First off, very happy to have been on the beta for the VR portion of this sim and sorry I have not provided feedback until this point. I really wanted to get my thoughts together before submitting.
Here are some of the issues I am concerned with.
a) Why do you need to have a desktop display running of both eyes AND in the headset? Most VR games I know have one viewport if any running on your desktop. Seems like some wasted performance here. Take a look at Beat Saber for instance…or Half Life: Alyx. None of these games are putting both eye renders on the desktop.
b) After a LOT of a/b comparisons of real life video and in-game rendering I can definitely say the cockpits in VR are too small. There needs to be access to a world scale if this is something that is struggled with between headsets. Let us change this setting. My apologies if this has been addressed elsewhere.
c) Pilot bone structure. For some reason to me when I am flying in VR and I move only my head, it seems the cockpit and the world are moving at different parallax. Like the plane is moving around me. It seems to me that this is a product of where the origin of movement is tied to. For instance,m I should be able to move my head and my BUTT does not shift position. This is exactly what it feels like in game to me. Something just feels really off about movement in the cockpit…and it is hard to explain. Maybe its the scale… I know that in FSX with the FlyInside VR add-on I did not get this. It felt like I was in the plane. To be completely immersed in this sim and for VR to be more than a gimmick at this point, I feel this needs to be addressed. Moving my head in the cockpit should not make the plane feel as though it is going up and down and side to side around me.
With all this said, I am offering the team my help with any of this. I can make testing videos, whatever you need. Just reach out. I am retired and have 24/7 to look at this stuff and get it sorted. I want this to be successful and am willing to work to get it there.
You could well be right with that thought. I’m using an Acer WMR headset at the moment and am pretty pleased with how it’s performing. Though, I am considering going to the Reverb G2 as the increased resolution would be good.
When flying at night where there are no clouds or habitation, I really notice the effect of the backlight in my G2 … even though it is apparently one of the better VR headsets in this respects.
Doesn’t look like anyone is currently using OLED technology, even the high end gear like the Pimax have gone away from it. There are some pretty interesting new developments in micro LED and OLED display technology … so let’s hope that in future generations of headsets we can get truly dark darkness!
It depends if Mirroring to the Desktop has any impact on Performance with MSF2020, most VR games/sims there is no effect on FPS at all as it’s merely a “copy” of the HMD renderer so it could well be the same here.
Although I agree it would be better to have the Dominant Eye Only displayed and even an option to not have a Mirror at all, personally all Menus working in Pancake Mode would also be better and just leave the Viewpoint (Pilot) for VR Rendering like what Race Sims are capable of.
I am using a Samsung Odyssey+ with Index Controllers as my setup. There are no limitations or issues in any other VR games, I have close to 100 including most AAA titles out there.
Do you by any chance run any other game or application using the OpenXR API?
Otherwise it is most likely you’re using SteamVR and WMR for SteamVR and there shouldn’t be much problems in this case and this might explain why you’re seeing this kind of problem only with FS2020 (because it is the sole OXR title you’re using).
NB: If you’re using SteamVR OXR instead of WMR OXR with the G2 and the O+, do you see any difference with these issues?
A) Fully agree. Either blank the screen to save GPU resources, or show an image someone can watch or that is recordable. The mirror of the VR headset is both useless and wastes GPU load. Or better yet, give the user the option (performance or second image).
B) Agree. Biggest immersion killer IMHO. The 172 interior is tiny…maybe 50% to 60% of real life. Yes I’ve been in one, yes I’ve flown one. If you move your head close to the yoke or switches, you can see the tiny toy scale of the yoke. Look down at the copilot seat. It’s a kindergarten chair. IRL, it’s not exactly a La-Z-Boy, but it is adult sized.
There’s a math error somewhere in the calculations. The G2 provides a value (in mm) for the IPD slider setting the user has set. Either use that correctly or give us a world size slider.
As a check, sit in the 172 cockpit in VR. Stand up, push your chair back, and walk toward the back of the plane by the rudder. Take off the headset. Look back at your desk. Is that how big a 172 is? Hint: a 172 is not 5 feet from pilot seat to tail.
It is very easy: there is no cost at all (so marginal measuring in micro second - no ‘s’ at the end).
Copying the render buffer to the front buffer is a buffer swap.
Copying the render buffer into another buffer is a blit operation for which there is dedicated silicon in the GPU doing just that.
This means if you don’t use it, you won’t ‘free’ resources for something else because it will be just silicon idle doing nothing otherwise (like Tensor cores in RTX being there and doing nothing if not used explicitly).
The only way this would cost anything is if the game rendering loop is waiting for a vsync or similar for presenting the rendered view, and if in doing so it is stalling the VR rendering. In this case yes, hiding the window from view should help, not because you won’t draw, but because the application won’t wait to draw.
PS: of course you can’t ignore the possibility there is a NVidia driver bug causing it stalling the VR direct mode when displaying anything in a window but honestly, I don’t see them introducing such regression bug in their drivers anytime soon.
Yesterday evening, Neofly sent me on a checkride in my PA44 Seminole to an airport that was unlit, at night.
Believe me, it was ■■■■ dark! I didn’t for a moment think that it looked any less dark than real night time and it was the most challenging landing I have done yet in FS2020. I had to overfly the airfield at around 50 feet to spot the lie of the runway, do a quick circuit and then find it again on approach, which was not easy. So I’m not sure I would agree with the not dark bit.
I know it’s difficult to compare what you see in the headset to what’s onscreen, but this doesn’t look too different:
Yes - no problem with the display or what MSFS shows. I was heading on an OnAir contract to northern Finland. It was really dark … apart from the celestial majesty of the Milky Way overhead. But there was this kind of vague light … sort of like someone was shining a torch at the back of your head…
Sure technology will fix it in time. my credit card is waiting and anxious!
: