DCS and Pimax

@ Voight: Thanks for your advise, i can buy a barely used pimax 8k here locally for half the price (that's why i am tempted).

I don't mind that it's not plug and play, i however want a flyable experience in DCS, if i have to invest in a new card i would be willing to do so, but then offcourse it needs to be playable (i read a lot of mixed reactions regarding this point).

Hmm... this seeing double, when rolling the aircraft or manouvering, I couldn´t follow.

Tried several times to follow the issue, but it´s not presence in the way it is described.

There is double vision in the way of ghosting, which appears with objects, like other planes, buildings or let´s say 3 dimensional objects on the ground and it could be observed on the horizon line.

I would say ghosting or double vision is an performance issue.

I´m no expert, just a user, but so far I observed less ghosting, with less frametime. Reducing the frametime by optimizing the settings did help me for reducing ghosting.

While there is a stable 36 FPS ( I run the Pimax at 72Hz panel frequency ) the frametime is still a variable. Frametime is the time the system ( more the GPU than CPU ) need to generate a new frame. While the CPU is much faster in framtetime than the GPU, the GPU becomes the limiting factor.

Optimum frametime in VR is 11ms. So, when flying over a city at stable 36FPS there is Reprojection, which doubles a formerly rendered image to give the impression of smooth 72Hz, this works fine, let´s say over the see, where are no 3-D objects, which make you aware of seeing the same frame twice.

But when flying close to a sykscraper or watching another plane or keeping the eye on the horizon line, you will see the doubled frame and it will stay longer to your awareness, the longer the time the GPU needs to generate the next real frame. I guess the effect is cumulative, as with reprojection, the next real frame will doubled as well and the effect of ghosting gets continously through observation.

In addition to regular flight speed or let´s say, the speed needed for the GPU for creating real and reprojected frames ( after each real frame, the next frame is a copy of that real frame through reprojection ) to give a smooth impression of movement in VR, manouvering or rolling the aircraft increases the time a new frame is needed to give the impression of smooth movement in VR, as the images while rolling or manouvering changes completely every millisecond.

As said, i´m no expert, it´s just my conclusion and I might be wrong, but to avoid this effect of double vision, I ´m pretty sure that it could be reduced - and that´s what I could have observed on my system - by reducing the frametime as close as possible to 11ms for the GPU.

With the application "fpsVR" the framtime could be measured while sitting in the cockpit in DCS.

@witwas

Oh, the 8k is additionally more demanding than the 5k+ of Pimax.

I wouldn´t go for the 8k, especially with a 1080 GTX.

Or let´s put it htis way, if the price for the second hand 8k would be in a range, in which you say "OK, let´s have a look on this device, but I´m not depending on it for my VR experience", then the 8k would be ok for me. But if you really want to use the 8k with a 1080 for DCS, then I would say: don´t buy it, as you won´t get happy with it for DCS, unless you´re running a RTX 2080Ti.

Yeah my bad, I was referring to the ghosting, it's like an old frame is being injected in with the motion smoothing or hyperpredicting a frame or some such.

And I'm rather particular with the levels of artefacting I am willing to tolerate.

The 11ms mark is for running at 90fps, the reason to drop down in hertz is to give yourself a slightly larger window, so when running at 72hz the optimal becomes a 13.88ms.

That would be ideal.

And as with all things DCS, its not so much the GPU that is the problem.

It's completely CPU bound still, because it is running only two threads, and the second only does audio..

Hence why one would se a substantial benefit from overclocking CPU frequency.

Unrelated:

People seem to thing the Index is going to run smoothly in games at 140hz, that's a frame time of 7.1ms, at a higher resolution, and with canted displays as well so a lot of games need adjusting for I'm sure, like the Pimax's.

Yeah, I'm getting my popcorn ready for that one.

You´re right, I think, refering the 11ms optimum frametime to 90Hz and 13.88ms to 72Hz...

So far I always measured lower framtime for the CPU and higher framteime for the GPU in the same moment, of which I think it is better than the other way around, as the CPU don´t become a bottleneck due to deliver data to the GPU for finally render the image.

Actually, that´s the reason, why I don´t overclock the CPU anymore, as I think the GPU is already at it´s limit at 80°C... maybe adjusting the fan speed could release a bit more freedom to the GPU. I´m a bit afraid of overheating or burning the chips. The CPU runs at 4.2Ghz on all 6 cores... maybe a benefit will be added, when Vulcan is implemented to DCS.

Yeah, me too think running the Index at full potential 144MHz is very challenging for any system, but I´m also curious about the visual fidelity 144MHz could provide. I would say, that Valve does have good reasons to design their new HMD this way, at least it´ll surely provide potential for the future of VR techniques & visual quality for VR.

BTW: "hyperpredicting" in my understanding is only marketing speech by Oculus, to make the people believe, they have invented some magic for their product. I don´t think, that there is any kind of prediction, but simply the copy of a former rendered image into the continuity of frames. The technique is the same with steamVR reprojection, maybe the algorythms differentiate and are improved by optimizied program codes with new updates for the function.

You´re right, I think, refering the 11ms optimum frametime to 90Hz and 13.88ms to 72Hz...

 

So far I always measured lower framtime for the CPU and higher framteime for the GPU in the same moment, of which I think it is better than the other way around, as the CPU don´t become a bottleneck due to deliver data to the GPU for finally render the image.

Actually, that´s the reason, why I don´t overclock the CPU anymore, as I think the GPU is already at it´s limit at 80°C... maybe adjusting the fan speed could release a bit more freedom to the GPU. I´m a bit afraid of overheating or burning the chips. The CPU runs at 4.2Ghz on all 6 cores... maybe a benefit will be added, when Vulcan is implemented to DCS.

 

Yeah, me too think running the Index at full potential 144MHz is very challenging for any system, but I´m also curious about the visual fidelity 144MHz could provide. I would say, that Valve does have good reasons to design their new HMD this way, at least it´ll surely provide potential for the future of VR techniques & visual quality for VR.

 

BTW: "hyperpredicting" in my understanding is only marketing speech by Oculus, to make the people believe, they have invented some magic for their product. I don´t think, that there is any kind of prediction, but simply the copy of a former rendered image into the continuity of frames. The technique is the same with steamVR reprojection, maybe the algorythms differentiate and are improved by optimizied program codes with new updates for the function.

Pretty good explanation of both Oculus ASW1 and ASW2 here:

https://www.roadtovr.com/oculus-launches-asw-2-0-asynchronous-spacewarp/

Note though DCS does not currently support ASW2.

Couple interesting excerpts from that article:

In ASW 1.0, Timewarp could only be applied to headset rotation, while any positional movement was left untouched. One of ASW 2.0’s big upgrades is that Timewarp can now also act on position, which means reduced positional latency too.

Oculus says that while ASW 1.0 performs well as long as an application can maintain 45 FPS, it can’t maintain smooth motion in cases where framerate drops even further.

With the new version, however, Oculus says that “even at [less than 45 FPS], ASW 2.0 can maintain an acceptable view in VR, as [Positional Timewarp’s] depth-based reprojection works more accurately for lower frame rates than ASW 1.0 extrapolation.”

It would be nice especially as demanding as DCS is in VR, if ED at some point could implement whatever is needed for ASW2 to work with it.

Would probably allow Rift users to run some higher detail while still getting acceptable performance.

Oh my god, today I need to work until midnight :)

Actually I don´t understand, why ASW 2.0 need to be supported by DCS?

SteamVR does provide this function already for some time, they name it smart smoothing.

As well, Pimax with the newest Pitool beta version provides this function and name it as well smart smoothing. With both it works fine with DCS and it is sensable, that the headmovement had become more smooth and no tearing in the cockpit with smart smoothing activated.

I don´t think, that smart smoothing will allow higher detail setting, but the experience while moving the head in VR is more improved.

Oh my god, today I need to work until midnight :)

 

Actually I don´t understand, why ASW 2.0 need to be supported by DCS?

 

SteamVR does provide this function already for some time, they name it smart smoothing.

 

As well, Pimax with the newest Pitool beta version provides this function and name it as well smart smoothing. With both it works fine with DCS and it is sensable, that the headmovement had become more smooth and no tearing in the cockpit with smart smoothing activated.

 

I don´t think, that smart smoothing will allow higher detail setting, but the experience while moving the head in VR is more improved.

Well it would be nice to be supported for the Rift users, still a few of us left.

:smilewink:

Also ASW2 should allow higher detail, as it appears would still be smooth if fps drops below 45 fps, like 40 or even 35.

Currently if drops much below 45 stutters will be an issue. So I adjust graphics settings to maintain at least the 45 fps min.

LoL... :)

No, I didn´t mean it the way, that Rift user should not be supported by DCS. :)

What I don´t understand is, why ASW 2.0 and its improvement need a special support by the application? I would expect ASW 2.0 to work with every VR application right away without the need of special implementation into each VR application separately. ... but I really don´t know, how it works in this regard, just wonder ...

LoL... :)

No, I didn´t mean it the way, that Rift user should not be supported by DCS. :)

 

What I don´t understand is, why ASW 2.0 and its improvement need a special support by the application? I would expect ASW 2.0 to work with every VR application right away without the need of special implementation into each VR application separately. ... but I really don´t know, how it works in this regard, just wonder ...

I believe it is called Depth Buffering that it needs - some games already have that support built in, and some do not. Hope I am remembering that correctly.

Oculus users can check if their game supports ASW2 easily.

Whilst in game press the Oculus button to bring up the Dash panel. If the game image stays displayed whilst Dash is open, it already supports ASW2. If the game image fades to gray background whilst Dash panel is up, it does not.

Ah, understand...interesting, ...

Wags mentioned, that there is good support by Oculus for ED, so maybe support for ASW 2.0 could be expected more sooner than later. Fingers crossed ...

Since ASW is not a Pimax solution, perhaps it's discussion deserves a non Pimax thread as opposed to this DCs and Pimax thread.

Anyone else experiencing freezing issues with the Pimax?

Every now and again (perhaps after half an hour of play) I will get a freeze in DCS with the 5k. The image in the headset will stay frozen, DCS is frozen for a few seconds, but recovers. The headset does not immediately recover and will be grey. Only after moving the headset around to wake up the tracking, will the headset resume as normal. I can continue flying after this, presuming the aircraft hasn't crashed.

In single player I can press escape to pause the game, but in multiplayer this is often fatal.

I am wondering what is causing this.

Could it be only due to a loss of tracking by the headset? (I'm using two HTC base stations)

Or maybe it's related to the micro-freezes that DCS frequently has? I got these a lot with the Rift, but tracking resumed after a few seconds. Maybe the Pimax has a harder time recovering?

I'm thinking it might also be my graphics card overclock. I'm going to try without OC and see it that's the issue.

Anyone else getting this?

Honestly, first thing I would try, might be to scale down your overclock primarily as a way to at least rule that out.

I haven't had any freezes or major stuttering in the pimax or rift, running an i7 8700k @ 4.7ghz 32GB @ 2900, initially rated for 3200, but being to kits of 16 I don't get post at 3200, and the occasional bsod at 3100-3000.

And a 1080ti that I so far has left on stock.

I do have some rather annoying ghosting on aircraft roll in the pimax but that's about it really.

I had some funny greyouts with my vive when I was fiddling with my base station layout. I have quite a small computer room and I think they were too close and I was getting FOV issues, or potentially base-to-base communication issues.

Make sure there are no nearby obstructions that could limit FOV, make sure they have clear line of sight to each other at all times and that there are no nearby reflective surfaces like glass cabinets or windows.

Guys, how does the pimax 5k resolution and fill factor compares against the rift s?

I mean at least in theory until some of you try it out. I know the pimax has greater resolution but also much better FOV. So how should it be in ternms of SDE and reading clarity against the rift S?

And finally, am i good with a 1080ti and 8700k?

Enviado desde mi SM-G950F mediante Tapatalk

Guys, how does the pimax 5k resolution and fill factor compares against the rift s?

 

I mean at least in theory until some of you try it out. I know the pimax has greater resolution but also much better FOV. So how should it be in ternms of SDE and reading clarity against the rift S?

 

And finally, am i good with a 1080ti and 8700k?

 

Enviado desde mi SM-G950F mediante Tapatalk

Not sure how many P5K+ owners will pony up for a Rift S for comparison.

Actually I think some will do. In here people buy heads0ets as if they were free xd, and actually people who want to play vr games other than sims or have children, might give the rift s a go.

Enviado desde mi SM-G950F mediante Tapatalk

Buenas,

for a comparison in practice, I think we have to wait until the Rift S is delivered.

Theoretically, with regard to display resolution, I would say, the higher the display resolution, the better clarity and less SDE.

**Some News**

If Pitool and/or the hardware goes open source, is this something ED can use to improve the sim with the Pimax?

I like very much Pimax company´s philosophy... it´s quite the opposite of using VR as a cash cow in a billion dollar market.

Anyhow ... not sure if going open source will bring real advantage for DCS with Pimax, but it´s for sure the right signal, like VR should be developed as an open standard.

Open source definitely can’t hurt, so I’m optimistic and count it as a win.

Sent from my iPhone using Tapatalk

W O W....This l@@ks interesting?????

Ya weird