Slightly OT, some cool realtime physics are shown here



Anyway I still assume lfs tyre physics will (is?) more realistic

Last edited by just2fast, Tue, 27 Oct 2020 22:26.

Quote from Bose321 :

... spend money to sit with a stupid helmet ...

The funny thing is in most of the games you don´t even sit but stand/move

Quote from cargame.nl :

But to get back on topic, nothing to expect from Microsoft then. Anything else, or is this all?

A small overview of current VR devices (german language!)



and coming AAA games
https://www.vrandfun.com/maste ... ing-aaa-caliber-vr-games/

Quote from Evolution_R :

But Valve just announced a long awaited Half-Life game... and it will be VR only.

A nice interview of Valve eployees why they made this game VR only

Now that the mirrors are adjustable I think they are much superior to the old 2D mirrors.
I like 'em

Nice patch! It´s crazy to see yourself in the mirror when moving the head in VR

However the new 3D mirrors makes it necessary being able to adjust the mirrors. When using a slightly different seat position it´s almost impossible to have a perfect look what happens on the track behind.

An alternative may be to calculate the best mirror setting according to the adjusted seat postition...

Quote from Scawen :

I am working on something.

I have a DK2 and today the drivers updated to 1.6.
No problem so far and lfs starts succesfully.
Are other VR games working?

Quote from just2fast :

One question though Scawen. The IPD slider is a great feature for the consumer version, but how did you solve the issue for yourself that it is not possible anymore to set the IPD in software for DK2?

I see you have a CV1 now?!

After some struggle to get runtime 1.3 running with my old cpu, lfs now runs with my DK2!

(I have some positional tracking issues, but that has nothing to do with lfs but rather with my hardware. I have this with other VR games too and also with the older runtimes)

One question though Scawen. The IPD slider is a great feature for the consumer version, but how did you solve the issue for yourself that it is not possible anymore to set the IPD in software for DK2?

Quote from Flotch :

Hope VR guys are enjoying Scawen's efforts !

Definitely!

I asked the same to the developers, but there seems no possibility with the newer Oculus SDK to turn it off.

I also found out that with the latest runtime 0.8 for some reasons I got judder too.
In the end I went back to runtime 0.7 and are now using the old dll which was compiled with SDK 0.7.
I still have some other isues, but I am not sure if it´s related to my PC or my DK2. However in some month I plan to built a new PC and upgrade to one of the consumer headsets (Vive or Rift).

Last edited by just2fast, Sat, 26 Mar 2016 19:41.

Quote from Degats :

The Logitech profiler has a Shift button function, that allows alternate controls to be assigned to other buttons while it is pressed (like the shift key on a keyboard).

If you assign one of the red buttons on your G25 wheel, then you can increase the number of hands-on accessible functions to 6 (from 4 originally, including the paddles).

For example, if you assign the left red button as Shift, then you could do:
Left/Right Paddle: ShiftDown/Up
Left Red + Left Paddle: VR Keyboard
Left Red + Right Paddle: VR Click
Left Red + Right Red: Spectate
Right Red: ???

Ok thanks, I will look into this!

Solved at least for me. I assigned SHIFT+S to the clutch pedal in the Logitech Software.

However my improvement suggestions for VR are still valid:

Quote from just2fast :

...
a) Wheel button doubleclick -> new assignment possibility
b) Wheel button combinations e.g. Button A + Button B -> new assignement possibility

or c) optional one button solution for "look and click"
...

Quote from Scawen :

That is pretty stupid...

Quote from Flame CZE :

It is stupid...

Quote from cargame.nl :

It is not stupid...

It´s useless to discuss if SHIFT+S is stupid or not, some public server do have this rules for some reasons. Either you accept the rules or you better drive on a different server (if there are any).

Quote from cargame.nl :

...The thing what just2fast is describing is that he has not enough keys on his wheel. Now, is this a problem of the programmer or the person which invests in a wheel. I think I know the answer. There are many Playstation a-like wheels available nowadays with loads of buttons. Go and grab one if you like racing

The wheel in my special case is a G25. Maybe not a highend wheel, but still a fairly good one. And yes a G25 has more than 2 button, but NOT on the wheel directly. They´re on the shifter module instead.
The point is, that in VR you have to find the buttons blindly. So putting commands on the shifter module makes only sense if they´re not time critical. SHIFT+S is time critical.

As cool VR is, it also has drawbacks and interacting with hardware/controllers is definitely one of them. Simply this are new challenges for users and developers.
From a users perspective of course I can buy different hardware, but I also can give feedback what on software side might can be improved. Especially if this affects also other users.

Scawen, is there a way to improve the interaction possibilities with few wheel buttons?

For example my wheel has only 2 buttons directly on the wheel. With the new "look and point" feature I have assigned them now to "virtual keyboard" (left button) and "crosshair click" (right button).
Before that, one of the buttons was assigned to "SHIFT+S" during Logitech profiler.

Today I got kicked from a server, because I wasn´t able anymore to quick exit the race after a crash.

My first ideas are heading in this directions:

a) Wheel button doubleclick -> new assignment possibility
b) Wheel button combinations e.g. Button A + Button B -> new assignement possibility

or c) optional one button solution for "look and click"

Let´s assume wheel button X is assigned to virtual keyboard. Then with the same button X you can interact with the menus. Clicking in the background simply leaves the VR Keyboard mode.

Quote from Lucas McFly :

Hello

when on the virtual keyboard - key is pressed "shift" characters change to "BIG" - and that is okay, but when you press the "CAPS" signs are still "small"

Regards

Same here, however most of the letters are still correctly chosen in CAPS mode.
EXCEPT this ones (german):

I´ve tested the new point and click functionality a little longer today. At first I thought I will assign ESC to the left wheel button and OK to the right wheel button (2 buttons only on the G25), but after a while I came to the conclusion that the virtual keyboard on the left button is the better choice, because there are simply more possibilities than with ESC.

In the end the implementation in this iteration is great as it is. Maybe in future we can interact in this way with the cockpit (ignition, horn, speed limiter etc...) ?

I am more worried about the new view reset mode. I fear it might be a little motion sickness inducing for VR newbies, because the whole world is moving. In the past I closed my eyes, did a reset of the view and opened my eyes again.

What about this idea? During reset keep the world as it is and render a second model of the world in wireframe mode which is perfectly aligned/attached to the HMD. After ok the old world fades out and the wireframe world fades with colors in? But to be honest I have no idea if this could work and is better than the current solution

Last edited by just2fast, Tue, 15 Mar 2016 23:28.

Quote from Scawen :

Test Patch 0.6M3 is now available.

For VR users it has a look and click feature and a virtual keyboard. You can assign Click, Escape and Virtual Keyboard to controller buttons so you can operate LFS without keyboard or mouse.

The virtual keyboard is also available for non-VR users, so you can control LFS without the keyboard. You will still need a mouse.

https://www.lfs.net/forum/thread/89160

I had only a few minutes to test the new "look and click" feature, but WOW now lfs really is 100% VR ready!

I will continue testing tomorrow.

One first side note maybe. I think the feeling for the crosshair would be better if it wouldn´t get stuck at the edge of the virtual screen. I would still render the crosshair outside the screen, but instead of rendering to infinity when lookin at 90° I would slowly fade out the crosshair after leaving the screen area. So let´s say at around 30%-50% (measured in % of screenwidth) outside of the screen the crosshair finally is invisible. Maybe it´s hard to implement and you have better things to do at the moment .

Quote from cargame.nl :

Also wondering if central position is most ideal position. But others need to say something about it I guess.

With an VR device I find it useful to have it in the middle of the screen, because you click "with the nose".

Last edited by just2fast, Mon, 14 Mar 2016 23:31. Reason : typo

Quote from Scawen :

One requirement is that it must be possible to operate the game with a controller only. For the last two days I have worked on a system that draws a crosshair in the view centre and allows you to click buttons using the space bar. Obviously this "Click" function needs to be assignable to a controller button. At first I thought this task would be a few hours but the complexity is quite surprising. Various adaptations and improvements have become necessary. I think the only other buttons that are needed are "Escape" and "Show Virtual Keyboard". These will soon be available in a compatible test patch.

Thank you!

Great efforts in this thread!

However as I state before, I think Googles Deepmind AI has the potential to be the perfect AI driver.


See this:
http://futurism.com/major-brea ... european-go-champion-5-0/

http://futurism.com/googles-ai ... t-ranked-go-player-world/

Even playing 3D games is already possible:
http://futurism.com/google-dee ... -doom-like-3d-maze-sight/

This includes racing:
http://futurism.com/videos/goo ... as-now-moved-to-3d-games/

Quote from Scawen :

Thank you for the test. Sorry to hear it doesn't help. I am using the 358.91 driver.

No problem as long runtime 0.7 works fine.

BTW with other games/demos I have the same performance issues under runtime 0.8 (358.91 didn´t help either), so at least it´s not lfs related!

Quote from Scawen :

I would be interested to know if you get better results with the LFSRiftVR.dll in the attached zip file.

This one is built with the 0.8 SDK and so requires the 0.8 runtime.

You just copy it into your dll folder. Better rename the existing one so you can go back if needed!

EDIT: I've uploaded two versions of the Rift DLL, both built with the 0.8 SDK. There is a slight difference between then but I can't perceive any difference and in fact they may perform just as well as each other.

Ok, I have installed runtime 0.8 again for testing, but same behavior as before. Both 0.8 dlls perform poorly. Between 35 to 75 fps with an avarage of around 50 fps, depending what has to be rendered.
In the starting grid for example with only 5 cars it drops down to 30 something.

Maybe it has to do with the NVIDIA driver in conjunction with 0.8?

It´s the latest driver version 361.43. Graphics card is a GeForce GTX 750 Ti.

Quote from Scawen :

Also here is an OpenVR DLL built with the latest OpenVR SDK and I was surprised to find out that it also works with the Rift (with 0.8 installed). I found this sometimes started up wrong though and had to exit and restart to get it right. I would be interested to know if this works without SteamVR needing to be installed (by selecting OpenVR instead of Oculus Rift). I have no idea.

Nope, OpenVR can´t be initialized in lfs (note: SteamVR is NOT installed).

Quote from Scawen :

No, it's not possible to disable vertical sync in Direct mode. I find LFS performs better in the Vive than the Rift even with the Vive's higher frame rate. One thing is that LFS can present its render target directly to the Vive software which renders the final image. In the Rift, you are forced to request a render target from the Rift software and that is a DX11 render target which cannot be used by LFS for rendering. So LFS must do an extra, wasted copy from its DX9 target to the DX11 one.

I don't know for sure if that is the real cause. I have a note to implement the 0.8 SDK support which is hopefully mainly renaming a few functions. Then we'll see if the 0.8 SDK works better than the 0.7 one.

It seems it had to do with the runtime 0.8. After going back to runtime 0.7 everything is fine and I can get some constant 75 fps.

Scawen, is it possible to disable vertical synchronisation in Rift direct mode? I bought a new dx11 graphics card, but can´t get a stable 75 fps in 0.6K. Between 50-75 fps only.
With my old GT 9800 and 0.6H (vsync off and extended mode) I got a stable 100 fps...

Quote from cargame.nl :

This is just before the peak now.Then the phase of 'trough of disillusionment' starts..

Well...

according to Gartner, VR is not before the peak, instead it left the "trough of disillusionment" in 2015

Last edited by just2fast, Fri, 8 Jan 2016 20:09. Reason : typo