fixed 2.9.6 quad views foveated Varjo Aero issue under 90fps

[mbucchia]

3 hours ago, gonvise said:

@mbucchia, when arrive the long awaited correct FOV submission (Thanks ED), and taking into account that you made a work-around to "bypass" this ED's bug in your QVFR... it will it still work or will it need some adjustment?

Good question, and it shouldn't matter whether the workaround is still active. Also, assuming ED didn't revert the app name change, then the workaround won't be activated anyway.

What matters more is to remove the unsigned DLL (for those who installed it) and reinstall the original (signed) qvfr.

Edited July 15, 2024 by mbucchia

[Panzerlang]

8 hours ago, mbucchia said:

Can you describe "fix"? Is it going to be the real thing (correct FOV submission) or just reverting the name of the app back to "DCS World"?

Thanks.

Hi Mbucchia. First I want to thank you for your passion and dedication to a piece of work you have freely provided that has given so many people so much pleasure. You, sir, are one of those rare individuals who selflessly give.

I have a question regarding my use of the Sandboxman's fix-file (XR_APILAYER_MBUCCHIA_quad_views_foveated.dll). With the original file in place (your install of quad view) I had micro jerking/stuttering of ground objects (buildings and trees etc) at low level as I looked left/right as I flew past them. With his replacement file, that micro-jerking totally disappeared. Absolutely smooth and flawless performance. 
As far as I can try to understand it my guess is the use of OpenXR Toolkit changes the properties of the .dll (XR_APILAYER_MBUCCHIA_quad_views_foveated.dll). Is that correct or is that file immutable (not able to change)?

TIA for any insight.

[mbucchia]

9 minutes ago, Panzerlang said:

Hi Mbucchia. First I want to thank you for your passion and dedication to a piece of work you have freely provided that has given so many people so much pleasure. You, sir, are one of those rare individuals who selflessly give.

I have a question regarding my use of the Sandboxman's fix-file (XR_APILAYER_MBUCCHIA_quad_views_foveated.dll). With the original file in place (your install of quad view) I had micro jerking/stuttering of ground objects (buildings and trees etc) at low level as I looked left/right as I flew past them. With his replacement file, that micro-jerking totally disappeared. Absolutely smooth and flawless performance. 
As far as I can try to understand it my guess is the use of OpenXR Toolkit changes the properties of the .dll (XR_APILAYER_MBUCCHIA_quad_views_foveated.dll). Is that correct or is that file immutable (not able to change)?

TIA for any insight.

The change in the unsigned DLL only enables the workaround for ED's bug like it did before. Your micro-stuttering issue is unrelated to any of it.

There is no connection between Quad Views and OpenXR Toolkit. 

[Panzerlang]

28 minutes ago, mbucchia said:

The change in the unsigned DLL only enables the workaround for ED's bug like it did before. Your micro-stuttering issue is unrelated to any of it.

There is no connection between Quad Views and OpenXR Toolkit. 

Thank you M. It'll forever remain a mystery then. Obviously I'll be very leery of installing ED's patch (or replacing the 'fixed' file with the original), although at some point it'll become inevitable.

Edited July 16, 2024 by Panzerlang

[Hunter Joker]

14 hours ago, BIGNEWY said:

correct FOV submission, the team have tested and it is looking good. 

thank you  

Hi, can You consider to release it as soon as possible as a hotfix, so detached from the rest of the others fix, cause from last patch VR is pretty unusable and I didn't even was able to see the new map from Friday...
Thx.

 

[BIGNEWY]

38 minutes ago, Hunter Joker said:

Hi, can You consider to release it as soon as possible as a hotfix, so detached from the rest of the others fix, cause from last patch VR is pretty unusable and I didn't even was able to see the new map from Friday...
Thx.

 

Hi, when we have a date for the next patch we will let you know but we are looking at other issues created by 2.9.6 also, as soon as we have the major issues fixed we will likely patch.

thank you 

[ColinM9991]

21 hours ago, BIGNEWY said:

correct FOV submission, the team have tested and it is looking good. 

thank you  

This is great news, thank you for sharing that.

Many thanks also to @mbucchia for his efforts in documenting and detailing the issue as well as providing the workaround that demonstrated the benefits of quad views over the past year. I believe this has helped to get the driving force of the community behind this ask, since people then saw massive improvements to their sessions.

This isn't for the sake of having a whinge but to really try to inspire a small change.

There are some really talented people in the DCS community (I'm not one of them). It's a shame to see people become frustrated with the bug reporting process when it's as detailed and articulated as mbucchia's report.

I wish there was some time spent retrospectively looking at how to improve bug triaging and prioritisation for things such as this. I completely understand that the development team have other priorities. There are many of us who actively work in development teams as developers, testers etc so we're well aware of that. We're also aware that it's important for teams to take feedback and improve processes - anybody feel free to correct me if I'm wrong here.

There needs to be an improvement around how bugs are prioritised and subsequently pulled in to be worked on - particularly in this case where a lot of the legwork has already been done by somebody like mbucchia. It shouldn't take a year to get to this point of having a resolution to a long standing implementation issue.

Nonetheless, it's great that a fix is on the way. It's clear that the collective minds at ED have really focused on this over the past week and the result is a net positive for the players.

Edited July 16, 2024 by ColinM9991

[Tomcat388th]

Question to the smart people.  In simple terms, "correct field of view submission" what does this mean? Will this help with performance any compared to what we have been using?

J

[mbucchia]

2 hours ago, Tomcat388th said:

Question to the smart people.  In simple terms, "correct field of view submission" what does this mean? Will this help with performance any compared to what we have been using?

J

No, it's purely a "correctness" change, at any point in time in DCS MT there may be data from 2 "in-flight" frames, and DCS was incorrectly mixing up the data used for the foveated rendering, causing the visual glitches when the foveated viewport followed your eye gaze. There is no performance implications with the change.

See full technical details:

 

[Ce_Zeta]

Thank you Mark for all your work and effort spent to improve VR experience. Also the honesty shown in this post. I hope that you can enjoy life and be fueled with new passion in your new job.

 

[Tomcat388th]

6 hours ago, mbucchia said:

No, it's purely a "correctness" change, at any point in time in DCS MT there may be data from 2 "in-flight" frames, and DCS was incorrectly mixing up the data used for the foveated rendering, causing the visual glitches when the foveated viewport followed your eye gaze. There is no performance implications with the change.

See full technical details:

 

Thank you Sir

J

[Werewolf_fs]

En 12/7/2024 a las 9:10, mbucchia dijo:

It's possible (but unlikely) that the game engine is now rendering left and right eye in a different order. I highly doubt this really works, all of the previous experiments with this showed that either FR wasn't really applied (no gain) or FR was incorrectly applied in certain situations (unexplained blurriness in the middle of the screen).

For VRS to perform better than QVFR, you must have had extremely low CPU headroom, which isn't the case of the large majority of QVFR users.

@mbucchia. You are absolutely right, and I just see the reason for this situation. With QVF the configuration (.ini) makes a slight supersmaplin, and for me it operates at 4259x4382, but when using VRS foveated in OpenXR Toolkit I operate at 3872x3984, so the performance is similar, although still slightly lower with VSR Foveated.

But with QVF what I do notice is a slight suttering that does not happen to me with foveated VRS. CPU limitations?.  Also the eye tracking of Foveated VRS is apparently faster, it is practically not noticeable if you look from one side to the other quickly, while the QVF is seen a few ms.

My configuration is, Intel i9 13900KF, 64gb, RTX4090.

Edited July 18, 2024 by Werewolf_fs

[lolof1]

HI Matthieu Bucchia and all 

In any case, apart from this indescribable mess, I thank you again Mr Bucchia for all your work for our common passion for simulation.
Spend your private time on your hobbies with your wife.
Now we're going to wait for Ed's patch which will work with #quadview or not, and we'll see when it arrives, which could last a very, very long time.
I haven't installed the #Sandfox #Dll because I don't want to have crash systems with my other simulators like #Iracing /Il2 Bob in my PC.
Especially iracing which uses #openXr natively, has an anticheatcod.
This could make these games unstable due to the #sandbox #Dll not being digitally signed.
To finish your inventory of the Pcvr simulation with headsets other than Meta send shivers down your spine!

Thank you again Matthieu for all your work.

 

 

aplushhhsss

[Ahdidaas]

Agree w/ the support to mbuccia -- QVFR is the reason I got a quest pro and am enjoying VR DCS so much.  Ty sir!

[Mainstay]

Kinda annoys me personally this cant be rolled out in a hotfix...... 

[LePaul]

4 minutes ago, Mainstay said:

Kinda annoys me personally this cant be rolled out in a hotfix...... 

Same, I can't fly DCS without my VR  

On a side note, they should hire @mbucchia as a developer!

[Phantom711]

@lolof1

One could use a mod manager like OVGME to switch between the regular version and the „dirty fix“ for QVFR.

[mbucchia]

Is DCS ST that bad btw? Cause that would be the easiest workaround, switch from MT to ST. Does ST still have the "cursor bug" double-vision?

[Glide]

I just noticed that the Max FPS slider is now available from the options menu while the sim is running.  Maybe it was there before, but it made me realize an easy way to tune performance while we wait for the next patch. 

Without any frame caps, or smoothers, or latency modes enabled, launch into the sim.  Let's say you are getting 66fps flat out.  Press Esc, drag the Max FPS slider to 60 (give your GPU 10%), and resume.  Doesn't matter what your graphics settings are or the mission load, you can use the same approach.  Hope this helps.

[PawlaczGMD]

Just to add on to the noise - with the current DCS performance, Quadviews is absolutely critical to play the game on high-end VR headsets. ST is unplayable in VR for me, and I don't think a GPU that can achieve decent FPS at decent settings without QV exists. I will try the fix posted here, but if it doesn't work, I can't play DCS until this is fixed. 

This tech should be supported officially, and I am more surprised that AMD themselves didn't pick it up rather than ED.

[Glide]

44 minutes ago, PawlaczGMD said:

Quadviews is absolutely critical to play the game on high-end VR headsets

I disagree.  I am running my Pimax Crystal at 66% resolution with DLAA on my 3080Ti, and I get 90fps in the main menu, rock solid 40fps across all my busy missions and 60fps on the light ones.  I could go higher with DLSS, but I don't need to because it's smooth as silk.  I'm not even using DFR or eye tracking.  The DCS engine has never been better, not saying it's perfect. 

[PawlaczGMD]

24 minutes ago, Glide said:

I disagree.  I am running my Pimax Crystal at 66% resolution with DLAA on my 3080Ti, and I get 90fps in the main menu, rock solid 40fps across all my busy missions and 60fps on the light ones.  I could go higher with DLSS, but I don't need to because it's smooth as silk.  I'm not even using DFR or eye tracking.  The DCS engine has never been better, not saying it's perfect. 

>66% resolution

Sorry, but that's low. I want to run above 100% in the center region, as it improves visuals massively. If I wanted to use 66% of the resolution I would have bought a $500 Quest 3 instead of a $1600 Crystal.

I'm not necessarily criticizing the DCS engine, but it is one of the heaviest games out there, and QV is an essentially free and massive performance boost. It's critical to the VR experience. 

[Glide]

6 hours ago, PawlaczGMD said:

If I wanted to use 66% of the resolution I would have bought a $500 Quest 3 instead of a $1600 Crystal.

The graphics card is the next upgrade.  One day.

[WipeUout]

On 7/19/2024 at 7:20 PM, LePaul said:

On 7/19/2024 at 7:15 PM, Mainstay said:

Kinda annoys me personally this cant be rolled out in a hotfix...... 

Same, I can't fly DCS without my VR

This

 

For the heck of it, I got the old track-IR back up yesterday and could handle it more than 2 minutes... Pancake Simming no more.

@mbucchia : Many thanks for your devotion to this community, you should get paid for what you did!!!

Edited July 22, 2024 by WipeUout
Typo

[AlessandroFuBu]

Hi all, do you think that Quad views is broked also in fixed foveated? I mean I see the good looking center square and the bad one in the peripheral but it means that I get some performance boost? I have just bought the quest 3 and just discovered the quad view from mbucchia and all the things that is going on so I never tried before the new dcs patch to make a comparison.