fixed 2.9.6 quad views foveated Varjo Aero issue under 90fps

[Rexxxy77]

Just tried taking off from Kandahar Afghanistan in the cold and dark quick mission (hornet). This is a heavily populated mission around the airport and depending one where I am looking I get below 90fps. 90 FPS and above all is well but below 90fps the dynamic foveated renderring box (centre) jumps when moving your eyes. very disorentating. completely fine at 90fps though.

[BIGNEWY]

quad views issue reported thank you. 

 

[Hunter Joker]

It seems also Pimax Crystal affected.

Inviato dal mio RMO-NX1 utilizzando Tapatalk

[T24_Martin]

same for me. Please fix with a hotfix, or the whole update can't be enjoyed in all its beauty!

[mbucchia]

@BIGNEWY this is the same bug I reported to you guys over a year ago, with incorrect FOV submissions in xrEndFrame().

There is a workaround to YOUR bug in Quad-Views-Foveated, but it looks for the application name being "DCS World", but it looks like you changed that string to "DCS", so the workaround doesn't kick in.

Please fix the real issue anyway... your code is violating the OpenXR standard.

Reported originally March 2023 with a very clear explanation (point 2):

 

Edited July 11, 2024 by mbucchia

[zildac]

@BIGNEWY this is the same bug I reported to you guys over a year ago, with incorrect FOV submissions in xrEndFrame().
There is a workaround to YOUR bug in Quad-Views-Foveated, but it looks for the application name being "DCS World", but it looks like you changed that string to "DCS", so the workaround doesn't kick in.
Please fix the real issue anyway... your code is violating the OpenXR standard.

Bit cheeky, but could you perhaps add a fix in your QVFR, if they haven't addressed something so fundamental called out over a year ago I'm not holding my breath for it to be sorted anytime soon

[mbucchia]

5 minutes ago, zildac said:

Bit cheeky, but could you perhaps add a fix in your QVFR, if they haven't addressed something so fundamental called out over a year ago I'm not holding my breath for it to be sorted anytime soon

I cannot release any new version because my signing certificate expired in February and the cost of digital signing has gone from $200 to >$1,000 in 2024, so I'm not getting a new one.

 

I'm also real real real tired of repeating myself ED

 

And specifically initial report point 2) here (March 2023):

 

Edited July 11, 2024 by mbucchia

[mbucchia]

Note that the bug is specific to ED'S multi-threaded code (they are submitting data from the wrong frame, a blantant violation of the OpenXR standard), so the ST version of DCS isn't affected (someone confirmed to me that it worked).

Also, if you are on Varjo, you could disable Quad-Views-Foveated and instead use Varjo-Foveated, because that tool did not have the conditional check on the app name (unlike Quad-Views-Foveated, the Varjo-Foveated tool is pretty much only useful for DCS, so I never checked which app is running before activating the workaround).

Edited July 11, 2024 by mbucchia

[zildac]

19 minutes ago, mbucchia said:

I cannot release any new version because my signing certificate expired in February and the cost of digital signing has gone from $200 to >$1,000 in 2024, so I'm not getting a new one.

 

I'm also real real real tired of repeating myself ED

 

And specifically initial report point 2) here (March 2023):

 

 

I completely understand, and appreciate your frustration. @BIGNEWY I know you've submitted this as a bug to the team, but can you please ensure they understad the context and the detail as outlined by @mbucchia and the fact that in the MT binary version they are doing OpenXR "wrong"  and are not compliant with the standard. There are a large number of VR users thast rely on this implementation by Matt to achieve acceptable performance in VR. It would be a real shame to let this one drag on, as it has been and absolute blessing for many many months.

@mbucchia Frankly, given what you've already done I'm sure the community who have been reaping the benefits of your work for the last couple of years would be able to cover the code signing cert cost plus some Kickstarter!

 

 

Edited July 11, 2024 by zildac

[javelina1]

rats....  and here I was all excited about the new patch, and planning to fly the Rhino after I'm off work.  (With my Aero)...

[markturner1960]

As title...firstly quad views..had to un install as it was iun useable with it on...weird distortions and jaggy vibrating images where the hi res window is looking.

Only getting 25 - 35 FPS in instant action now when before, it would be 36 - 60 

i thought the patch was delayed to prevent showstoppers like this? 

[Bones308th]

I have the Aero. Not updating until this is addressed. 

[zildac]

Unfortnately:

 

[VR Flight Guy in PJ Pants]

It seems to me QVF was trying to track but failed, resulting the screen "shaking".

I concur the overall FPS has dropped when there is a lot going on.

[zildac]

12 minutes ago, Bones308th said:

I have the Aero. Not updating until this is addressed. 

You could just use Varjo Foveated, it still works. Quad Views Foveated has the issue.

[mbucchia]

1 hour ago, mbucchia said:

There is a workaround to YOUR bug in Quad-Views-Foveated, but it looks for the application name being "DCS World", but it looks like you changed that string to "DCS", so the workaround doesn't kick in.

FYI this is the specific line that checks for DCS and enables the workaround to their incorrect multi-thread frame submission logic:

https://github.com/mbucchia/Quad-Views-Foveated/blob/957ff0327185ea29acc41fc86c4cdc3caf428fb9/openxr-api-layer/layer.cpp#L175

The new version of DCS seems to report "DCS" instead of "DCS World".

As mentioned above, while this looks like an easy change for me, it isnt:

1) ED needs to provide the real fix, as explained in March 2023 and again in December 2023. If the real issue had been addressed at the source, this wouldn't be happening today (since there would not be a need for a game-activated workaround). Changing the name in QVFR only opens the door to it breaking when ED decides to use "Digital Combat Simulator" as the app name in the future. Always activating the workaround creates a maintainability issue where the workaround could potentially break non-DCS apps.

2) I don't have a code signing certificate today, so any release I'd make would break anti-cheat software (not in DCS, just ANY OpenXR game installed on your machine), which isn't something I'm going to do (cause then that is more fallout coming in my direction, for something that isn't even my problem). Acquiring a code signing certificate with Cloud Signing is >$1000 AND also several hours setting up a token vault and all of the other goodies that the digital government now requires.

 

Not really sure what is the point of me spending hours of my time writing all this this stuff and explaining best practices and making detailed bug reports, pointing out issues, if this work just goes down the trash can without probably being read at all

Edited July 11, 2024 by mbucchia

[zildac]

5 minutes ago, mbucchia said:

FYI this is the specific line that checks for DCS and enables the workaround to their incorrect multi-thread frame submission logic:

https://github.com/mbucchia/Quad-Views-Foveated/blob/957ff0327185ea29acc41fc86c4cdc3caf428fb9/openxr-api-layer/layer.cpp#L175

The new version of DCS seems to report "DCS" instead of "DCS World".

As mentioned above, while this looka like an easy change for me, it isnt:

1) ED needs to provide the real fix, as explained in March 2023 and again in December 2023. If the real issue had been addressed at the source, this wouldn't be happening today (since there would not be a need for a game-activated workaround). Changing the name in QVFR only opens the door to it breaking when ED decides to use "Digital Combat Simulator" as the app name in the future. Always activating the workaround creates a maintainability issue where the workaround could potentially break non-DCS apps.

2) I don't have a code signing certificate today, so any release I'd make would break anti-cheat software (not in DCS, just an OpenXR game installed on your machine), which isn't something I'm going to do (cause then that is more fallout coming in my direction, for something that isn't even my problem). Acquiring a code signing certificate with Cloud Signing is >$1000 AND also several hours setting up a token vault and all of the other goodies that the digital government now requires.

 

Not really sure what is the point of me spending hours of my time writing all this this stuff and explaining best practices and making detailed bug reports, pointing out issues, if this work just goes down the trash can without probably being read at all

 

Understood and hadn't really taken the wider impact of an unsigned release into consideration. As an aside, for me even Varjo Foveated is borked. Just tons of:

2024-07-11 21:37:33 +0100: xrEndFrame failed with XR_ERROR_HANDLE_INVALID messages in the log after the

2024-07-11 21:37:31 +0100: Creating swapchain with resolution: 2048x2048 message

A DCS timeout for me I think!

46 minutes ago, mbucchia said:

Note that the bug is specific to ED'S multi-threaded code (they are submitting data from the wrong frame, a blantant violation of the OpenXR standard), so the ST version of DCS isn't affected (someone confirmed to me that it worked).

Also, if you are on Varjo, you could disable Quad-Views-Foveated and instead use Varjo-Foveated, because that tool did not have the conditional check on the app name (unlike Quad-Views-Foveated, the Varjo-Foveated tool is pretty much only useful for DCS, so I never checked which app is running before activating the workaround).

 

This doesn't appear to work for me

2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1908x1908
2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1908x1908
2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1824x1824
2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1824x1824
2024-07-11 21:37:31 +0100: Application is using Quad Views for this session.
2024-07-11 21:37:31 +0100: Creating swapchain with resolution: 2048x2048
2024-07-11 21:37:31 +0100: xrEndFrame failed with XR_ERROR_HANDLE_INVALID

 

Edited July 11, 2024 by zildac

[mbucchia]

9 minutes ago, zildac said:

Understood and hadn't really taken the wider impact of an unsigned release into consideration. As an aside, for me even Varjo Foveated is borked. Just tons of:

2024-07-11 21:37:33 +0100: xrEndFrame failed with XR_ERROR_HANDLE_INVALID messages in the log after the

2024-07-11 21:37:31 +0100: Creating swapchain with resolution: 2048x2048 message

A DCS timeout for me I think!

This doesn't appear to work for me

2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1908x1908
2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1908x1908
2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1824x1824
2024-07-11 21:37:28 +0100: Creating swapchain with resolution: 1824x1824
2024-07-11 21:37:31 +0100: Application is using Quad Views for this session.
2024-07-11 21:37:31 +0100: Creating swapchain with resolution: 2048x2048
2024-07-11 21:37:31 +0100: xrEndFrame failed with XR_ERROR_HANDLE_INVALID

 

 

Reminder that Varjo-Foveated requires you to turn off OpenXR Toolkit, in case that's your issue here. 

[Hawkeye_UK]

3 hours ago, BIGNEWY said:

quad views issue reported thank you. 

 

How can you not have anyone with a Pimax Crystal on your testing team.  Painful - i have a Quest 2, Reverb G2 and a Crystal sat on my desk for god's sake and im not even an IT company or have any professional interest in this sphere.

You like to say your embracing the future and VR is important (as per wags interview some time back) how can you credibly say this without even testing the basics - ONCE AGAIN the internal testing team have let game breaking issues through to release, and its not like we even have the option to revert to "stable".

What is going to be done, and more importantly when.

[zildac]

2 minutes ago, mbucchia said:

Reminder that Varjo-Foveated requires you to turn off OpenXR Toolkit, in case that's your issue here. 

Thanks Matt, I found the issue. I checked OXRTK and DCS.EXE was indeed enabled , because guess what, it used to be called DCS World and it was automatically enabled as a result of the name change! That's going to catch a few people out I suspect.

 

[mbucchia]

2 minutes ago, zildac said:

Thanks Matt, I found the issue. I checked OXRTK and DCS.EXE was indeed enabled , because guess what, it used to be called DCS World and it was automatically enabled as a result of the name change! That's going to catch a few people out I suspect.

 

Ahaha yes that's right. This rename is also going to be another source of issues for that scenario

[zildac]

2 minutes ago, mbucchia said:

Ahaha yes that's right. This rename is also going to be another source of issues for that scenario

Definetely!

[Bones308th]

38 minutes ago, zildac said:

You could just use Varjo Foveated, it still works. Quad Views Foveated has the issue.

Thanks for the reply man. How do I change it if I have to? Sorry man, I have no idea. I just turn it on and go! lol What is the difference between the two?

 

[zildac]

2 minutes ago, Bones308th said:

Thanks for the reply man. How do I change it if I have to? Sorry man, I have no idea. I just turn it on and go! lol What is the difference between the two?

 

If you have a Varjo headset, just uninstall Quad Views Foveated and install Varjo Foveated and set the config to your liking:

https://github.com/mbucchia/Varjo-Foveated/releases/tag/0.2.0

All the info re config can be found in the Wiki including wich DCS settings need to be disabled or cause issues.

https://github.com/mbucchia/Varjo-Foveated/wiki

 

And make sure you disable DCS and DCS World in OpenXR toolkit (if you have it installed)!!!

 

 

 

[edmuss]

1 minute ago, Bones308th said:

Thanks for the reply man. How do I change it if I have to? Sorry man, I have no idea. I just turn it on and go! lol What is the difference between the two?

 

Uninstall quadviews foveated and install varjo foveated

Performance is typically slightly better with varjo foveated anyway. Ensure you disable openxr toolkit!

Read @mbucchia's excellent wiki on his GitHub page and it will all be good. If you have 3080/90 I would recommend 39ppd in varjo base, 1.0 focus and 0.6 peripheral in the config file with DLAA to quell the shimmers.