wju
-
Posts
151 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Posts posted by wju
-
-
not sure if it is intention, but as said in title: modifier defined in module is ignored by UI layer until is defined also within UI
an example (see screenshots below)
in module:
custom modifier: VKB_BTN42
buttons binded:
VKB_BTN42+JOY_BTN21
VKB_BTN42+JOY_BTN23
VKB_BTN42+JOY_BTN24
VKB_BTN42+JOY_BTN22in ui layer:
no custom modifier
buttons binded:
JOY_BTN21
JOY_BTN23
JOY_BTN24
then working butons are
JOY_BTN21 - from UI
JOY_BTN23 - from UI
JOY_BTN24 - from UI
VKB_BTN42+JOY_BTN22 - from modulenot working:
VKB_BTN42+JOY_BTN21
VKB_BTN42+JOY_BTN23
VKB_BTN42+JOY_BTN24these last three combos starts to work only when modifier VKB_BTN42 is defined in UI-layer, even though it is not used in UI.
Is it intention?I use only two modifier DCS-wide, i.e. it is no problem to add it to UI, but some pilots may need more...
thank you for your attention
-
-
VR is hardware, truth; regardless most of VR´s problems come from its SW implementation.
On the other hand, VR capability is DCS´s the core feature for all VR pilots.
-
1
-
-
significant CPU boost in my favourite CPU_Test mission, i.e. FW190A, Normandy Quick Action, Airfield attack - just before attack
well done!
-
2
-
-
23 hours ago, Supmua said:
I'm very skeptical about significant *GPU* performance jump by just upgrading CPU, it's more likely to be related to the final fps but this has to do with both GPU and CPU frametime numbers. I'm upgrading my CPU from 9900K to 5800X3D soon and I would be shocked if there is that much improvement in GPU performance. Already have all the 9900K/3090 and 9900K/4090 frametime numbers from various missions waiting to crosscheck.
please do not forget to share your frametime numbers after upgrade!
-
3
-
-
On 10/12/2022 at 2:09 AM, M1Combat said:
Yeah I don't think you'd see as much improvement as I did anyhow. You seem to have 6GB vid mem... Is that a 970 or something of that era?
No, I have 3080Ti &12GB VRAM
OpenXR overlay shows "used VRAM", whatever is means..
-
12 hours ago, M1Combat said:
Well the actual framerate numbers aren't really what I was trying to point out... It's more about the Delta FPS if you will. How much was gained in a few limited scenarios. As you can see...
thanks, I just wanted to know, how bad is my "stone age CPU" (intel i7 7820X 4.7GHz & 32GB RAM)
Since this spring, I am hesitating whether wait another "two weeks" for multicore or go to CPU upgrade.
the screenshot is with the same DCS settings as yours, OXR 100%, toolkit - all improvements off; I am pretty CPU bound here
-
On 10/9/2022 at 1:27 AM, M1Combat said:
Tower View... 28FPS and some telemetry on the left. 3800X@4.3Ghz 12GB3080 64GBMem@3600
Tower View... 38FPS and some telemetry on the left. 5800x3d@4.45Ghz 12GB3080 64GBMem@3600
Ocean View... 75FPS and some telemetry on the left. 3800X@4.3Ghz 12GB3080 64GBMem@3600
Ocean View... 79FPS and some telemetry on the left. 5800x3d@4.35Ghz 12GB3080 64GBMem@3600
would You be so kind and share Your settings? (i.e DCS, OXR..)
-
1
-
-
-
On 9/14/2022 at 3:14 AM, m0jo said:
Hi, here is my take on it.
Instead of Bluetooth, this is using Nordic Gazell protocol with a small dedicated receiver running QMK. Only the receiver need to be flashed (via USB) if some key mapping need to be changed. This support keyboard buttons, mouse click, mouse wheel (with acceleration) and joystick buttons press.
The CR2032 hidden on the top of the finger should last a few years since the microcontroller is always sleeping and don't need to stay in sync with the receiver. There is no delay after pressing a key, this is very similar to pre-Bluetooth wireless keyboard/mouse.
I also added some extra wireless buttons on the VR headset with the same design (but different PCB) for things like centering the view and catapult salute
would you be so kind and share more details? I am very curious...
-
well, my espruino based "very ugly proof of concept" version is working! (left finger only so far, second espruino is on its way)
the only issue was caused by my own stupidity; I have forgotten to plug bluetooth antena to motherboard and therefore spent two hours by thinking experimenting why espruino cannot be connected...
As you can see, I have decided to build "ring" version with lipo battery (230mAh), as I hate gloves, battery can be smaller in the future, but this one was on my shelf
the latching button swtitches on/off, where off can also be a charging mode; charging pins are not soldered yet.
the only weird thing is when mouse becomes "white dot"; then it behaves oddly, i.e. moves with headset in oposit direction, (I understand why), but one can get use to it...
anyway, guys @frenzon, @Sielu thanks for sharing!
-
4
-
-
8 hours ago, Sielu said:
Nope, this option in WMR settings:
ah, these I have also off
-
8 hours ago, Sielu said:
I've also had an issue with the latest version of Gemini and second monitors messing fingers up. WMR Was creating 'extra' fake monitors to handle 'legacy' windows apps every time it started, which was really chewing up Fingers. I could get it to work by re-starting fingers AFTER WMR, but disabling the extra monitors setting in WMR fixed it for good.
meaning the "HKEY_CURRENT_USER\SOFTWARE\Microsoft\Windows\CurrentVersion\Holographic" trick or something else?
wired mouse version based on Arduino works fine, but the wire is annoying, maybe I am "the most ugly solution award" winner
anyway, espruinos ordered.
-
1
-
-
well, I have found the root of my problem - the second monitor. When switched off, then FingersApp works as expected. When is Fingers running, it even does not mind when I switch second monitor back on. Mouse cursor behaves as expected.
Now its the time to build a mouse (wired version at first ) based on an Arduino.
-
1
-
-
Hi,
@frenzonnice work with Fingers App! Thanks for Your effort!
I took my Leap out of the shelf, where it had been rotting for years, printed the mount for my Reverb G2, installed latest Gemini, installed your Fingers 0.9.3, enabled BlueTooth (I do not use it, it was also rotting since ordovik on my PC) and viola! It works!
Unfortunately with ugly offset to right, i.e. mouse pointer is shifted 1/4 screen size to right and therefore it is very difficult to use mouse pointer in DCS as it is very close to right hand edge of visible area, see the screenshot,
Mouse pointer is cca at red point, as PrintScreen does not include it. This right offset is present for both hands.
Is there any setting or calibration of Fingers app?
I have found user.config in
\\AppData\Local\FingersApp\FingersApp.exe_Url_oiz1xmo2tljpyqkdzivdgwc423rzymoi\1.0.0.0\but there is nothing to meaningfully setup.
-
downloading already...
-
2
-
-
6 minutes ago, edmuss said:
A10C easy instant action, it's my typical go-to to roughly guage performance as it's a pretty set mission and frametime variations are minimal to none (barring MFDs being turned on and off etc.). Only running DCS, 0.75% resolution on both to keep things like for like, no mods/upscaling/FFR/sharpening.
Essentially start the mission, fly down below the cloud level and point it at the green smoke, record frametimes, change settings, rinse and repeat. It's not scientific or taking an average, just a fairly comparable snapshot of what's happening in the grand scheme of things. Doing a static cold/ramp scenario would probably garner more repeatable results but we don't spend 90% of our time on the floor in DCS, it also normally biases the load to be more CPU intensive (in my experience) so again not truely indicative of a normal flight situation.
Obviously last night something had gone majorly wrong with my GPU/drivers and had really screwed up the OXR MR whereas everything else was consistent, solved it at 2am and it's now running the same frametimes
As per my very first post in the thread, there isn't much in the actual peak fps improvement, it really shines in the smoothness and stability of the frametimes though; testing carried out by the varjo users showed a very significant increase in the low percentages. Steamvr was jumping 4-5ms constantly all the time, introducing stutter with no reprojection active and still hitching a little with reprojection. Now I'll happily admit that the reprojection implementation in OC_ACC isn't quite as good as SVR yet but that should come with some more development time
agreed, reliable ms/fps measurement is not the simple
the best would be ingame implemented standard benchmark...
well, OXR is the right way to go on, no doubt. I have played with it yesterday for the very first time.
for me personally, caveats are the lack of kind of NeckSafer and not so polished MR
on the other hand, I was always struggling on complex multiplayer servers, with may units etc.
I have tried server with many units yesterday with OXR, where I was always limited by poor CPU frametimes (above 18ms).
Now, with OXR and MR locked it to 30fps it is very playable. Nice experience.
-
3
-
-
On 3/31/2022 at 1:29 PM, edmuss said:
Will do, it will be later on tonight though
Off the top of my head my CPU normally sits around 9-11ms I think with OXR based on a simple mission.
edit: @wju did some back to back testing,
I think my OXR MR is currently bugged as I've got jelly eels all over the place again and post gpu overhead is currently blowing it out of the water
OC_ACC MR off
CPU 10.3ms GPU 15.5OC_ACC MR on
CPU 10.9ms GPU 20.5
CPU 10.4ms GPU 15.2SVR MR off
CPU 9.6ms GPU 15.5SVR MR on
CPU 9.7ms GPU 15.3Figures are nominal but +/- 0.5ms give or take, apart from the OXR MR which whilst still pulling 50fps is reprojecting down to 22hz bracket.
I don't use MR as it is and SVR to OXR performance is virtually identical without it, CPU is slightly higher but that's not an issue. What you can't see from those numbers however is that SVR MR off was stuttery at >60fps when looking round the cockpit whilst at the same time being more blurry than OXR which was far smoother and much better clarity.
Once the OXR MR bug is fixed then I would say that performance is likely to be comparable across the board.edit: sorted the GPU overhead issue, all the same frametimes now, SVR still looks like I have a guaze sheet over my head and it stutters around the cockpit.
nice, thanks
in what mission/map/module did you take the measurement?
-
2 hours ago, wju said:
I can confirm CPU significant frametime impact in OpenXR, unfortunatelly negative
I did a short test, for both openXR and steamVR, motion reprojection OFF and forced ON.
Instant action, Cau/FA18/Ready on the ramp, looking strait ahead - it is my usual simple test,frametimes measured:
openXR MR off: CPU 14,8ms GPU 10.8ms
openXR MR on: CPU 15,0ms GPU 13.6mssteamVR MR off: CPU 9.6ms GPU 13.9ms
steamVR MR on: CPU 9.7ms GPU 14.4msmeasurements are repeatable with deviation +/- 0.3ms,
SS in both steamVR and OpenXR set to 60% (my usual steamVR value)
well, from my point of view, the real benefit of all that openXR fidling is the ability to run VR reprojection locked on 30Hz
(I have Reverb G2 and fly MR on only) but openXR is so CPU hungry and lack of VRNeckSafer...the choice is, as always, based on personal preferences..
culpa mea, sorry guys, I have switched CPU/GPU figures for steamVR, the correct result are as follows:
frametimes measured:
openXR MR off: CPU 14,8ms GPU 10.8ms
openXR MR on: CPU 15,0ms GPU 13.6mssteamVR MR off: CPU 13.9ms GPU 9.6ms
steamVR MR on: CPU 14.4ms GPU 9.7mssteamVR measured by fpsVR, openXR values read via frame tining overlay
-
31 minutes ago, edmuss said:
Could be the CPU but I wouldn't have expected there should be that much in it.
may I kindly ask You to perform your own measurement? or did you any already?
the best both MR on/off
-
17 minutes ago, edmuss said:
Curious that you've gained so much CPU, but you have plainly demonstrated that you're 3ms faster on GPU which to be fair is a pretty good uplift. OXR MR off you've gone from 72 fps with SVR to 92fps with OXR, in DCS I would call that pretty significant. I would hazard that something's either not configured right or has gotten sideways on the installtion - resetting the SVR openxr runtime from WMR to SVR and back again has been shown to improve some oddities in the system and for the 6 seconds it takes to do so might be worthwhile
the 3ms GPU gain is present only in case MR off, which is not relevant for me.
When is MR on, GPU gain is still there, but far less, see figures.
well, it is all about my system (poor CPU maybe?), my settings, my preferences. Others can have it vastly different.
-
On 3/30/2022 at 1:14 AM, Pikey said:
Cons: FPS is lower, it struggles worse at low FPS than SteamVR at the same FPS, can't use fpsVR, its a mod and resets with updates, appears to be more CPU bound and performed badly in MP, some of the settings in the Dev tools dont seem to work well or at all!!! edit: steamVR figures have switched CPU/GPU values, the coorect ones, see few posts below!!!
I can confirm CPU significant frametime impact in OpenXR, unfortunatelly negative
I did a short test, for both openXR and steamVR, motion reprojection OFF and forced ON.
Instant action, Cau/FA18/Ready on the ramp, looking strait ahead - it is my usual simple test,frametimes measured:
openXR MR off: CPU 14,8ms GPU 10.8ms
openXR MR on: CPU 15,0ms GPU 13.6mssteamVR MR off: CPU 9.6ms GPU 13.9ms
steamVR MR on: CPU 9.7ms GPU 14.4msmeasurements are repeatable with deviation +/- 0.3ms,
SS in both steamVR and OpenXR set to 60% (my usual steamVR value)
well, from my point of view, the real benefit of all that openXR fidling is the ability to run VR reprojection locked on 30Hz
(I have Reverb G2 and fly MR on only) but openXR is so CPU hungry and lack of VRNeckSafer...the choice is, as always, based on personal preferences..
-
1
-
-
On 1/31/2021 at 9:04 PM, Jharvis said:
you could try this, i deleted all models and removed the background. For the few minutes of testing it looks like complete black:
DCSWorld\Scripts\DemoScenes\sceneVR.lua
thanks for sharing - I am maybe a bit late to reply, but I had found my own solution before You have posted your scene. Yours is more simple, than mine, as I use empty dds file.
So, in accord of Occam's razor, yours is better.
-
as said many times elsewhere - more options
modifier defined in module is ignored by UI layer until is defined also within UI
in Controller & Assignment Bugs - General
Posted
the issue is not global/local but the behavior of UI-layer; i.e. you have to define all modifiers from all modules also in UI-layer to make settings behave correctly