I've just received my Pimax 8K X, and have spent some time tinkering with it.
I'm going to keep on editing this post as I test various settings.
For reference, I have a 3090 and 5900X with 32GB RAM... etc.
I almost exclusively fly the Tomcat, so my performance is focussed on that. In addition, due to its steam gauges, I need better visibility.
Test 1:
Just did full basic settings. I basically followed the preset I had for my Reverb G2, except with the 8K X I'm running native resolution. It was OK, with FPS being around 35 - 45. I had FOV wide, no smart smoothing, parallel projection disabled, render quality 1, fixed foveated rendering close. 75 Hz. It was OK. GPU went up to 99% usage, frametimes around 25 - 30ms.
Test 2: I tried smart smoothing. GPU usage went down, experience didn't feel good. Have now disabled it.
Test 3: Enabled MSAA X2. Visuals significantly better, and can read small numbers on HSD clearly. An impact of a few frames, but consider it necessary.
Test 4: Trying normal FOV. Doesn't feel like much of a penalty for using it at all. Had an average of 45 FPS or so, which felt very usable.
Test 5: Conservative FFR. Didn't notice too much of a performance degradation. FPS improved by 3 - 4 FPS.
Test 6: Balanced FFR. Generally OK. Little worried about shimmering at the bottom, but it was at the edge of the lenses where this are a lot more peripheral. FPS another 3 - 4, generally sitting around 50 in my test scenario, maybe a little above.
Test 7: Aggressive FFR. Made things too blurry and started to interfere with the ability to read small text. Though it provided 2 - 3 FPS more, the blur wasn't worth it. Shimmering too. Staying at balanced.
Test 8: Ketegys VR Shader. This added a solid 7 - 8 FPS to stay at 58 - 60 FPS, and generally speaking, was very good. That said, I did lose a little visual quality. I suspect this is conflicting slightly with FFR. Due to this, I will try to going back to conservative FFR for the next test.
Test 9: Updated firmware on headset to allow 90Hz support. No noticeable performance differences in game. (Just wanted to make sure the firmware didn't massively shift something).
Test 10: Went back to conservative FFR with Ketegys shader. Visual acuity improved a little, and curiously performance improved by a frame or two. For the sake of testing, I am going to switch off FFR again and see if that has any further impact.
Test 11: Switched off FFR. Saw about a 5 FPS drop. Visual acuity notably better in other areas, but not 'necessary.' Going to switch back to conservative FFR as a compromise.
Test 12: Worked on IPD offset and my personal IPD. My IPD is 68mm, so I set that as the baseline. I was feeling that when I looked at the HSD of the F-14 I was getting eyestrain. I fiddled with the IPD offset settings, and -5, 5 feels a lot better, as well as easier to read. I felt less eye strain immediately. I tried the whole range, but I could see the inside of the screen if I went too far. Happy with offset, I revisited my IPD to see if clarity changed by moving the lens. It was at its best at 68mm, which in theory suggests I did this correctly...
Test 13: Set to 90Hz. No noticeable performance difference or impact.
Test 14: 90Hz with Smart smoothing... I'm conflicted on this. The idea behind this is that there should be no problem that my card can render 45fps, and then it would fill in the gaps. For some reason it just dropped all the way down to 30FPS, which is just undesirable. Unless smart smoothing gets better and hooks to 45 FPS, I won't use it. Shame really.
Test 15: Render quality 1.5. Obviously better image quality, but dumped FPS(around 30). Shame since I could read gauges better.
Test 16: Render quality 1.25. Still too much of a performance impact it seems.
Test 17: Going to try FHolger AMD FSR mod. Performance went up to about 70 - 80 FPS, which is great, but visual quality really dropped. In particular, with FFR you get a shimmering effect which just makes it unworkable. Going to disable FFR and see if this improves anything. Legibility of numbers degraded.
Test 18: FSR, FFR disabled. Doesn't solve the problem with reading numbers, but shimmering on the outside has stopped. Impact 2 - 3 FPS. I realise that FSR does its own form of FFR, so I think in this scenario it's best to have either one disabled. In my view, it seems you can leverage more out of FSR.
Test 19: FSR, radius set to 1. Didn't make a visual difference, with a small frame penalty. I'll set radius back to 0.7 or so to be a stand-in for FFR.
Test 20: Set Pimax render quality to 1.5, and use FSR. This could just be just a roundabout approach of just doing the default resolution, but it could be an approach towards getting visual acuity and having a little performance. It was about 40 FPS. Gauges were easily legible, but performance was not adequate.
Test 21: Set Pimax render quality to 1.25. This gave about 55 FPS average. Text was adequately legible. It was nice not to have shimmering in the corners, that would usually be cause by FFR. I think this is a rough equivalent to FSR disabled, FFR conservative, render quality 1.
Test 22: FSR disabled, FFR conservative, render quality 1. FPS sat at around 60, so generally superior performance. Legibility may have been marginally worse, but workable. I'm conflicted, principally because with FSR the overall image quality is better and you don't get shimmering. The Conservative FFR setting is ugly and a bit too brutal. If there was a midpoint between the FSR quality and FFR, that would probably be best.
Test 23: Went back to try FSR, with FFR disabled. For some reason performance has jumped with FPS at around 70, with same image fidelity as before. Clearly I'll stick with FSR over Pimax efforts.
On another note, does anyone find whether changing any thing in Steam's VR settings actually alters anything?
For reference: I also dialled down the red colours to -4, and contrast to -1. Everything feels too red!