community advice Dynamic Foveated Rendering - Everything in one page

[mbucchia]

5 hours ago, CidTheViking said:

I have an odd issue with DFR, I have installed all fixes but it seems that DFR will work for a short period and just stop. The error I am getting is

2023-10-20 07:45:58 -0400: xrBeginFrame failed with XR_ERROR_CALL_ORDER_INVALID
2023-10-20 07:45:58 -0400: xrGetActionStatePose failed with XR_ERROR_ACTIONSET_NOT_ATTACHED
2023-10-20 07:45:59 -0400: xrGetActionStatePose failed with XR_ERROR_ACTIONSET_NOT_ATTACHED
2023-10-20 07:45:59 -0400: xrGetActionStatePose failed with XR_ERROR_ACTIONSET_NOT_ATTACHED

Not even sure if this is related.

And just repeats that last line until it times out.

I can't use partial logs. Need full logs.

4 hours ago, witwas said:

@mbucchia: thanks for all the work.

So if I read correctly, we need to uninstall the old version and install the new one because of 2.9?

You don't need to uninstall anything. Installing a new version overwrites the previous one.

> Just released a new build 1.1.3 that fixes the remaining issues with DCS 2.9.

Edited October 20, 2023 by mbucchia

[Mr_sukebe]

Quick update.

With the 1.1.3 version I was getting a LOT of jitter with my QP.  I reverted back to version 1.1.2 and that resolved it.

The frame rate was fine.  It seemed to be set off by physical movement of my head.

Edited October 20, 2023 by Mr_sukebe

[mbucchia]

1 hour ago, Mr_sukebe said:

Quick update.

With the 1.1.3 version I was getting a LOT of jitter with my QP.  I reverted back to version 1.1.2 and that resolved it.

The frame rate was fine.  It seemed to be set off by physical movement of my head.

 

The difference between 1.1.2 and 1.1.3 is a single trivial operation that happens within the first second of loading. I'd recommend you repeat that experiment a few times, because there is no logical explanation to your statement.

[twistking]

Hello,
i've got three small questions. I've read all the projects history overview in @mbucchia's github, but i don't understand how DCS does not end up completely cpu bottlenecking itself. The CPU has to prepare not 2, but 4 different viewports. Is the renderer in DCS smart enough to know that one pair is just a zoomed in variation of the other pair and therefore the amount of additional computation is so low?

Is there a simple reason why there is no way to directly interface a quadview-renderer with openXR? i'd thought the whole thing openXR has going, is being a standard for everything VR/AR. Why would they not support it natively?

Do i understand it correctly, that the original varjo foveated was an extension to the openXR standard (like a addon?) and the "quad views" from the github is also an extension to openXR that interfaces with DCS in the same way, but then uses the openXR compositor to feed a regular stereo image to the f.e oculus quest. So DCS interfaces as it would with the varjo extension and the oculus quest interfaces normaly with openXR and the magic happens in the middle. Is this somewhat correct?

Thanks.

Ps: Great work by the way! I can't test it myself unfortunately because my pc is too weak for DCS in VR, but it seems to be a great piece of software!!!

[mrsylvestre]

Hi, @mbucchia, thanks for the quadviews update. Just in case this helps:

• QV 1.1.0, DCS 2.8 and SteamVR's OpenXR runtime or VDXR 0.6.0 -> OK
• QV 1.1.0, DCS 2.9 and SteamVR's OpenXR runtime -> OK
• QV 1.1.0, DCS 2.9 and VDXR 0.6.0 -> Crash on DCS start-up
• QV 1.3.0, DCS 2.9 and VDXR 0.6.0 -> OK

(System specs in signature, QV in FFR mode, obviously)

Happy camper this morning !

 

[Hippo]

On 10/20/2023 at 10:24 AM, Hippo said:

My very quick first impression with 2.9 is that DLAA (TAA too, but at the cost of too much blurring) noticeably reduces shimmering (I usually use MSAA x4 as I can't bear jaggies and shimmers), and that it seems to make any DFR peripheral area shimmering much less noticeable than before.  Placebo or does anyone else feel the same?  Might it be feasible to decrease the size of the focus area further?  I will test when I have more time, but wondering if anyone else has already played around with this?

 

Sorry, after further testing I was mistaken, and afaict there is no difference.  However, I'd swear that since I first tried this when mbucchia released for Quest Pro, either my tolerance or something else has changed.  I am able to set a 0.25 focus region (for a massive performance improvement) and whilst I am aware of shimmers in the periphery I can happily live with them - there is no way I felt this was the case when I first tried this out on release.

As to 2.9 DLSS, DLAA, TLA, I'm afraid it's all very much 'meh' for me, with any perfomance improvements simply not worth the visual artifacting/ degradation.  As always, there's no such thing as a free lunch (though quad views foveated is getting very close).

EDIT: Arrrrgh.  Someone help pls.  I think I'm going nuts.  I'm seeing different results between restarting sessions.  I think a restart might be necessary between changing MSAA x 4 -> DLAA.  I'm not even sure if DLAA is the cause, but I have definitely had sessions where I can detect no shimmering in the periphery, and sessions where I can.  DFR is definitely working as it is very perceptible in the menus and I am aware of the "square".  If the effect is real, this is a very big deal.

To check: in the F-18 engine gauge shimmers are very noticeable when the engine gauge is in the periphery, but I have had sessions where it doesn't shimmer at all.

EDIT Again:  Almost certainly down to DLAA.  Shimmers in the periphery eliminated vs MSAAx4.  Needs a mission restart (at least), i.e. changing the setting in-mission doesn't appear to do it.   In normal usage (i.e. not using DFR) I can discern no difference in quality or performance between MSAAx4 and DLAA.

 

Edited October 21, 2023 by Hippo
Note to self. Stop posting at the same time as you test. Test first, wait a few hours. Then post.

[slughead]

6 hours ago, Hippo said:

In normal usage (i.e. not using DFR) I can discern no difference in quality or performance between MSAAx4 and DLAA.

Try looking at moving aircraft. The trailing edges and sometimes leading edges smear/melt like a firey phoenix with DLAA/DLSS.

[SMH]

1 minute ago, slughead said:

Try looking at moving aircraft. The trailing edges and sometimes leading edges smear/melt like a firey phoenix with DLAA/DLSS.

Yeah, as expected. I haven't quite understood why everyone was in such a hurry to make their picture look poorer. 

[Hippo]

3 hours ago, slughead said:

Try looking at moving aircraft. The trailing edges and sometimes leading edges smear/melt like a firey phoenix with DLAA/DLSS.

Thanks.  I didn't go out of my way to check moving aircraft.  Sad if so as it was all looking so promising.  EDIT:  Have now checked and indeed am seeing artificating / ghosting.  But imho nothing so bad as to not be worth using DLAA + FSR (so far).

Edited October 21, 2023 by Hippo

[mbucchia]

Sorry but can you guys go discuss this on one of the 15 threads about DLSS

[EazyE]

I've switched from the Crystal to the Varjo and now I'm getting 20 - 30fps less. I left everything the same in game, mostly everything left around the default VR settings and 1.0 PD. Is this normal?  

[dburne]

13 hours ago, EazyE said:

I've switched from the Crystal to the Varjo and now I'm getting 20 - 30fps less. I left everything the same in game, mostly everything left around the default VR settings and 1.0 PD. Is this normal?  

 

Running Open XR with DFR/Eye Tracking? I use the original version 2 of DFR.

[TimSell75]

Hello,

I have just got Dynamic foveated rendering working with the Pimax Crystal. Installed everything according the TallyMouse (@GayCavalier ) manual on YouTube. Installation worked perfect, the FPS gain is phenomenal (from 45 to 90+)!

The only thing I noticed, and what is a bit annoying is, that I can see the edges of the foveated box in my left display (not visible in the right). It is a subtile change from the focus area box to peripheral, but the box edges are noticible and move with my eyes.

I have altered the Transition thickness, sharpening settings and peripheral resolution with the companion app, but it did not change. It is mostly visible when I look at a uniform background like the sky.

Has anybody noticed the same and found a solution?

Thanks so much in advance!

[DayGlow]

I've found screen space shadows makes the foveated border much more noticeable. 

[EazyE]

12 hours ago, dburne said:

 

Running Open XR with DFR/Eye Tracking? I use the original version 2 of DFR.

As far as I can tell. The eye tracking is working, I can see the blurriness in my peripheral so I believe DFR is working. No matter the settings I plug into quad views I get no change in fps. 

[witwas]

Try Night Vision, you will notice it right away. (there will be a rectangel gray box with a green circle, the box moves when you move your eyes).

If not, it is not working.

[mbucchia]

4 minutes ago, witwas said:

Try Night Vision, you will notice it right away. (there will be a rectangel gray box with a green circle, the box moves when you move your eyes).

If not, it is not working.

NO. The night vision was fixed by ED in 2.9.

It is NOT a correct way to tell if eye tracking is working.

The correct way is to follow the troubleshooting instructions on the project page.

Edited October 23, 2023 by mbucchia

[witwas]

Great to hear it is fixed!

[markturner1960]

Going to re post these questions from the original DFR thread as it seems to have got lost down on page 2: 

Can I just clarify what the recommended ( from MBucchia ideally as I guess he would know!) way to go about supersampling is? Above, someone mentioned setting everything at one, and only ramping up the resolution in the QVFR app.....seems sensible from a theory point of view? . Currently, I am running 1.3 in the occulus app and some increased supersampling in the QVFR app only, none anywhere else......Guess I can try it and see but just wondered on the tech knowledge .

In the same theme, same applies for changing the tangent multiplyer in ODT....can this be handled with the same effect by QVFR or not?

Another query...I also see mention of people making a profile in OTT tool and setting the monitor resolution to as low as possible in the profile.......Is there "Official" word on wether this is recommended? As I also read that using OTT is not a great idea as it is an old, unsupported app now.....

 

Its so confusing ! 

[Hippo]

12 hours ago, TimSell75 said:

The only thing I noticed, and what is a bit annoying is, that I can see the edges of the foveated box in my left display (not visible in the right). It is a subtile change from the focus area box to peripheral, but the box edges are noticible and move with my eyes.

Yes, I get it on Quest Pro.  It's been a thing since release.  IIRC the dev has said it's something that can only be corrected by ED making changes in DCS.  Perhaps  @mbucchia can confirm?  Since DLAA has made a significant difference with reducing shimmering, it really would be nice to get this working perfectly.  Is it worth requesting changes from ED, and how to request?  What would they have to do?

Personally it doesn't bother me that much as it's rarely very distracting.  Much more noticeable in the F-18 than the AH-64 where I can't see it at all.

Edited October 23, 2023 by Hippo

[TimSell75]

vor 10 Stunden schrieb DayGlow:

I've found screen space shadows makes the foveated border much more noticeable. 

I have tested since. Switching off SSS and all Antialiasing does not change anything unfortunately. What I do not understand is, why I only see the foveated border in my left lense, the right one is perfect. Maybe @mbucchia has an explanation for this?

Bloom, Lens effect are all off

Edited October 23, 2023 by TimSell75

[dburne]

36 minutes ago, TimSell75 said:

I have tested since. Switching off SSS and all Antialiasing does not change anything unfortunately. What I do not understand is, why I only see the foveated border in my left lense, the right one is perfect. Maybe @mbucchia has an explanation for this?

Bloom, Lens effect are all off

 

 

Have you tried re-calibrating eye tracking in your headset?

[TimSell75]

vor 45 Minuten schrieb dburne:

 

Have you tried re-calibrating eye tracking in your headset?

Yes, and I can see that it is working as the box moves around. Also frames are almost double

[dburne]

1 hour ago, TimSell75 said:

Yes, and I can see that it is working as the box moves around. Also frames are almost double

 

Any way to tell if it is an issue only with DCS or do you see that in other VR Apps?

[TimSell75]

vor 14 Minuten schrieb dburne:

 

Any way to tell if it is an issue only with DCS or do you see that in other VR Apps?

Unfortunately I do not have other Apps that support DFR at the moment.