divinee

If you change something from your surroundings you must reset the environment data and teach it again. Otherwise the goggles don't know where they are pointing.

The only solution for OP's problem is to decouple mouse totally from the VR viewport. Solution what VirusAM tries to explain doesn't work if you cannot keep your head still enough to hit buttons.

This is what everyone has been trying to explain.. Moving head also moves mouse cursor in all cases -> mouse cursor is coupled with vr viewport and if you shake your head it is impossible to hit any buttons and that's what OP tried to explain. The mouse should be totally independent like shown in the YT video posted by Taz1004. This have been on the wishlist for few years now and quite a lot of people have been asking for this. Maybe @NineLine or @BIGNEWY have some info about the situation?

VirusAM Please try this: Put your cockpit mouse cursor on top of the button you want to press and slightly move your head. The cursor also moves if you dont move the mouse to compensate your heads movement.

After ticking that option the mouse movement is still coupled with head movement. You can move mouse separately but moving head also moves the cursor.

I managed to get to the computer and updated dcs from previous to latest OB (18.11.). Seems like last patch fixed almost totally all vram problems what i had. I still have to run more tests but looks very positive. Even had about 5fps bump on the perf. edit: I run about 3 hours mp testing in very hevy PvE mission with syria. Managed to get that vram leak and coupe times pressing F10 still almost blacked out the vr headset (went black for 10s but woke up). But still the performance was generally veery much better than in previous OB.

Lots of people are compensating this now with incredibly expensive GPUs but i think that this is something ED really should be looking at with high priority. No point in investing almost 2000€ for gpu because game is bugged and handles vram badly. Last summer and at the beginning of autumn we didn't have this kind of problem and about after the release of Syria it popped up. Lots of people have been reporting exactly same symptoms after that. For the explanation that "Syria just eats more vram", yes it does, but this leak now exists also in Caucasus and PG. It just takes longer to leak enough vram to make that blackout/stutter etc. happen.

If you can, try to first spawn into a tomcat in caucasus, look around for a while and then crash the plane. After that go to some heavy mp server with syria map and spawn into a hornet or f-16. After spawning look at fpsvr gpu frametimes and keep hitting F10-F2-F10-F2. If it makes that same stutter and blackout it’s propably because vram leak. I’ve been trying to find out consistent way to reproduce this for recording and documenting but still ”nojoy”. Sometimes that works and i can reproduce the vram leak and after that the symptoms are absolutely same as yours. That other topic about steamvr crashing had propably few different cases mixed in to the same topic and OP’s problem was because of windows hardware acceleration or something like that. Edit: This kind of problems has been around whole autumn and lots of people are reporting same kind of issues. Would be nice to hear from ED that are they inspecting something and is there anything specific what we should be testing to help solving this problem?

Nice test, thanks! Are you sure about that resolution? I've understood that SteamVr SS % and DCS PD work differently. First you must calculate resolution after DCS PD -> (2160*1.3)*(1200*1.3)=4 380 480 pixels (or 2808x1560). Then you use SteamVR SS% to multiply that pixel amount -> 1.5*4 380 480=6 570 720 pixels (3439*1911). Someone wiser correct me if my calculations aren't correct.. btw, are you able to test i5 CPU with about 5ghz OC? Would be interesting to see how it compares.. edit: I learned that Vive's native resolution is propably not at the 100%? So we have to calculate the amount of increase in supersampling and my calculation is not correct. The result is propably bigger depending on the native SS% number..

I’ve had that ”off” all the time. Tried to put it on for testing and had massive stutters so i couldn’t make proper testing. I’m still quite certain that the root cause for the crash is bad vram handling and vram leaks (which occurred at the same time as Syria).

Vram usage in syria is always maxed out. Even with everything low and 0.8PD. Windows requires 800mb for some recovery files so i put 800 to 800. I didn’t have time to test with bigger page file. i’m not absolutely sure about this and it is always possible that i wasn’t able to trigger that bug causing crashes. Edit: I will try today with bigger page file and post the results.. Edit2: I wasn’t able to replicate the results anymore. I have make more testing..

I just had a time to make some testing. i went to a big PvE server which had Syria running and spawned in F-16 in Incirlik. I started to hit F10-F1-F10 several times. Same time i checked fpsvr frametimes and memory usage. GPU memory stayed at about 7,7GB but frametimes started to go up from 20ms to more than 100ms and didn't return to original 20ms. If i'd keep doing that i know that the steamvr will flip. After that i removed windows page file to 800mb and restarted computer. Then i repeated the same procedure but this time GPU frametimes stayed at about 20ms. This only worked in Syria. Can anyone else try this?

I can quickly answer for this from my part. I haven't had time to run all these tests but i hope this helps My system: i5-9600k @5GHz with AIO cooler Asus RTX2080 Super OC 8GB 32gb 3000mhz ddr4 1tb M2 ssd hp reverb 1) HMD goes black BUT i can use hand controllers to get mirror view visible in my HMD so i can land after that crash. Perf is really bad in that situation and fps is like ~15-20 2) DCS is still running after crash 3) Problem is with steamvr crashing because WMR softwares and DCS keep running (like in part 1) I mostly fly very big PvE missions and crashing is not the biggest problem because if i stay out of F10 map i haven't managed to make the game crash (yet). The most annoying thing is that everytime i die or change slot i have to restart everything to clear vram. Otherwise i will get huge fps drop, stutters and fpsvr shows GPU frametimes go from 15-20 to about 30+. Cpu frametimes stay 9-13. After that fps drop the risk of crashing steamvr when pressing F10 increases a lot. Also tried with everything low and 100% SteamVR SS / 1.0PD without help. Last spring/summer i was able to fly from supercarrier and change slots with even 200% steamvr SS and high textures without any problems. That "vram memory leak" started after the patch when Syria was released (or possibly next? not 100% sure) and i'm quite sure these crashes and bad vram usage are releated. Edit: I forgot to mention that vram leak happens with all maps, not just with Syria

Same as this? https://forums.eagle.ru/showthread.php?t=287897

I also have this and few guys in our community also has this same problem. Crash mostly happens when opening F10 map in Syria and started occuring couple patches ago.. Also, i'm not sure if its connected to this but if i'm staying F10 map for too long and return back to F1 view, my fps goes down from 45-50 to 35 and stays there. Same fps drop occurs when i change slot or die in MP. After that fps drop it's about 90% probability that pressing F10 will crash steamvr. If i dont' change slot, die or press F10 i can run with 45-60fps even very hevy multiplayer missions without steamvr crashing. I've run with same settings since spring without any problems before this and tried all possible repairs, cleanings, resetting shaders, shadermods, drivers etc..

+1

Absolutely want this! Would be nice to make some pitching decks landings!

I tried to reproduce this few times after latest OB and seems like there isn’t fps drop anymore. Thank you ED!

I had time to make more testing and i can confirm now that the fps drop happens only when using brakes. If i take off and land again the fps drops even more. For me the fps drop still happens whatever direction i’m looking at and only thing what restores the situation is restarting the mission. Weird thing is that only gpu usage is going up and if i don’t touch the brakes i can keep a steady 45fps in freeflight quick mission.

I noticed at the weekend that fps drop possibly occurs only when using wheelbreaks after landing. I'm not 100% sure and i will make some more testing. My testing has been without wake turbulence, online and offline.

No improvement after latest beta update. If there is something i can do to get more data about this issue, let me know. I’m more than happy to help if it helps to get my favorite module playable again.

I still have this same problem with the latest OB update

I made some more testing yesterday. I updated windows and nvidia drivers to latest versions which didn't help. Also tried with no mirrors and didn't notice any difference. Then i made belly landing and didn't have any noticeable fps drop. So the issue is possibly somehow related to landing landing gear touching the airfield. EDIT: If i land gears down to the side of the runway on the grass, the fps drop does not occur.

I also tried with clean mission, both Persian Gulf and caucasus and still had fps drop..

Made some more testing and noticed that after landing, the stutter starts at the halfway of runway. Same thing online or offline and only with hornet. I checked with fpsvr and when the stutter starts, gpu frametimes go to something like 60ms from 20ms. Cpu frametimes stay normal ~20ms. If i change to spectator after landing and look straight up to sky cpu frametimes drop to about 10ms and gpu frametimes stay at the 60’s. Edit: As long as i stay airborne i dont have any problems. I can run very long stints and game runs very smooth Edit2: I made a youtube clip about this. As you can see, something happens when rolling on the runway. Frametimes slowly go from 20 to 60. What i didn't notice before was that the cpu frametimes also go to ~40.