PSA: Reverb G2 small sweet spots, observations and solutions

I agree with this and why I’ve commented not only the Asobo developer behind the implementation of the mouse cursor must be left eye dominant, but so is the entire test team… :grin:

I my personal case yes I have one of the two eyes needing slightly more correction than the other, but not much, and like you said, it shouldn’t show as much difference when comparing 2 headsets like G2 and Index.

Yesterday, following up your post in the Cliff House, I’ve tried capturing some photos which would best represent the differences I see with and without glasses. I’m not entirely fond of the result so I’ve kept publishing them for now but I’ll do. In trying to adjust the G2 for this photo shoot and taking notes of the differences in order to make sure I can convey the same in the photos, it was clear to me I can enhance the sharpness in “forcing” on the eyes (trying one at a time then both), like when I try to read up close I have to force as well. When you’re forcing on the eye, you do feel it and you do know you’re doing it, as much as you do know when you let your eyes relaxing to their natural focusing “strength”. And this is where right after I’ve tried again the Index for comparison and it is clear to me: I just let go my eyes relaxed and I can see naturally in the Index, whereas if I don’t force I can’t see as clearly in the G2 (let alone chromatic problems due, I believe, in a difference in calibration between the G2 optics and the WMR anti-CA).

Now you have the same correction than me, the same headset, yet, you can see clearly in the G2 and I can’t… Did you manage to fly say 1H00 or 1h30 straight since the last time? Did you feel any particular fatigue in the eyes?

1 Like

Typo? Mean “fond”?!

The only fatigue I’ve noticed is when I try to fly a more complex plane or crank up the graphics settings and get micro-jittering or juddering and it doesn’t take very long for a headache from that to set in! Of all the planes I’ve tried, the Shock Ultra, maybe because it can be flown at the ~slowest speed?, gives my best eye experience with the highest graphics settings but I’ve barely scratched the surface of what I might try. For me, I think the novelty of VR is still overwhelming with enthusiasm any objective judgements I might try to form on long-term experience. I’ve been to Hong Kong a number of times IRL but it’s been almost a decade since I was last there. Just to fly around the bay and see what’s changed in that decade is incredible (I hope that city gets a photogrammetric update - flying over Victoria Peak looks pretty yukky in terms of details right now!). Since Hong Kong is one of the most scenic and physically blessed locales in the world, it might pay MS dividends in terms of enthusiasm for MSFS to upgrade the visual quality of Hong Kong flying sooner rather than later!

1 Like

typo, thanks!

1 Like

There you have it! By the index finger test, I’m currently left eye dominant and have been most of my life, as far as I can tell (I asked my right eye be operated on first for cataract surgery for that reason!). So if the sharpness of the VR image and the VR mouse cursor position is somehow related to eye dominance, I fall into the leftist group that designed VR for MSFS (urban legend starting here, how long before picked up QAnon!? While politics is verboten on the site, I hope mild political jokes are not!). :slightly_smiling_face:

1 Like

I’m not sure for mild politics, but religion and philosophy started 18H ago :joy:

How Realistic is VR - #180 by KevyKevTPA

PS: there are interesting developments in the posts following I encourage reading too!

1 Like

I am a VR McScrooge! Bah! Humbug! :wink: How Realistic is VR - #197 by JALxml

1 Like

Here are mines for info:

SteamVR Runtime Version: 1.16.2
vrcmd Version: 1.16.2

Device 0 - holographic.WindowsHolographic - HP Reverb Virtual Reality Headset G20 by WindowsMR - HMD - generic_hmd
		D3DAdapterIndex:   0
		 RecommendedRenderTargetSize: width=3172, height=3100
		 Left Projection: left=-1.169062, right=0.991684, top=-1.057655, bottom=1.054540
		Right Projection: left=-0.993433, right=1.169252, top=-1.054049, bottom=1.057184
		Lens Center: (0.000000, 0.000000), (0.000000, 0.000000)
		Eye To Head Left
0.999994 -0.001135 0.001575 -0.032917
0.001133 0.999995 0.001227 0.000006
-0.001576 -0.001226 0.999994 -0.000132
		Eye To Head Right
0.999994 0.001133 -0.001576 0.032917
-0.001135 0.999995 -0.001227 -0.000007
0.001575 0.001228 0.999994 0.000133
		Tracking universe: 42185081794400
Driver holographic : 0 displays
Driver htc : 0 displays
Driver indexcontroller : 0 displays
Driver indexhmd : 0 displays
Driver lighthouse : 0 displays
Driver oculus : 0 displays
Driver oculus_legacy : 0 displays
Driver gamepad : 0 displays
Driver null : 0 displays
Dashboard pointer device: 0

And here is the G1 I had in June (glad I’ve saved these just in case):

SteamVR Runtime Version: 1.13.2
vrcmd Version: 1.13.2

Device 0 - holographic.WindowsHolographic - HP Reverb VR Headset VR1000-2xxx0 by WindowsMR - HMD - generic_hmd
		D3DAdapterIndex:   0
		 RecommendedRenderTargetSize: width=2208, height=2160
		 Left Projection: left=-1.112804, right=0.944776, top=-1.001175, bottom=1.013526
		Right Projection: left=-0.948403, right=1.110657, top=-1.006503, bottom=1.011904
		Lens Center: (0.000000, 0.000000), (0.000000, 0.000000)
		Eye To Head Left
1.000000 0.000000 0.000000 -0.031500
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
		Eye To Head Right
1.000000 0.000000 0.000000 0.031500
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
		Tracking universe: 27718518337900
Driver holographic : 0 displays
Driver htc : 0 displays
Driver indexcontroller : 0 displays
Driver indexhmd : 0 displays
Driver lighthouse : 0 displays
Driver oculus : 0 displays
Driver oculus_legacy : 0 displays
Driver gamepad : 0 displays
Driver null : 0 displays
Dashboard pointer device: 4294967295

Now I’ve to figure out a quick way to represent these matrices graphically, like here:

HMD Geometry Database - HMD Geometry Database

Sure, sure, it’s always my fault! Just for the record, my wife would agree with that 1000%!

:rofl: :rofl: :rofl:

1 Like

Another thought that enters my head as to how a disparity between eyes could cause problems with visual sharpness in sweet spot or elsewhere is the use of the interpupillary distance, the distance between both pupils. What would be better is if HMD headsets could be adjusted for monocular PD’s, a.k.a., Dual PD’s (the distance from a pupil of the left or right eye to nose bridge). Monocular PD’s are not necessarily the same for both eyes and I should hope if you go to a really good optician, they use monocular PD’s, rather than the IPD to make and position your lens centers for expensive prescriptive glasses, etc.

How to measure pupillary distance (PD) - All About Vision (describes monocular PD determination as being better than IPD).

One way of getting an idea of whether there is a significant difference in your monocular PD for each eye (other than calling your optical care provider or trying to measure it directly yourself with a ruler or app) would be to close one eye when doing a vision sharpness test with your headset. For example, with the G2, adjust the physical IPD slider looking only through one eye. Do it a number of times to get an average (averages are more accurate than individual determinations for a Gaussian variable by ~1/sqrt(n), where n is the number of determinations, so an average of 4 determinations is 2x as accurate as the value of a single determination). Do the same for the other eye. If you get the same IPD with both eyes, then you’re monocular PD is the same for both eyes.

So it would be interesting if the folks that think their headset is very sharp have identical monocular PD’s whereas folks that find the vision not so sharp have significantly different monocular PD’s for each eye. Wonder if any monocular PD adjustments, physical or software, will ever be offered?

This is a good point as well! The headsets are also missing vertical adjustment but not everyone might have both eyes at the same height either.At least a software only vertical offset per eye could help there.

In my case I’ve tried everything (I guess): rotation around the nose (diff height), sliding the headset left right (PD instead of IPD), and up down + IPD (alignement to lenses centres), vertical rotation as well (toward forehead and away from the chick bones, or the reverse). Nothing does.

When perfectly aligned, which I can discern by comparative adjustments like you are indicating, there is no CA but every thing is a little blurry. Forcing on the eyes can restore some sharpness indicating to me a closer focal distance than expected or experienced with the Index. When rotating (forehead/chick bones) I can manage to get a sharper picture with little eye effort, but there is visible CA and I can see the anti-CA filter spreading RGB too far apart. I believe it is sharper though because of the rotation putting my eyes slightly above or below the center line and therefore view the image at a different focal length, because of the lens shape at this location light rays are converging closer (or farther - at this hour I can’t think much LOL). But because the focal distance is not the one it has been calibrated for, I see the anti-CA working too much. NB: a WMR dev has confirmed me the G2 headset anti-CA is calibrated per-headset prior shipping and can’t be user-adjusted.

1 Like

NB: a WMR dev has confirmed me the G2 headset anti-CA is calibrated per-headset prior shipping and can’t be user-adjusted.

That’s an odd thing to say as there definitely must be an “adjustable” software component as the new introduced in last november WMR anti-CA filter is applied to older headsets like Odyssey+ and Reverb G1 aswell.

Very late edit: Also the hardware IPD adjustment on G2 can be seen as a user-adjusting the anti-CA filter or, as I said earlier, the filter will be applied incorrectly to the headset.

Yeah, I know that the G2 panels need time to warm up, but I have now played about 45 minutes worth of oing pong and it wasn’t getting better.

It’s either the panels themselves being not low-persistence enough compared to what the Oculus uses or there’s another SW bug. Did a bit of research and there are people complaining in the Star Wars Squadrons community that they experience the same thing and are getting more smooth output at 60Hz. Will give that a go and report back.

PS: the reddit answer was posted publicly in the release thread if the latest WMR for SteamVR release announcement, but I have no clue on how to link a reddit thread using the reddit android app, hence the screenshot.

1 Like

Thank you for sharing. It seems we’re in contact with the same people :wink:

There is a known “issue” (not really one but let’s call it like this) with the G2: it takes a few minutes for the panels to warm up. During this time, you can have some ghosting images when moving (like a moving white square over black background leaving some trailing faint images of the square). I don’t know how long but I wouldn’t be surprised it’d take at least 2 or 3 mins.

When in 60Hz you should notice any bright scene flickering a lot. When in 90Hz it won’t. Actually I personally find 60Hz unbearable in FS2020 daytime flying, but for night time it is not bad, except with EFIS screen flickering and less fps giving less reprojection images to compensate fast movements.

Yeah, definitely not 60Hz, that flicker is pretty prominent. Seems to be a SteamVR issue… Running the same sims and games through ReVive is buttery smooth… need to try running the same stuff through SteamVR/OpenVR on the Oculus headsets tomorrow and see if that problem persists… G2 is definitely on 90Hz but feels like 60Hz.

All those WMR/WMR for SteamVR/OpenVR layers are getting on my nerves. It’s time OpenXR sees some adoption outside of MSFS2020… :smiley:

1 Like

@CptLucky8, how, where, and why would you change those settings. (From 90Hz to 60Hz as an example.)

it’s here. But it’s gonna make your eyes bleed…

@BernhardBerger, I think I know how to properly use WMR (via the Mixed Reality Portal) with FS2020, but try as I might, I haven’t been able to figure out how to use the SteamVR option. When I try to start MSFS from the SteamVR “waiting room”, nothing happens except the app starts running, but seems inaccessible from within Steam.

I’m sure these are just the growing pains of a noob (that would be ME), but can you help a stupid noob out? I’m not even sure I completely understand what you guys are even saying with all this stuff. Like “ReVive”. No clue even what that is, much less how it’s applicable to VR with MSFS.

■■■■ noobs!!!

Sorry for a little off topic. Does the WMR software have to run/startup when SteamVR is the active OXRuntime?

**`

To follow up: it seems to be a SteamVR problem on my side, unrelated to the G2 or WMR on a few games. iRacing and other sims run just fine.

1 Like

@JALxml (and others)

NB: please makes sure to view the pictures at their full res (100% zoom).

I’ve tried to take through the lens photos showing up what I see. It is very hard to get exactly the same because the camera lens is not the same as my eyes, therefore, I’ve sometime post-process a little the photos to compensate and otherwise I’ll pinpoint the specific pixels to look at in the photos.

The problem so far

What I see in the G2:

When I place the headset in such a way there is no chromatic aberration on white lines and text, it appears blurry like in this picture.

NB: just look at the upper half of the following picture, starting at the “The Blu” icon at the bottom, ignore everything below it because there is too much CA.

What I see with +1 correction glasses

When I wear glasses I can see much clearer in the center (slightly more than the picture below) and the disk of clarity in the center is wider.

NB: the photo is showing some CA but there is nearly none with the glasses, except to the edges due to the glasses themselves.

Changing the point of view

When I wear the G2 headset in a position where my eyes are off center, slightly above or below, I can see clear sharp text but with chromatic aberration like this.

NB: just look at the “Quality of your visuals” text.

I believe in positioning my eyes above, I can see with a different focal distance due to the shape of the lens putting my eyes farther from its surface. It also requires looking “slanted” which I’ll try to illustrate below.

Normal viewing:

      /
     /
 < --|------------------------------------- A
     \
      \

Slanted viewing:

      /
 < --/....._______
     |            `````--------......._____
     \                                     ```` ------ A
      \

This is why I get maybe sharper text because it changes focal distance, but conversely I also get chromatic aberration because it is not calibrated for off-center viewing.

Partial Solution

So I can either see no-CA but blurry unless I add +1 correction, or I can see sharper but with CA without +1 correction.

This is where the undocumented Dwm registry keys are helping:

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\Dwm\ExtendedComposition\ColorDistortion
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\Dwm\ExtendedComposition\ColorDistortionB
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\Dwm\ExtendedComposition\ColorDistortionG
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\Dwm\ExtendedComposition\ColorDistortionR

NB: they are all REG_DWORD

If I set ColorDistortion to 1 (enable) and I change the scaling of blue to 0.995 (value = 995) and scaling of red to 1.005 (value = 1005) I can get this:

NB: this corrects the red blue fringes from the shot above but in the headset it is really sharper than this.

With this I’ve been able to fly for 1H+ without eye strain and if I forcefully relax my eyes it is slightly but only slightly becoming blurry, which is what I’d expect. However the disk of clarity is maybe a little smaller than when wearing +1 glasses,

Conclusions so far

If I wear the headset so that the anti-CA filter is properly aligning the RGB places producing white lines without fringes, the text is blurry. If I wear the headset offset from center so that text is sharper I get CA fringes. If I adjust the R and B scaling I can compensate the fringes mostly alright. In the 1st case wearing +1 glasses makes the image as clear as the 2nd case with RB adjustment (nearly).

This poses a few questions:

  • do you see the same R/B fringes when it is sharp with your G2?
  • If you do see the fringes and reposition the headset vertically until they are no longer visible, is the text becoming blurrier?
  • Is there any other G2 user here which experience matches what I’m reporting here?