Yay, my 8kX came one day earlier than expected, shipping went smoothly, all good! (that customs number seems to have worked for me).
Here some initial impressions after some hours of experimentation. Will update this as new impressions become available.
My VR background: Have an original Vive and a 5k+, have been using the latter for the last year and am quite happy with it.
Disclaimer; I might not have found the optimal supersampling settings yet - see below.
great (better than 5k+):
- the added details in native mode are very obvious - looking really good!
- almost no visible sde
- better weight distibution
- no “snow effects” experienced yet, even without special cable care (I sometimes see randowm white pixels with my 5k+, have to make sure that the cable isn’t close to any electromagnetic source - think my cable isn’t shielded too well. If this is ensured then the snow goes away on my 5k+ and all is ok. Not needed on the 8kX, works out of the box.)
- ok, also a little room for improvement but definitely better than with the 5k+: still seeing noticable aliasing on straight lines, even in native mode (the resolution isn’t “retina” yet, so this probably has to be expected). But e.g. texts on Virtual Desktop are readable a lot better now. Not yet monitor level but quite usable. Could be related to the supersampling-oddities though, could imagine that aliasing could be improved a lot with well adjusted SS and a card that can handle it, see below.
ok (like 5k+)
- distortion is similar to the 5k+: noticable but imho not too disturbing - normal FoV is still the best setting for me. Perhaps dynamic correction via eye tracking can get rid of that at some point? Don’t know, don’t care too much either.
- colors, contrast and brightness appear similar, neither better nor worse - thus ok.
- 75 Hz mode doesn’t disturb me e.g. in Alyx in a first test (I notice a difference in Eleven Table Tennis though - see below)
room for improvement:
- noticably heavier than the 5k+ (but fortunately the better weight distribution somewhat helps there)
- sweetspot appears slightly smaller than 5k+. But I can find a spot where the center is sharp for both eyes. Could be related to comfort kit. Or perhaps also to the added sharpness - you see much more of a difference between “almost sharp” and “really sharp” on the 8kX - as the 5k+ doesn’t have that level of sharpness, no matter how you adjust it, this might not be as obvious
- 75 Hz is noticably less smooth for Eleven Table Tennis, prefering the 5k+ in 144 or 120 Hz for that game
- 119 Hz upsampling mode has quite disturbing motionblur effects, would prefer the 75 Hz native mode for sports games when just having the 8kX. (Frametimes are far below critical levels)
- in 119 Hz upsampling mode I see a copy of the screen content from the middle on the very left side. This also happens in Pimax Home (the thing with the Pimax logo) and does thus not appear to be SteamVR related. This happens no matter which FoV is selected.
Does not appear in 75 Hz native mode.
- noticable mura (particularly visible on white surfaces). Are there perhaps correction files similar to StarVR? Would probably have to be calibrated per device…
- using the comfort kit for sports games leads to sweating much faster, lenses also start to steam up. A short break fixes this.
- the on/off switch led doesn’t go blank when switching the headset off, it turns and stays red.
I’m still somehow puzzled when it comes to supersampling settings. From the numbers everything looks correct: The PiTool factor is used to adjust the resolution of the 100% level of SteamVR global SS and per game SteamVR SS is relative to global SS. The resulting resolution numbers shown in SteamVR make sense.
But somehow I don’t see any difference in visuals nor frametimes. The only obvious difference appears when switching to Upscaled mode - then visuals get worse and frametimes go down (as expected and good, as we need to reach 119 Hz now).
Have set maxRecommendedResolution": 16384 and experimented with renderTargetMultiplier in the steamvr settings. Was also restarting SteamVR and even the HMD and services, still no difference.
The only supersampling settings that had an immediate impact visually and performance wise were first the one in Virtual Desktop. Setting this too high (particularly with MSAA x8) is reliably killing my framerates. But it starts to look really nice.
And secondly the settings in Alyx, when using “-console -vconsole +vr_fidelity_level_auto 0 +vr_fidelity_level X” with X between 1 and 6.
But these two likely circumvent the SteamVR and PiTool settings and do some custom supersampling.
Or could it be that SteamVR does some restricting based on used hardware? (“I know you don’t want that factor, it would just judder, I’ll do you a favor and just ignore these settings for you”). I have a 1080 TI. But as I don’t see any difference between SteamVR SS x0.5 and x5 that would mean that my actual setting would have to be capped close to x0.5 (for PiTool setting x1). And why proclaiming high resolutions then when they are in fact not used?
Have to experiment further, any recommendations would be appreciated.
First conclusions: Text on virtual monitors (like in Virtual Desktop or BigScreen) and games like Alyx look noticably better on the 8kX than on the 5k+. Currently I’m glad that I kept the 5k+ though (mainly for sports games where refresh rate makes a difference for me and it’s less sweaty). But for most things the 8kX is definitely an upgrade. And if the 119 Hz mode works as well as the 120/144 Hz mode of the 5k+ at some point (which should hopefully be fixable via software updates) then there wouldn’t be much reason left to futher use the 5k+. All in all I’m happy with the device and glad that I pledged for the 8kX! Some tuning would still be helpful, but the 5k+ also went a long way when comparing the current experience to the initial one. So I’m optimistic and will keep the 8kX.