DCS and Pimax

[Pinello]

Remember that the Reverb does not sport as wide a FOV as Pimax. It's strength is concentrated in 1st gen FOV and very high resolution.

 

 

I know, thats why we need a version with the Reverb resolution and a 150-170 FOV..

[Euan Emblin]

You will be waiting some time for the tech to catch up with that expectation in a consumer level product.

[boedha68]

You will be waiting some time for the tech to catch up with that expectation in a consumer level product.

 

Another three years.. just as rift cv1 to rift s. :D

[hansangb]

Another three years.. just as rift cv1 to rift s. :D

 

 

I would say 18 months. Typically, first to second gen is slower than 2nd to 3rd. I'm hoping!

[Pinello]

You will be waiting some time for the tech to catch up with that expectation in a consumer level product.

 

 

It was called the Pimax 8KX, and was a part of the Kickstarter. But Pimax has gone very quiet about this version..

[Euan Emblin]

More of a long term goal, they are just not there yet with the technology to bring such a product to the consumer market.

[Rosebud47]

For better clarity/less blurriness set in Nvidia control panel "no scaling".

Nvidia control panel-->Adjust desktop size and position-->scaling mode--->no scaling

[Rosebud47]

New official Pitool Version 1.0.0.132 available for download.

 

It notes:

Issue Fixed：

 

Fix the occasional ghosting problem caused by the high GPU occupancy

 

https://pimaxvr.com/pages/pitool

 

...sounds good, looking forward to check with DCS in the evening.

[Buckeye]

DCS and Pimax

 

New official Pitool Version 1.0.0.132 available for download.

 

It notes:

Issue Fixed：

 

Fix the occasional ghosting problem caused by the high GPU occupancy

 

https://pimaxvr.com/pages/pitool

 

...sounds good, looking forward to check with DCS in the evening.

 

 

 

Has been out for a while now in Beta. Using it for probably a month or so with no issues.

 

 

Sent from my iPhone using Tapatalk

[cordite]

Yes, but it seems they added that ghosting fix.

[Buckeye]

Yes, but it seems they added that ghosting fix.

 

 

 

Ah, I see! Guess I will check this out next time I’m on the PC then. Thank you.

 

 

Sent from my iPhone using Tapatalk

[Rosebud47]

So far I don´t see much of a difference with ghosting than before.

With ghosting in DCS it is like before: the lower the framtetime, the less ghosting.

 

But found another interesting entry in the profile.json file ( located in C:/user/´username´/AppData/local/Pimax/runtime/... )

 

As follows:

 

"fov_outer_adjust_degree" : -2.32,

"lens_separation" : 0.08988272398710251

 

The profile.json reflects the settings you could make in Pitool, like display timing selection ( Hz ), contrast, brightness, etc.

 

The above two lines are not in the Pitool user interface, maybe there will be sliders or options in a future release of Pitool.

 

Anyhow, "field of view outer adjust degree" is set by default to "0" - I´ve edited to "-2,32", which defines the aspect ration of the large field of view, so now the render target could be set to the native resolution of the displays ( like the XTAL does ) by following setting:

 

Pitool render target: 1.0

Large FOV mode

Profile.json: fov_outer_adjust_degree : -2.32

SteamVR SS: 30% ( edit: forgot to mention that I´ve set the maximum render resolution to 4096 instead of 8192, of which I thionk is more meant for the Pimax8k not for the 5k+ )

= 2560 x 1440 render resolution

 

In consequence, there is no longer an overhead rendered through the process, which results in better performance. Measured a benefit of approx 10% - 15% or concrete approx 3ms plus in frametime ( aldo depending on the plane you´re flying! ) .

 

The point is, that when there is already an overhead with any render target, which will be multiplied as same as the render target or supersampling gets multiplied.

 

But now the aspect ratio of the large field of view could be properly defined, so there is no overhead anymore for the benefit of performance.

 

Does this make sense?

 

Anyway, if you use the Pimax in large field of view mode, I would recommend to set fov_outer_adjust_degree to -2.32

I guess the default value of "0" is more in accordance to the aspect ratio of normal fov mode.

 

The render resolution resulting of readjusting could be observed through steamVR supersample settings.

 

EDIT:

 

just made some further checks:

 

Pitool 1.0 = SteamVR 30% render target: 2560 x 1440 / coefficient 1.77

 

Pitool 1.25 = SteamVR 30% render target: 3200 x 1800 / coefficient 1.77

 

Pitool 1.5 = SteamVR 30% render target: 3840 x 2164 / coefficient 1.77 ( <--- interesting: native resolution of the Pimax 8k panels: 3840 x 2160 )

 

Pitool 1.75 = SteamVR 30% render target: 4484 x 2524 / coefficient 1.77

Edited June 1, 2019 by - Voight -

[Pinello]

In 2019, Pimax time/schedule..:

[DDSSTT]

Anyone else having issues with NVGs in the F-18... For me they seem to be too far apart and not centered, but up to the right / up to the left... Does anyone know how to move the location the NVG display?

[Highwayman-Ed]

"fov_outer_adjust_degree" : -2.32,

 

You Sir, are my new favourite poster on the forums :thumbup:

[Wicked.-]

Today I got a chance to mess with some of the options on the 5K+. My system specs are below but today I was using a new Asus RTX2080Ti instead of the 1080ti. I have a few questions. The first is IPD and fitment of the unit. My IPD I believe is about 58 and I can just use the unit without using IPD offset but the image is never really sharp regardless of where I set it. I don't wear prescription glasses but I do wear 1.75 readers to read but I've never needed them in the Rift CV1. Does anyone else need readers but not need them with the 5K?

Second question. Is judder always present? No matter how low I go on settings it's always somewhat present even though I'm getting a solid 45FPS with the DCS frame counter. Judder, I'm assuming, is when static objects seem to vibrate for lack of a better word, when rolling the jet or flying by. I've followed and tried most of the recommended settings listed in this thread but haven't been able to get rid of it.

 

Thanks for all the previous work and suggestions in this thread. Otherwise I'd have tossed it by now lol.

[ram0506]

Second question. Is judder always present? No matter how low I go on settings it's always somewhat present even though I'm getting a solid 45FPS with the DCS frame counter. Judder, I'm assuming, is when static objects seem to vibrate for lack of a better word, when rolling the jet or flying by. I've followed and tried most of the recommended settings listed in this thread but haven't been able to get rid of it.

 

 

I have the same problem. No chance to get rid of it. I think it`s related to Pimax`s smart smoothing, which still needs a lot of optimizations.

[Highwayman-Ed]

I've seen a few people complaining on the Pimax forums about juddering when using the smoothing as well and I have on occasion seen it myself and chose not to use it instead using a lower refresh rate, but it's much better since the 132 Pitool release and I haven't noticed it.

[Rosebud47]

@Highwayman-Ed

 

Actually I wouldn´t like to interrupt you posting more compliments, but thank you very much for this. :)

 

-2.32 is accurate with render target setting 1.0 and 1.25 in Pitool. To meet the exact aspect ration with a higher render target 1.5, 1.75 or 2.0, there would be need a 3rd or maybe 4th position after the comma.

 

Some thoughts on judder and ghosting ( double image effects ) in DCS:

 

Special VR games and experiences are build from scratch with minimal load for the render process respectively the GPU to get a smooth experience in VR. Also most VR experiences or games use some kind of teleporting system to move your position in the 3D World and/or snap turning to move around axis. As long as you don´t move or only slowly move in VR you´ll get a smooth experience, if the graphical fidely reflected in detailed graphics like high polygone, high amount of 3D structures and high res. textures does not exceed the capability of the GPU to change the rendered picture in time due to movement.

 

DCS is quite the opposite as it does give you quite accurate the feeling of speed in movement by flying a jet, means the rendered picture needs to change very often when flying low level.

 

-There are more structures to render over a city, than over the sea.

 

-The closer you get to the ground, the more often the rendered picture need to change to maintain the feeling of your speed.

 

-when you fly by another plane both objects move with high speed over a griund which also has to be rendered while moving

 

- most demanding is rolling in your jet while looking to the left or right of the cockpit and not straight, as the rendered picture need to change superfast in every degree of the roll to give you a smooth experience. Some jets can do a 360° roll within approx 1 second.

In this one second of rolling the render process needs to create let´s say for every degree a new image to give you the perseption of a smooth roll. The figures are not accurate and are just an example to visualize the problem.

 

So the payload for the GPU and render process receives a massive impact with superfast movements in DCS VR while having to render high polygone environments.

 

First step for a smooth experience in VR is to keep the maximum framerate of 90fps or 45fps with reprojection/ASW/smart smoothing.

 

With stable 45fps the supporting algorythms generates a copy of rendered frame, which is faster, than the frame would have been actaully rendered by the GPU, and fills with the copied frame the 45 real rendered frames to reach a perceived 90fps. This works fine as long as you don´t move too fast through the 3D environment or so saying the 3D environment does not need to change to fast to give you the impression of a fast movement.

If you´re moving faster than the render process delivers new real and copied frames, you perceive the copied images as a doubled image or ghost image every let´s say 20 milliseconds.

 

So, fist to mention is: Try to keep a stable and average framerate under every circumstance, which should be half of the Hz your VR headset is set up, to maintain a constant and fluid continuity for the interchangeable real rendered frames and copied frames, which in sum should create a smooth perception of the 3D environment in VR while moving.

If you´re dropping under the fixed 45fps things get much more worse to your perception.

 

It´s more delicate when coming to frametime, which is a more real measurement for the performance in VR, as the FPS should be fixed at 45fps in anyway.

BTW: I would expect measuring the FPS with the build in fps counter in DCS, that you only measure with it the frames on your monitor not the frames in the VR headset. You would need a special VR framecounter to get measuring the fps in VR.

 

The framtime is the time needed for the render process to deliver one frame to the headset.

11 milliseconds is the focused value in framtime when processing 90Hz headsets to reach the 90fps target.

 

If the render process could reach 11ms to create one frame continuously at a frequency of 90Hz/FPS then you should´n perceive any judder or ghosting/ double images. Remember the payload to create one frame is much more in a high polygone to redner scene, than over the sea or at 30 000 ft heights.

 

So, while the render process could deliver one frame at a frametime of approx 11ms over the sea, it would need maybe 20ms to create a frame of a city with many polygones, details and eye candy. So while you get the impression of a smooth image over the sea, over a city you will get judder and ghosting as long as you don´t keep the framtime as much close to 11ms as you can get.

 

When the frequency is set to 90Hz ( to reach 90FPS ), but your frametime is over 11ms, one frame stays longer visible as the continuity in frequency has to wait for the next frame longer than 11ms. The higher your fametime is, the longer each frame stays visible to your perception, resulting in ghost images in the worst case. If you can´t stay even at stable 45fps you might get jiddering or stutter effects generally and in addition the perception of double images at fast movements gets much more worse.

 

...oh, have to go back to work - please add or correct if I´m wrongly thinking.

[A Hamburgler]

Who’s ready for the VR improvements today.

[Icebeat]

 

Pitool 1.0 = SteamVR 30% render target: 2560 x 1440 / coefficient 1.77

 

Pitool 1.25 = SteamVR 30% render target: 3200 x 1800 / coefficient 1.77

 

Pitool 1.5 = SteamVR 30% render target: 3840 x 2164 / coefficient 1.77 ( <--- interesting: native resolution of the Pimax 8k panels: 3840 x 2160 )

 

Pitool 1.75 = SteamVR 30% render target: 4484 x 2524 / coefficient 1.77

 

could you please explain why are you using SteamVR 30%?

Thanks

[Rosebud47]

To get the render target resolution as noted.

If you set steamVR not to override the render target, it will always adjust its settings automatically, when you change the render target in Pitool.

[volk_50]

"-2.32 is accurate with render target setting 1.0 and 1.25 in Pitool. To meet the exact aspect ration with a higher render target 1.5, 1.75 or 2.0, there would be need a 3rd or maybe 4th position after the comma."

 

 

I had trouble reproducing this "-2.32" setting at Pitool 1.0 etc. -- SteamVR's scale just wouldn't get exactly to the 2560 scale, always a little off. I tried to to calibrate the value to dial-in but to no avail. It was always a little off and would stop scaling completely after 3-4 precision digit.

[Wicked.-]

@Highwayman-Ed

 

Actually I wouldn´t like to interrupt you posting more compliments, but thank you very much for this. :)

 

-2.32 is accurate with render target setting 1.0 and 1.25 in Pitool. To meet the exact aspect ration with a higher render target 1.5, 1.75 or 2.0, there would be need a 3rd or maybe 4th position after the comma.

 

Some thoughts on judder and ghosting ( double image effects ) in DCS:

 

Special VR games and experiences are build from scratch with minimal load for the render process respectively the GPU to get a smooth experience in VR. Also most VR experiences or games use some kind of teleporting system to move your position in the 3D World and/or snap turning to move around axis. As long as you don´t move or only slowly move in VR you´ll get a smooth experience, if the graphical fidely reflected in detailed graphics like high polygone, high amount of 3D structures and high res. textures does not exceed the capability of the GPU to change the rendered picture in time due to movement.

 

DCS is quite the opposite as it does give you quite accurate the feeling of speed in movement by flying a jet, means the rendered picture needs to change very often when flying low level.

 

-There are more structures to render over a city, than over the sea.

 

-The closer you get to the ground, the more often the rendered picture need to change to maintain the feeling of your speed.

 

-when you fly by another plane both objects move with high speed over a griund which also has to be rendered while moving

 

- most demanding is rolling in your jet while looking to the left or right of the cockpit and not straight, as the rendered picture need to change superfast in every degree of the roll to give you a smooth experience. Some jets can do a 360° roll within approx 1 second.

In this one second of rolling the render process needs to create let´s say for every degree a new image to give you the perseption of a smooth roll. The figures are not accurate and are just an example to visualize the problem.

 

So the payload for the GPU and render process receives a massive impact with superfast movements in DCS VR while having to render high polygone environments.

 

First step for a smooth experience in VR is to keep the maximum framerate of 90fps or 45fps with reprojection/ASW/smart smoothing.

 

With stable 45fps the supporting algorythms generates a copy of rendered frame, which is faster, than the frame would have been actaully rendered by the GPU, and fills with the copied frame the 45 real rendered frames to reach a perceived 90fps. This works fine as long as you don´t move too fast through the 3D environment or so saying the 3D environment does not need to change to fast to give you the impression of a fast movement.

If you´re moving faster than the render process delivers new real and copied frames, you perceive the copied images as a doubled image or ghost image every let´s say 20 milliseconds.

 

So, fist to mention is: Try to keep a stable and average framerate under every circumstance, which should be half of the Hz your VR headset is set up, to maintain a constant and fluid continuity for the interchangeable real rendered frames and copied frames, which in sum should create a smooth perception of the 3D environment in VR while moving.

If you´re dropping under the fixed 45fps things get much more worse to your perception.

 

It´s more delicate when coming to frametime, which is a more real measurement for the performance in VR, as the FPS should be fixed at 45fps in anyway.

BTW: I would expect measuring the FPS with the build in fps counter in DCS, that you only measure with it the frames on your monitor not the frames in the VR headset. You would need a special VR framecounter to get measuring the fps in VR.

 

The framtime is the time needed for the render process to deliver one frame to the headset.

11 milliseconds is the focused value in framtime when processing 90Hz headsets to reach the 90fps target.

 

If the render process could reach 11ms to create one frame continuously at a frequency of 90Hz/FPS then you should´n perceive any judder or ghosting/ double images. Remember the payload to create one frame is much more in a high polygone to redner scene, than over the sea or at 30 000 ft heights.

 

So, while the render process could deliver one frame at a frametime of approx 11ms over the sea, it would need maybe 20ms to create a frame of a city with many polygones, details and eye candy. So while you get the impression of a smooth image over the sea, over a city you will get judder and ghosting as long as you don´t keep the framtime as much close to 11ms as you can get.

 

When the frequency is set to 90Hz ( to reach 90FPS ), but your frametime is over 11ms, one frame stays longer visible as the continuity in frequency has to wait for the next frame longer than 11ms. The higher your fametime is, the longer each frame stays visible to your perception, resulting in ghost images in the worst case. If you can´t stay even at stable 45fps you might get jiddering or stutter effects generally and in addition the perception of double images at fast movements gets much more worse.

 

...oh, have to go back to work - please add or correct if I´m wrongly thinking.

 

 

I'm flying on a brand new system tonight. Specs in my sig. I guess my concern is that while the Rift is a lower res and narrow field of view, even at high settings of 1.8PD/Shadows Low/Terrain shadows flat/No MSAA/No SSAA (I saw zero effect from this) and most everything cranked up except grass, chimney smoke and pedestrian traffic it is silky smooth. With the PIMAX I can never get rid of the judder without making the image look like ass. I'm still trying all your suggestions but I'm assuming Pitools is going to need a few more updates before it gets worked out.

[Rosebud47]

Meanwhile I´m pretty sick of testing, doubting and fiddling with Pimax. I´ll keep it for experimenting, but reserved an Index and will get one as soon as available.

 

After three years of VR, I really appreciate the large FOV before regular HMDs, but am also very interested in 120Hz/ 144hz mode. Let´s see how this goes ... ;-)