VR Optimization

[Germane]

Hey guys, when the Viggen or F14 was developed VR was not so a big deal and you notice that no great focus was put on it. But more and more players use VR and I hope you have a special focus on it to optimize the Eurofighter for the VR experience.

[Lurker]

I think that this is a reasonable request. However considering how much detail and effort Heatblur pours into all the 3d modeling and texturing of all their modules, it might be a bit too much to ask of them. If it's not too much additional work, it might set a nice precedence and would improve the exeprience of even to 2d users, with minumum spec PCs and graphics cards. 

[Steph21]

Gero is using VR and already mentioned they will look to make it a good experience in VR

[Germane]

Am 25.11.2021 um 14:34 schrieb Steph21:

Gero is using VR and already mentioned they will look to make it a good experience in VR

Ah I didnt know that. Sounds good!

[westr]

I hope so F14 cockpit isn’t very taxing for me in VR. 

[IronMike]

We, uh, put a lot into VR optimization for both Viggen and Tomcat. But a trade is a trade: you can have it look great, and require a strong setup. Or you can have it look not so great and require a not so strong setup. We will put the same VR optimization that we put into the Tomcat and Viggen. I don't think some realize how much more detail is in our modules, and thus how much it actually is optimized for VR (or else).

[Germane]

vor 9 Stunden schrieb IronMike:

We, uh, put a lot into VR optimization for both Viggen and Tomcat. But a trade is a trade: you can have it look great, and require a strong setup. Or you can have it look not so great and require a not so strong setup. We will put the same VR optimization that we put into the Tomcat and Viggen. I don't think some realize how much more detail is in our modules, and thus how much it actually is optimized for VR (or else).

With VR optimization I dont mean performance optimization. I mean the optimization of the VR "experience". Things like the correct positon of the head and and a good graphical representation in vr. The planes in dcs look very different in vr. Some look very good and others much less so. Some effects / textures simply look different in 2D than in VR. The Tomcat, for example, is beyond reproach in 2D. in Vr, however, a few things stand out negatively.

[draconus]

2 hours ago, Germane said:

Things like the correct positon of the head and and a good graphical representation in vr. The planes in dcs look very different in vr. Some look very good and others much less so. Some effects / textures simply look different in 2D than in VR. The Tomcat, for example, is beyond reproach in 2D. in Vr, however, a few things stand out negatively.

Can you be more specific? What is wrong in the VR Tomcat cockpit or how the head is in wrong position?

[Lurker]

2 hours ago, Germane said:

With VR optimization I dont mean performance optimization. I mean the optimization of the VR "experience". Things like the correct positon of the head and and a good graphical representation in vr. The planes in dcs look very different in vr. Some look very good and others much less so. Some effects / textures simply look different in 2D than in VR. The Tomcat, for example, is beyond reproach in 2D. in Vr, however, a few things stand out negatively.

To me this is nitpicking. The only thing that can be a little bit off is the scale of the VR cockpit, which can be easily changed within the DCS graphical options. 

[IronMike]

On 12/15/2021 at 2:32 PM, Lurker said:

To me this is nitpicking. The only thing that can be a little bit off is the scale of the VR cockpit, which can be easily changed within the DCS graphical options. 

The scale and proportions are down to, iirc 1mm or less.

From what I gathered so far, one of the complaints we get the most about VR is readability of certain stencils, texts, etc or gauges. But mind you: a navy jet is much more prone to weathering than a non-navy aircraft, for one, thus cockpits look less worn, and secondly, VR also caps at some point still, where making stuff more readable means also distorting it beyond realism. In time users should be acquainted with the cockpit enough, to just now which MASTER test position means what, etc, too.

[Dragon1-1]

I think that the readability might be a problem with how DCS does mipmapping. In another sim, I noticed that numbers on a gauge that's far away from the viewpoint are much less readable than on one that's close by, despite the apparent size being very close. DCS doesn't have any aircraft with gauges out on the wing, but I'm pretty sure something similar is happening here. A mipmapping algorithm optimized for pancake is not going to work the best in VR.

[Lurker]

14 hours ago, IronMike said:

The scale and proportions are down to, iirc 1mm or less.

 

I was referring to the setting that Eagle Dynamics calls IPD scale. Depending on the values set there cockpits can look bigger or smaller in VR and this has nothing to do with Heatblur, but sometimes the correct value that looks "good" in one jet, is completely off in another and a lot of people are unfamiliar with this setting in DCS World

[Cobra847]

6 hours ago, Lurker said:

I was referring to the setting that Eagle Dynamics calls IPD scale. Depending on the values set there cockpits can look bigger or smaller in VR and this has nothing to do with Heatblur, but sometimes the correct value that looks "good" in one jet, is completely off in another and a lot of people are unfamiliar with this setting in DCS World

Generally the IPD will scale things up or down because it simulates the spacing between your eyes. If one cockpit looks too big or too small with the same IPD as your natural pupillary distance, it means that one or more cockpits are mismodeled; or you think something in reality is bigger/smaller than it actually is (this is frequent )

 

The Eurofighter is a newer jet; as such it will have a different character than the Viggen or the F-14 and thus may be naturally easier to use in VR; however it will be chock full of the requisite Heatblur texturing quality and mesh accuracy.

Edited December 17, 2021 by Cobra847

[Jayhawk1971]

I believe to have read somewhere that the "IPD" setting under the VR section in DCS is actually a misnomer, and it actually means something like "world scale". 

Since I have no real reference as to what the actual scale of a cockpit is, I just left the whole thing off for the Tomcat, which does feel to be right on the money regarding proper dimensions (having never sat in a Tomcat). However, subjectively the A-10 seems to be too spacious, whereas the Viper feels a little too cramped.

Edit: I did have the opportunity to sit in the front seat of a CF-18B, but that was a loooong time ago, but I did seem to remember the cockpit to feel smaller in the real thing compared to the DCS Hornet.

My Vive Pro has a dial on the bottom of the headset where I can manually set the IPD, which I have measured using a mirror, a measuring tape, liberal amounts of Whisky and a Voodoo Priestess. 

Edited December 17, 2021 by Jayhawk1971

[Raven (Elysian Angel)]

On 12/15/2021 at 3:28 PM, draconus said:

or how the head is in wrong position?

HeatBlur themselves have admitted the default VR head position is lower than the pilot's head would be IRL. My guess would be is that it's some sort of compromise to make certain things easier and/or more easily visible such as IFR. 

Personally, I moved the VR head position up and then resaved that as default, but I have to lean forward quite a bit while refuelling if I want to keep my eye on the tanker properly.

[draconus]

22 minutes ago, Raven (Elysian Angel) said:

HeatBlur themselves have admitted the default VR head position is lower than the pilot's head would be IRL. My guess would be is that it's some sort of compromise to make certain things easier and/or more easily visible such as IFR. 

I see it other way. After recentering I can barely see the heading tape on the HUD under the rail. Since we have the seat height regulation I don't see a problem. The head is definitely in the right place.

[IronMike]

On 12/17/2021 at 7:47 PM, draconus said:

I see it other way. After recentering I can barely see the heading tape on the HUD under the rail. Since we have the seat height regulation I don't see a problem. The head is definitely in the right place.

Raven is right here, we moved the head down on purpose. IRL the pilot literally had the canopy bar in his face, however IRL it is much easier to squeeze your body down in the chair lean left right (not just with your viewpoint, but the pivoting point of your body), to lower a shoulder, to glance from the corner of your eyes, etc. If we would put the head to the correct height, it would be really disturbing in the sim, to the point where folks would think it is wrong, just like you just thought as is, is correct, while it is not quite.

Edited December 18, 2021 by IronMike

[Jayhawk1971]

vor 15 Stunden schrieb IronMike:

IRL the pilot literally had the canopy bar in his face,

...whereas "Maverick" could barely look over the dashboard.    (or did somebody find him a box?).

So for VR users, if they want the most authentic experience, would you recommend to raise the head position back to the height of the canopy bar?

[draconus]

7 minutes ago, Jayhawk1971 said:

So for VR users, if they want the most authentic experience, would you recommend to raise the head position back to the height of the canopy bar?

You'd seat correctly but then the HUD will not be adjusted to that position.

[Jayhawk1971]

Am 19.12.2021 um 14:18 schrieb draconus:

You'd seat correctly but then the HUD will not be adjusted to that position.

Adjusted as in obstructed by frame/handle, or adjusted in the sense that the (simulated) HUD-projection is off?

Guess I'll have to play around with that next time I fire up DCS. 

Also, since there are basically two ways to adjust the seating position (the numpad keys, or moving around and pressing the recenter key in VR, which allows one to, for example, "stand" in the cockpit): I wonder if the effect is the same with either method. 

[draconus]

2 hours ago, Jayhawk1971 said:

Adjusted as in obstructed by frame/handle, or adjusted in the sense that the (simulated) HUD-projection is off?

Guess I'll have to play around with that next time I fire up DCS. 

Also, since there are basically two ways to adjust the seating position (the numpad keys, or moving around and pressing the recenter key in VR, which allows one to, for example, "stand" in the cockpit): I wonder if the effect is the same with either method. 

Yes, same effect with any method - the HUD will be either obstructed or you will not see all of it but ex. crosshairs will still point correctly.

[J-1775]

On 12/17/2021 at 1:35 AM, Dragon1-1 said:

I think that the readability might be a problem with how DCS does mipmapping.

Hi Dragon. You seem to know something about that! I wonder how MIP loading is done in DCS. Is ist done all automatically (if a DDS includes mipmaps) or can I control it somewhere for my own EDM models? Especially I wonder if there is any correlation between shape LODs and mipmaps?

[Dragon1-1]

I don't know how it is for DCS specifically, but usually, mipmaps are within the DDS file. LODs are a different thing, and would typically use a different texture altogether, though again, I have no idea how DCS in particular does this.

[J-1775]

Thanks. That was actually the starting point for my question: Is it better to save my texture files with MIPs and let the system do all the deciding when to swap MIP subfiles or shall I save extra smaller files for each of my LODs and associate them to the appropriate LOD. It would appear to me that the first way should be the optimized way, even quality-wise? It'll also be much less complicated than associating  different file names to each LOD. What ya think?
Since I use NVIDIA's DDS texture exporter tool I was just wondering who exactly decides when the subfile swap happens. Is it the game (DCS), is it DX11/Vulcan or maybe even NVIDIA's driver?

Edited February 19, 2022 by J-1775

[TOViper]

On 12/15/2021 at 3:30 AM, IronMike said:

We, uh, put a lot into VR optimization for both Viggen and Tomcat. But a trade is a trade: you can have it look great, and require a strong setup. Or you can have it look not so great and require a not so strong setup. We will put the same VR optimization that we put into the Tomcat and Viggen. I don't think some realize how much more detail is in our modules, and thus how much it actually is optimized for VR (or else).

Hello Mike!
Seeing the pictures of the cockpit makes me smile and fear.
I am long time user of VR, playing a sim on a screen became a no-go (i would then rather go playing guitar or biking instead of playing on a screen, honestly).
The newest AH-64D e.g. consists of MANNNY triangles to be calculated, and in VR this number doubles, and thus it currently kills VR user's fps with "low" to "medium" setups (like mine).
Which optimizations are you applying to the EF module? LOD? Textures? Number of Triangles? Other secret tricks?
No hurry for answer, just for interest.

Kind regards
TOViper

P.S.: Please be so kind and look to my P.M. (if you haven't seen it so far)