VR - Cumulus Clouds

[Gildahl]

I'm on Valve Index.  I doubt its a setting since I was able to duplicate on vanilla XP.   In the other thread where this was discussed, the developer and I theorized that maybe it was due to the Vive Pro he was using.  While I don't know about the Pro (since I don't have one), I tried testing using my old original Vive and I can at least say that the problem is unquestionably less noticeable--I think, because the resolution is so bad that the eyes don't find it hard to converge the big gray blobs.  Still, when approaching the ground in fog, even on the OG Vive it is still easy to see the "mirage" effect and the banding that rotates when I rotate my head (easiest to evaluate if you pause the simulator).  On the Index, however, its a more serious story, since there is enough cloud detail for the eyes to want to converge--and when they can't it makes me go cross-eyed.  It's interesting that you were able to use VR without these issues on an earlier version of XP/SMP  In 11.50, I've tried in both Vulkan and OpenGL, so I don't think its the graphics API--and I know that I don't have any of these issues with the standard XP clouds and fog.  

Edited June 16, 2020 by Gildahl

[Jonnti46]

The problem is I don't know if it is Xplane version updates that caused this, or if it is updates in SMP - I suppose I could roll back to SMP 3 (although I've never actually tried going backwards with SMP. No even sure if technically that is possible). Not entirely sure about this. 

[Gildahl]

Hey Jonnti46, have you tried b11 yet?  I just tried it, and at first blush it seems that LR may have done something to improve this.  My testing under overcast skies has been too limited to be completely sure, but give it a try and see if it makes a difference for you.  

[Edit]  Not fixed.  I just took a flight in the Tetons using real weather (which included some dark overcast type cumulus) and I'm back to the serious head-turning weirdness.  Oh well.  :-(

Edited June 17, 2020 by Gildahl

[Jonnti46]

Same. I discovered I can't roll back versions either due to the way x-aviation system works, I even tried SMP V3 (which would load and install) but when I switched to VR there were no clouds at all - so obviously that is not a useful test. I've heard X-Enviro does not have this issue, (but who knows for sure) and they say no possible refund if it does. It feels a bit like changing deck chairs on the Titanic, you will still sink in the end!  Probably another set of issues with xEnviro (which is not as proactively developed). Hopefully eventually they can recreate this in VR and resolve it - it certainly was work fine with some version of x-plane and SMP - I just can't pinpoint exactly when it all went wrong. I think it was around version 1.41 of xplane. Being also real world pilot, clouds are somehow important to me, and SMP really got the edge - just not in VR! 

[sundog]

FYI, X-Enviro does not support VR at all as far as I can tell.

Maybe it's just my eyesight; everything looks a little blurry to me in my Vive Pro and maybe that's preventing me from seeing the issue. I'll spend some more time trying to reproduce it when I can, though. I do wonder if the issue is specific to certain headsets though; I've never heard a complaint from a Vive Pro user.

[Gildahl]

Thanks!  I guess we really need you to see this...and once you do, I think you'll agree its unmistakable.  

From this thread, I guess we can at least say it occurs with OG Vive, Pimax, and Index and that higher resolutions make it that much more obvious.  Also, keep in mind that many cloud formations do not exhibit this issue at all.  For example, I flew for about 15 minutes in the Tetons yesterday (using real weather) in and around heavy cumulus clouds including ones with dark underbellies completely convinced that beta 11 had totally fixed the problem..no convergence or "head turning" issues at all.  I was actually pretty excited about that.  But then I flew into a canyon that was exhibiting more of an overcast/foggy texture and as I approached the convergence (just in that area) got worse and worse, and along with that, the "head-turning" issue too (i.e. where turning my head back and forth causes pieces of clouds and fog bands to rotate weirdly in place).  So it definitely seems like just one specific kind (or class) of cloud texture that is doing this.

Edited June 18, 2020 by Gildahl

[sundog]

13 hours ago, Gildahl said:

Thanks!  I guess we really need you to see this...and once you do, I think you'll agree its unmistakable.  

From this thread, I guess we can at least say it occurs with OG Vive, Pimax, and Index and that higher resolutions make it that much more obvious.  Also, keep in mind that many cloud formations do not exhibit this issue at all.  For example, I flew for about 15 minutes in the Tetons yesterday (using real weather) in and around heavy cumulus clouds including ones with dark underbellies completely convinced that beta 11 had totally fixed the problem..no convergence or "head turning" issues at all.  I was actually pretty excited about that.  But then I flew into a canyon that was exhibiting more of an overcast/foggy texture and as I approached the convergence (just in that area) got worse and worse, and along with that, the "head-turning" issue too (i.e. where turning my head back and forth causes pieces of clouds and fog bands to rotate weirdly in place).  So it definitely seems like just one specific kind (or class) of cloud texture that is doing this.

Yes, I suspect it's the overcast clouds in particular, as they use very large billboards - and billboards and VR generally don't get along well. You might try experimenting with different settings in SMP for the overcast representation. Ultimately this may not go away until we have an entirely new way of depicting these clouds to offer, which is something we're starting to experiment with here.

Well, there is one exception - if you're *not* using real weather, setting the overcast in SMP to "solid procedural" or "broken procedural" would probably clear this up. But when real weather is active, the weather is usually too complicated to be represented with those options, and we automatically use a different one instead. Still, for VR, I'd set it to "solid procedural" - it couldn't hurt.

[Jonnti46]

I've tried solid procedural in the past and had the issue - but I will certainly re-test with this new version of XP and SMP. 

[sundog]

2 hours ago, Jonnti46 said:

I've tried solid procedural in the past and had the issue - but I will certainly re-test with this new version of XP and SMP. 

Solid procedural will likely only help when you are not using real-world weather, or in some very specific circumstances if you are.

[Gildahl]

Just tried some more testing.  Turned real weather off, turned procedural on, set low visibility and just sitting on the runway I could see the issue immediately.  So even without real weather, it can be quite serious.  Anyway, I've uninstalled SMP again and will try again with the next rev. as it remains too distracting for general use.

[DJSean00]

FYI both myself and my friend using latest Skymaxx in VR have the exact same problem. It was foggy in YBBN this morning and I had fog and clouds coming through my cockpit (felt like into my eyes).

I might need to drop this for now, wait like @Gildahl. However happy to do what I can to help identify the issue and support to make the changes needed.

[sundog]

Hmm... while I tested flying through overcast conditions, I don't think I explicitly tested low visibility in VR. Maybe that's the piece I've been missing. Under low visibility, the far clip plane of the projection matrix changes, and if Laminar is not passing the correct matrix to us, it would explain what you're seeing.

I'm going to be away from home for a week or so, but once I get back I'll test that. If that is indeed the case, I'll likely need to work with Laminar as it may be a bug on their end.

[Jonnti46]

I did re-test with unlimited visibility - it made no difference to the main problem. I also selected the recommended cloud sets suggested, same issue. Such a shame you can't recreate it - that makes it very tricky to sort out. 

[sundog]

I suspect there are two separate issues being discussed.

- Rotation of overcast cloud particles when moving your head around when you're near the clouds. That's something I know about and have done my best to minimize. Ultimately a new rendering technique will be needed to eliminate that entirely.

- Left/right eye divergence issues. That's the thing I can't get to happen.

But, I don't normally fly in VR so it's likely that I just haven't flown enough to experience it. I'm still stuck out of town (have to move my elderly Mom closer to us) but will spend some more time on this once I'm back in the office.

[Jonnti46]

I understand, no problem, this is not important compared to some things going on in the world currently. I just wanted to let you know that I don't think visibility is a factor. Any cu cloud approaching from below, as you get close I'm sure you will see it, it "switches" suddenly. I can be looking dead ahead, not moving my head and it will happen. It is very easy to recreate (here at least). I can do a video if that would help - but will let you see if you can see it at a later date. 

[TheHybridCockpit]

Can confirm the cloud effects near cumulus with Oculus Rift S on Beta11 Xplane. 
 

Just purchased Skymaxx Pro as UWXP failed in VR, working seamless with ASXP. Now SMP is far better than UWXP in terms of interface and stability. But also not usable in VR for now. My description of the problem is : the 2d cloud layer feels like glued or sticked onto my goggles and moves with my head. If the distance increases from the clouds, all is perfect. For an Airliner, its essential to have solid fly through effects, so I wonder what makes the standard laminar clouds better here?

Anyway, thanks for your support in advance, looks like a dedicated team behind SMP, great to see.

 

THC

 

[Gildahl]

I'm glad you are able to see the rotating cloud particles.  That is certainly one of the fundamental issues.  As far as the convergence issues, I *suspect* it has similar roots.  I tried to document this in a picture on the last version's release thread while flying over Boston.  When flying in foggy conditions, for example, I find the fog (not always, but sometimes) non-continuous, meaning that there is banding such that the fog density increases with distance in steps.  Like the "cloud particles" these "bands" rotate with the rotation of my head and more critically, the left and right eyes see these bands in different positions.  The way I *think* this leads to what I call a "convergence" issue is that (as shown in my picture), you might have a building that appears in the left eye as shrouded in a dense layer of fog, but because of the rotating density banding, the right eye might simultaneously be seeing the building much more clearly (for example, if I paused the simulator and then blinked my eyes back and forth, left to right, I could see the building as alternately in the fog and out of the fog).  Thus, even if the parallax of the building is correct, the competing fog density differences in right vs left eyes make the brain cry foul and makes it nearly impossible to focus.

[sundog]

12 hours ago, Gildahl said:

I'm glad you are able to see the rotating cloud particles.  That is certainly one of the fundamental issues.  As far as the convergence issues, I *suspect* it has similar roots.  I tried to document this in a picture on the last version's release thread while flying over Boston.  When flying in foggy conditions, for example, I find the fog (not always, but sometimes) non-continuous, meaning that there is banding such that the fog density increases with distance in steps.  Like the "cloud particles" these "bands" rotate with the rotation of my head and more critically, the left and right eyes see these bands in different positions.  The way I *think* this leads to what I call a "convergence" issue is that (as shown in my picture), you might have a building that appears in the left eye as shrouded in a dense layer of fog, but because of the rotating density banding, the right eye might simultaneously be seeing the building much more clearly (for example, if I paused the simulator and then blinked my eyes back and forth, left to right, I could see the building as alternately in the fog and out of the fog).  Thus, even if the parallax of the building is correct, the competing fog density differences in right vs left eyes make the brain cry foul and makes it nearly impossible to focus.

That "banding" sounds like the cloud/terrain blending slider in SMP isn't up enough. I know the docs recommend setting it to 0 to maximize performance in VR, but that's exactly the trade-off you make by doing that. Might want to fiddle with that if you haven't already..

[Gildahl]

You had mentioned this before, and yes, I had tried setting it to max, but didn't fix the problem (as I recall, density between bands was a bit more graduated, but still clearly visible).

[TheHybridCockpit]

It’s not a deal breaker, but it also comes across latest reviews, so we should be fair in requesting a fix.

seen this one? 
 

SMP vs UWXP

 

Regards, THC

[sundog]

No, I don't look at reviews. I'm aware that there are situations where things don't look perfect in VR. I've been working to improve it all along and continue to do so.

[Jonnti46]

I don't suppose you've managed to get any further with this issue? Using real weather and now the winter is fast approaching, all those lovely clouds are much more in the cockpit than in the summer! I see other things were fixed but this strange situation of the clouds being re-projected incorrectly in VR is ever present. I see from other posts work was done directly with Ben at Xplane to correct issues when there was cloud issues inside the cockpit in normal mode (non VR), is it possible to investigate with them re this issue as well? It is hanging around for so long, I know it is just us VR guys suffering which is probably small, but when MS sorts out their VR I fear I might be jumping ship! It is such a shame as it basically means the whole immersion is lost as you get close to the clouds. 

[sundog]

On 10/6/2020 at 7:02 AM, Jonnti46 said:

I don't suppose you've managed to get any further with this issue? Using real weather and now the winter is fast approaching, all those lovely clouds are much more in the cockpit than in the summer! I see other things were fixed but this strange situation of the clouds being re-projected incorrectly in VR is ever present. I see from other posts work was done directly with Ben at Xplane to correct issues when there was cloud issues inside the cockpit in normal mode (non VR), is it possible to investigate with them re this issue as well? It is hanging around for so long, I know it is just us VR guys suffering which is probably small, but when MS sorts out their VR I fear I might be jumping ship! It is such a shame as it basically means the whole immersion is lost as you get close to the clouds. 

We do have some VR fixes slated for our next release, so stay tuned. I'm hopeful it will help, but for some reason I've never had clouds in the cockpit in VR with my own setup, so I'm not 100% sure this particular issue has been fixed.
 

[rawdmon]

I get the same issue which I had also reported to Frank in a different thread and he stated that he was working on several VR related fixes (as he mentioned in his last note).  We'll have to wait and see if this ends up being fixed in the next release whenever it's available.

@Frank If you want to me to beta test it to check and see if it's resolved prior to release let me know.  I can reliably reproduce the issue in the current version.  I get this with the fast clouds as well btw.

Edited October 8, 2020 by rawdmon

[Jonnti46]

Its not clouds in cockpit - it is clouds that go strange in the near distance, which makes your eyes feel like you are going "cross eyed" its to do with the rendering and I presume reprojection in VR. It is really strange that you never see it, I can't think of any reason why it would be not be showing up in some headsets other than others - I have a feeling perhaps you are not looking for the same thing? Or looking for something different? Clouds in cockpit is something slightly different and might be misleading, they feel like they are in cockpit (perhaps), but they are not, the clouds switch / elongate in their form, and due to that lose the 3d effect which I think causes your eyes to do strange things / see double. It is most noticeable approaching them CU or stratus layers.