New VR Features.

[montify]

Hello, i saw this Video today:

 

 

 

Can you tell me what is Pixel Density means? And is this a good improvement?

 

And when we get this VR update? :)

Edited June 5, 2016 by SkateZilla
Fixed YT Embedding

[SkateZilla]

Hello, i saw this Video today:

 

 

 

Can you tell me what is Pixel Density means? And is this a good improvement?

 

And when we get this VR update? :)

 

 

It's Basically Virtual Super Resolution.

 

Instead of Rendering 2160x1200, You're rendering an image larger than the display and scaling it to the display

[montify]

THx for this info, with this i can read the labels better, i mean this is possible to render more pixel that the screen have?

[sniperwolfpk5]

Yes it will improve readability

[Fab]

This is gonna be a game changer for us VR players !!!

 

I know "2 more weeks" and I know "It will be there when its done" but is there an ETA on Pixel Density and Mouse control ?

[npole]

It's called super-sampling.. and you can do it already it now by enabling it on your gfx control panel... and it is a (expensive) antialiasing method.

The issue is: we are already doing bad fps with hi-end cards, increasing the resolution will make it even worse, and you know that everything lower than 90fps (or 45 fps with retro projection) is an issue with VR (ghosting).

 

I see people super excited about this, and I don't understand the reason: you can try it already today... and you will be disappointed by the frame rate, especially in multiplayer.

 

I think the priority should be given to optimize the performance, and they should look seriously to the nVidia serie 10 API and their new view port rendering, because it's gonna give a huge boost (almost doubling the fps), reducing the workload of the stereo rendering they are using now... it's gonna work only with the 1070 and 1080, but at the moment is the only solution we have to have some rock solid fps.

[npole]

THx for this info, with this i can read the labels better, i mean this is possible to render more pixel that the screen have?

 

No, unless you manage to enter in a parallel universe.. :)

In two words: with the super-sampling, the pixel color transition is better optimized, because the antialiasing is calculated using a much larger image than the original one (in example 2X) and then scretched to the original resolution.

It gives a better result.. but it's extremely resource expensive, for obvious reason: a 2X image (in width and height) is 4 times bigger than the original picture.

[RaidPW]

I'd love to know the specs of the machine he's running that on. If they're the ones he mentions in his forum sig ("i7 4770K, 8 GB RAM, GTX 980ti, Windows 10, 1TB SSD, TM Warthog stick with CH Pro Throttle"), then he's getting vastly better performance than I am despite having only a marginally better CPU, a roughly equal GPU and less RAM. I really hope that's down to performance improvements in the graphics engine.

[SkateZilla]

It's called super-sampling.. and you can do it already it now by enabling it on your gfx control panel... and it is a (expensive) antialiasing method.

The issue is: we are already doing bad fps with hi-end cards, increasing the resolution will make it even worse, and you know that everything lower than 90fps (or 45 fps with retro projection) is an issue with VR (ghosting).

 

I see people super excited about this, and I don't understand the reason: you can try it already today... and you will be disappointed by the frame rate, especially in multiplayer.

 

I think the priority should be given to optimize the performance, and they should look seriously to the nVidia serie 10 API and their new view port rendering, because it's gonna give a huge boost (almost doubling the fps), reducing the workload of the stereo rendering they are using now... it's gonna work only with the 1070 and 1080, but at the moment is the only solution we have to have some rock solid fps.

 

Super Sample Anti-Aliasing (SSAA) on nVidia or AMDs Control Panel is NOT the same as Pixel Density on the VR API.

 

 

As for NVidia's 1080 Viewport/VR Features, those are Part of nVidia's GameWorks API, which is not used by ED.

Edited June 5, 2016 by SkateZilla

[npole]

Super Sample Anti-Aliasing (SSAA) on nVidia or AMDs Control Panel is NOT the same as Pixel Density on the VR API.

 

What's the difference?

 

those are Part of nVidia's GameWorks API, which is not used by ED.

 

I know they are not used, and considering the enormous benefit it gives.. they should really consider to implement it (there's no downside doing so).

[SkateZilla]

What's the difference?

 

 

 

I know they are not used, and considering the enormous benefit it gives.. they should really consider to implement it (there's no downside doing so).

 

 

Theres a difference, I'd post a detailed write up, but its pointless.

Mainly, if an Engine isnt designed to support SSAA Specifically, it will have a Negative impact on performance,

Where as PD/VSR/DSR is simply rendering the game at a larger resolution and using the VPU's Scalar to downsize and output at the screen's native resolution.

 

Why pay nVidia for a Gameworks Developer License and re code the edge engine so only 2 GPUs can use said features, while at the same time nerfing the performance for 900 series and below as well as AMD.

Edited June 5, 2016 by SkateZilla

[npole]

Theres a difference, I'd post a detailed write up, but its pointless.

 

Pointless?

 

Why pay nVidia for a Gameworks Developer License and re code the edge engine so only 2 GPUs can use said features, while at the same time nerfing the performance for 900 series and below as well as AMD.

 

You don't pay to use it (and it is VRWorks in this specific case). And you are not "nerfing" the other GPU: you are giving an advantage (actually an huge advantage) to those who (and will) own the new GPU's, while the others will continue to perform the same. It's called progress. I would invite you to read about the single pass rendering, to figure out what's the incredible benefit we would have.

 

I would remind that those cards will be owned by more than 7 players of 10 (75%) from now to the next year...it's not just "two cards".

Edited June 5, 2016 by npole

[adrianstealth]

Will this just be for version 2 .

 

I have 2 but use 1.5 because I prefer the 1.5 map unless everything will become available in 2

[SkateZilla]

Pointless?

 

 

 

You don't pay to use it (and it is VRWorks in this specific case). And you are not "nerfing" the other GPU: you are giving an advantage (actually an huge advantage) to those who (and will) own the new GPU's, while the others will continue to perform the same. It's called progress. I would invite you to read about the single pass rendering, to figure out what's the incredible benefit we would have.

 

I would remind that those cards will be owned by more than 7 players of 10 (75%) from now to the next year...it's not just "two cards".

 

VR Works is part of Game Works,

 

And Apparently you havent seen what happens to games that run GameWorks API Code on GPUs that dont support it.

[npole]

VR Works is part of Game Works,

 

As I specified: it's VRWorks in this specific case, it's a set of completely new API dedicated to the VR frameworks.

 

And Apparently you havent seen what happens to games that run GameWorks API Code on GPUs that dont support it.

 

The graphic cards which will not support this specific feature will work normally, the ones which will support it, will take advantage of it. It's a feature given to those who can benefit of it, assuming that the developer won't be lazy to not optimize the code for anyone else as well.

If you know something more about VRWorks (and specifically about the single pass rendering and the new view ports, please post it here.. otherwise it's speculation).

Both Unity and the Unreal Engine have introduced the support to VRWorks, do not take advantage of it, it is a wrong move: NVIDIA cards are used by th 75% of the players... it's not two ppl.

[SkateZilla]

Trust me, Game Works (and the VRWorks Components), is not the Magic Pill they make it out to be.

 

This is no different than asking ED to integrate WaterWorks for better looking oceans..

[Phantom88]

I'm so very Happy DCS is taking such an active step forward into The VR Market.

[DerekSpeare]

I'm so very Happy DCS is taking such an active step forward into The VR Market.

 

Me too! It's evident that ED/Wags and company understand the urgency and necessity for proper VR support.

 

@adrian - from what I know the future "V2.5" will harmonize the maps in V1.5 and V2 into one package. I am guessing that they will then include the Hormuz map as well.

[npole]

Trust me, Game Works (and the VRWorks Components), is not the Magic Pill they make it out to be.

 

I don't have to trust you, I have looked at the API: the view port rendering delegation to the GPU saves 80% of the rendering calls if compared with a standard stereo rendering.

I think your assumption are based on speculations (because of what happened in the past: physx, HairWorks), not about the real facts: I'm suggesting to apply a technology that would give a huge benefits to the nvidia users, that are actually the 75% of the gamers and where the percentage is even higher if we talk about the simmers. The 1070 and the 1080 (and the possible future 1080Ti), will be the new 970, 980 and 980Ti, you can't ignore a technology because of speculation, you have to actually looking at the tech. I'm not asking you to "like" VRWorks, I'm asking the DCS coders to look at the API if they have the occasion to do so... like others have done already (it's not that Unity and Epic are the last comers...).

 

This is no different than asking ED to integrate WaterWorks for better looking oceans..

 

This is not about the look, this is about having the VR working properly: having a solid frame rate is essential to make it to work, and DCS is struggling on this front, you need to take advantage of the new tech.. you can't say: "we don't need it...", without really understand what's about.

[SkateZilla]

I can start posting marketshare %'s to.

 

A Neutral developer is not going to adopt an API that will cost them money to develop and integrate and only give a performance increase to less than 15% of the total GPU market, this includes Entry Level, Professional Market, Etc etc.

 

Im telling you the Functions of VR Works (ie Multi-Projection), are available without using their API, and will work on both AMD and nVidia, as well as any other DX11 GPU.

 

Oculus can easily integrate VRWorks, LiquidVR and GPU Open VR Functions into their API, which will allow any Application using their runtimes to render directly to their HMD to use which ever the GPU Supports.

 

That being Said, VR Implementation is not the bottleneck, as the bottleneck exists for users on normal Screens as well.

I can Run 1 Viewport, or 3 Viewports w/ Dx Offset and lose next to no performance, and maybe 15-20% performance going from normal to Stereoscopic 3D.

 

VRWorks will only be used by Companies that Develop Mass Use and Mass Licensed Engines (Unity Engine, Unreal Engine, Etc Etc).

I Doubt you see it integrated into Independent Company's Engines.

Edited June 6, 2016 by SkateZilla

[X93355]

Anyone else see the FA18C icon between the F86 and FC3 on Matts screen? Or is it me being wishfull?

[Milopapa]

And when we get this VR update? :)

 

Looking at these two posts from Wags, I'd say there is a good chance it will come this Friday.

 

We plan to update 2.0 next Friday. I have asked the team to make sure the AI fix is included.

 

This week we internally added the option to control the cursor with either the mouse or head movement.

 

Hopefully, this and other VR improvements like pixel density selection will be rolled to a release version soon.

 

Thanks

[Fab]

I don't have to trust you, I have looked at the API: the view port rendering delegation to the GPU saves 80% of the rendering calls if compared with a standard stereo rendering.

I think your assumption are based on speculations (because of what happened in the past: physx, HairWorks), not about the real facts: I'm suggesting to apply a technology that would give a huge benefits to the nvidia users, that are actually the 75% of the gamers and where the percentage is even higher if we talk about the simmers. The 1070 and the 1080 (and the possible future 1080Ti), will be the new 970, 980 and 980Ti, you can't ignore a technology because of speculation, you have to actually looking at the tech. I'm not asking you to "like" VRWorks, I'm asking the DCS coders to look at the API if they have the occasion to do so... like others have done already (it's not that Unity and Epic are the last comers...).

 

 

 

 

 

This is not about the look, this is about having the VR working properly: having a solid frame rate is essential to make it to work, and DCS is struggling on this front, you need to take advantage of the new tech.. you can't say: "we don't need it...", without really understand what's about.

 

Nvidia is used by more than 70% of the simmers ? Now that sounds like a assumption based on speculation, of a fanboy

[npole]

Nvidia is used by more than 70% of the simmers ? Now that sounds like a assumption based on speculation, of a fanboy

 

nVidia at the moment has a marked share of 71% overall (Q1 2016) for the desktop market.

The above percentage is higher for the mid-end/hi-end solution (for the previous gen: 970/980/980Ti-TitanX; to be substituted by 1070/1080/1080Ti).

The simmers (FSX/X-Plane/DCS/...), because of the hardware demand, are normally concentrate in the upper market, where nVidia is stronger (you can make a poll yourself to figure it out).

That's the reason of why the overwhelming majority of simmers own a nVidia.

"fanboy"? (nor I have the age anymore to make console wars, gfx wars, CPU wars.. and this childish things anymore.. :) ), I adopt both the solutions, where it is convenient to do so.

 

A Neutral developer is not going to adopt an API that will cost them money to develop and integrate and only give a performance increase to less than 15% of the total GPU market, this includes Entry Level, Professional Market, Etc etc.

 

Implementing the new viewports is a two week man job (for a game that already supports the multi monitor), and projected to the future it will give an advantage to your customers at "zero cost", it would be naive to not look at it.

 

Im telling you the Functions of VR Works (ie Multi-Projection), are available without using their API, and will work on both AMD and nVidia, as well as any other DX11 GPU.

 

Oculus can easily integrate VRWorks, LiquidVR and GPU Open VR Functions into their API, which will allow any Application using their runtimes to render directly to their HMD to use which ever the GPU Supports.

 

If you look at how it work: the single pass rendering is hardware driven, it cannot be done externally (not until the DX will permit it), that's why it give a such boost in performance, it's because it splits the "viewports" before the DX call.

You can say AMD could do the same (probably they will do...), and in fact you could implement their solution as well as soon it's ready.

None have asked to limit the others, I have suggested to look at the new tech because the boost in performance is huge.

 

That being Said, VR Implementation is not the bottleneck, as the bottleneck exists for users on normal Screens as well.

I can Run 1 Viewport, or 3 Viewports w/ Dx Offset and lose next to no performance, and maybe 15-20% performance going from normal to Stereoscopic 3D.

 

Indeed, we have a performance issue.. and it's especially evident while in multiplayer, and this issues are amplified when in VR because the fps is a crucial factor. While on a normal monitor you will notice the low fps but you can continue to play; in VR having a frame drop will make the game unplayable, so why not looking at the new solutions?

 

VRWorks will only be used by Companies that Develop Mass Use and Mass Licensed Engines (Unity Engine, Unreal Engine, Etc Etc).

I Doubt you see it integrated into Independent Company's Engines.

 

It's a "free" solution, so again there's really no practical reasons to not adopt it. I know Project Cars is currently looking at it (from their forums), and they uses a independent engine.

 

I honestly struggle to understand why someone is so reluctant to embrace a new technology, if there's some technical limitation of the current engine, it can be understandable.. but to ignore something that could help you and your customers to have a better performance.. hmm... dunno it sounds wrong. I'm not asking you to sell your soul to the devil.. I'm asking the devs if there's a chance to look at it.

[Milopapa]

Honestly, instead of VRWorks, what I'd love DCS to support is DX12 or Vulkan. Most of the performance issues come from DirectX Draw Call overhead, that's why we see both CPU and GPS sit unused at 40% while struggling with framerates.