change to video settings low, I am not able to run at 80 fps... are we saying that my hardware is not enough to run it at that framerate?
Your cpu core speed matters a lot. I got a 9600k@48ghz + gtx1060 and I can run a constant 90fps. Can't run the exact settings above though. I presume you've got a rift s, and that should not be a problem. Think you are running in asw mode atm.