I’m perfectly aware of the support status of the OXRT. However, without it at any Foveated Rendering Scale other than 0.50 with Pimax Crystal OG, the image in VR is affected by jitter, when DFR is on in MSFS2024.
If you know any solution to remove the jitter, please share it. I’m only using OXRT as a workaround because it allows setting the custom resolution to the value which can be divided without a reminder by a Foveated Rendering Scale factor (0.25 - division by 4), which eliminates the jitter.
As I described in my test results, with OXRT disabled, the jitter is still there, so it’s not OXRT which is causing it, but some interaction of Pimax and Asobo resolution manipulating with DFR on in the sim. I hope either Asobo or Pimax will eliminate the bug (present also on SU3 but with slightly different way of rounding when dividing the resolution by the Foveated Rendering Scale factor), which will make OXRT truly redundant (other than it’s ability to display FPS).
By the way I’m not using the DFR in OXRT, nor QuadViews in PimaxPlay, only the native DFR/QV in MSFS2024.
My point remains simply that your use of OXRTK is just an extraneous variable that does nothing to help debug the regression.
By all means if it helps you as a workaround great, as a beta tester I’m not particularly interested in finding workarounds as opposed to identifying clear regressions so they can be fixed.
The key difference may be that you apparently are running Pimax Crystal at 1.2 setting (super sampling) in PimaxPlay while I’m running it at 1.0. I will try 1.2 later today.
I’m using OXRT only as a worakround tool to apply the resolution which is divisible by 4 (0.25 Foveated Rendering Scale) without reminder. TAA slider doesn’t allow setting vertical and horizontal resolution independently. And for normal flying In using DLAA not TAA.
We have created an internal ticket to see if our team already has this logged, and if not they will attempt to reproduce the issue and create a new bug report. This item is now marked as feedback-logged. If there is an existing bug report or one is created, we will move this thread to bug-logged.
Either you are lucky and not affected by this jitter/image vibration. Or it might be that you don’t perceive it - VR perception may differ between people. It’s not about stuttering, FPS drops or hesitations when you fly or move you head. It’s image vibration even when you sit still looking straight ahead. It’s most visible in the distance, for example on a line of trees, but if you look closely you will find it in the entire image, even in the cocpkit or in the menus as well. For me it’s obvious and very annoying image vibration.
Or maybe I and some other guys who reported this issue are just unlucky and we have someting in our systems which is causing this jitter.
I tested with various TAA values and various “Image Quality” settings in PimaxPlay and the jitter is still there, even with OXRT disabled, unless I apply the OXRT resoultion change workaround described several posts above.
I can’t speak directly to what is causing this, but I can definitely say it’s awful when it occurs.
It came back to bite me in the behind today. I had set my Pimax Super to medium resolution in Pimax Play for some reason and hadn’t remembered to set it back to high before starting up SU4 beta with render scale at 0.25…which usually works.
Bam, everywhere I looked in the scenery I had image shake. I’m thinking, “What the &%$$$!!?”
I checked the Toolkit and it showed some weird lower resolution.
So I switched from DLAA to TLAA and the resolution of the foveated area was not a neat divisor. Restarted the sim with the Super at “native” resolution of 6240 x 6280 and everything was fine once more.
This bug is definitely not hard to spot with a Super and it’s “super” annoying.
Okay, I think I may have noticed this but will need some more experimenting. It’s very subtle. It’s more apparent on the ground. Once in the air everything is moving around so much I noticeess.
Is it that only 0.5 and 0.25 work or is it dependent on the render resolution?
Depends on the render resolution. Seems like the jitter appears in SU3 and SU4 beta when any of the below conditions is met:
(horizontal resolution) x (Foveated Rendering Scale) = non integer number
OR
(vertical resolution) x (Foveated Rendering Scale) = non integer number
In SU4 there are also some rounding errors in the multiplications above which were not present (or were different) in SU3, these errors make obtaining integer result more difficult. Crystal Super and Crystal OG have different native resultions, so the issue manifests itself differently. Also depends on the resolution (Games → Image Quality) set in PimaxPlay.
I did log a ticket with Pimax (before I found this post) hopefully they can put some pressure from there side. Since a basicly can’t use a value other then 0.5 I’m loosing frames for no good reason and knowing how much “value” a frame more is (especially when running a 5090) it’s freeking frustrating.
You may try finding the resolution combination which results in integer numbers. You can change the resolutions in the modified OpenXR Toolkit (allows setting horizontal and vertical resolution independently). Then you can play with various Foveated Rendering parameter (0.5, 0.25, 0.2) looking for integer resolution of the foveated area (best to see it in TAA mode as it shows the resolution in the MSFS VR Graphics mode immediately as you change the FR setting).
I created a ticket for this issue with pimax and referred to this forum post to make sure they understood it.
They replied that it will be fixed in the next version.
They also gave me a workaround on the spot (I cannot share because it involves specific files replacemnet)
I would suggest, wait for the next version or create ticket with Pimax and refer to this post.
Happy VR-ing!
(I also created a ticket for the “not connecting on boot” of the super (making me disconnect USB to make it work) , this is also resolved in next version.)