Jump to content

Rifter

Members
  • Posts

    116
  • Joined

  • Last visited

Everything posted by Rifter

  1. Sorry, ich war für Monate hier nicht aktiv. Vielleicht nutzt es trotzdem noch dem einen oder anderen. https://github.com/drosoCode/Offculus Hier gibt es ein Tool mit dem man ein Backup der eigenen Oculus-Installation vornehmen kann. Das lässt sich dann auf einem anderen Rechner wieder reinstallieren. Könnte man im Prinzip auch von Hand machen, aber das Tool ist bequemer, insbesondere hinsichtlich dem zwingend erforderlichen Registry-Backup. In der hosts-Datei unter C:\Windows\System32\drivers\etc kann man dann die Verbindung zur Außenwelt kappen: # localhost name resolution is handled within DNS itself. # 127.0.0.1 localhost # ::1 localhost 127.0.0.1 securecdn.oculus.com 127.0.0.1 secure.oculus.com 127.0.0.1 id.oculus.com 127.0.0.1 www.oculus.com 127.0.0.1 www2.oculus.com 127.0.0.1 scontent.oculuscdn.com 127.0.0.1 graph.oculus.com Für ein Rollback zu einer älteren Runtime muss man sich im Internet bedienen. Das sind dann Backups von anderen Usern. Oculus bzw. Meta bietet das natürlich nicht an. Ich nutze eine Version von 2019, von der an die Performance in DCS nach und nach in den Keller ging. Da DCS keine erweiterten Features von Oculus nutzt (wie z.B. ASW 2.0), funktioniert das einwandfrei. Auch OpenXR funktioniert mit dieser alten Version fast problemlos (der Turbo-Mode vom OpenXR-Toolkit bringt die alte Runtime allerdings zum Absturz).
  2. I would also love to know what kind of available background knowledge went into the modelling of the Tomcat ACLS system since there is not much information around. And the stereotyped ‘it was not 100% reliable’ is something that sounds like a welcomed excuse. The only information I ever stumbled over was the story of a F-14A pilot who said that a non-working ACLS system was due to factory faults: Wrong sized beacon antennas and wrong (‘backwards’) mounted coupling actuators for the pitch commands to the horizontal stab which led to pitch up manoeuvres when pitch down was needed and vice versa. But when the aircraft was set up properly on the technical side the ACLS would fly ‘rails’ to a 3 wire.
  3. Was bei Diskussionen rund um VR immer wieder gerne übersehen wird: Bei der Bildwiederholfrequenz in VR geht es nicht primär um die Frage, ab welcher Bildfrequenz ein Nutzer das Bild als “flüssig” empfindet, wie auf einem 2D-Monitor. Die entscheidende zusätzliche Anforderung ist die Überlagerung des bewegten Bildes mit der Kopfbewegung des Nutzers. Dies (und praktisch nur dies!) hat überhaupt zum Wert von 90 Bildern pro Sekunde als Grundanforderung bei VR Brillen geführt. Genau 90 müssen es zwar nicht sein (die 80 der Rift S funktionieren auch), aber unterhalb von ca. 70 Bildern pro Sekunde ist es rein technisch/physikalisch schon nicht mehr möglich, bei raschen Kopfbewegungen eine dazu passende Bildinformation zu liefern, da die Bildwiederholrate bzw. die benötigte Bildinformation der Kopfbewegung hinterher hinkt. Ich kenne Leute, die eine frenetische Abneigung zur VR-Zwischenbilderzeugung pflegen (Motion Reprojection) und sich in VR regelrecht bestimmte Kopfbewegungstechniken angewöhnt haben, um das Problem zu umschiffen, dass sie eigentlich mit viel zu wenig Bildern pro Sekunde unterwegs sind. Die gucken dann in bestimmten Situationen nicht zur Seite oder bleiben mit dem Blick im Cockpit, bis es wieder halbwegs flüssig wird. Und erzählen dann, dass eigentlich alles ganz “flüssig” läuft, weil ja 60 Bilder pro Sekunde für einen kontinuierlichen Bewegungseindruck ausreichend sind.
  4. It’s ‘IT-23 targeting display’ for the Su-25T - not TDC slew.
  5. Have a look at this tool: https://github.com/drosoCode/Offculus You can create your private standalone version of Oculus runtime with your present v47 or any desired version (provided you have backups of them). I’m running my Rift S with version 1.43 from 2019 as a standalone system without any connection to Oculus/Meta. Fetched the old version from the web, a couple of people archived them. Stunning performance and no compatibility problems. Even OpenXR is working fine although it can only be switched on and off by registry. See also old thread here:
  6. That option is old. It is for so called single pass stereo rendering. It renders both left and right eye images at the same time in VR. It enables the GPU to share culling for both eyes. The GPU iterates through all objects in the rendering scene for only one time for the culling and then renders the reminding objects that were left over after the culling process. It can boost VR performance significantly. But sometimes it comes with rendering errors (different light effects for left and right eyes for example). This feature was never completed or approved by ED.
  7. When registry is set to none for OpenXR Runtime (= default) the Oculus runtime will be used when starting DCS in ST or MT and an Oculus headset is recognised. So I assume Oculus runtime is default for Oculus headsets (since 2.8.3.38090). Haven't tried latest update yet.
  8. Perhaps the smaller guys are modelled correct according to the time they lived whilst the bigger ones represent todays average male size? But even if that would be the case the difference in size seems quite a bit off…
  9. Breaking news: Taz1004 the Tasmanian Texture Devil did some devilish good optimizations with impressive VRAM reduction. https://forum.dcs.world/topic/323252-dcs-optimized-textures/
  10. Every VR ecosystem has its own implementation of motion reprojection - it is part of their IP. There is no cross-usage of those implementations possible. One would have to reverse engineer a VR system to do that. Motion reprojection can be set to permanently on/off or can be set to be activated when a certain amount of fps is not reached. Those settings can also be set outside of a VR ecosystems (in SteamVR or OpenXR Toolkit) - which obviously confuses a lot of people. Historically DCS supported OpenVR (SteamVR) and LibOVR (Oculus). DCS never supported WMR so there was no other way to use a WMR Headset (like a G2) than via SteamVR. That had always being a pain because the WMR plugin for SteamVR was not a masterpiece performance wise. It was a kind of ‘lost in translation’ thing. With DCS now supporting OpenXR all of that should be history because all VR ecosystems have OpenXR implemented and can now be run native by DCS. So if motion reprojection on a G2 performs better with SteamVR (OpenXR) than with WMR (OpenXR), there must be something wrong with the OpenXR API implementation on the WMR side. It would actually mean that the chain DCS —> OpenXR —> SteamVR (WMR OpenXR) —> WMR motion reprojection ( = OpenXR via SteamVR as a fallback) works better than DCS —> OpenXR —> WMR (OpenXR) —> WMR motion reprojection (= native WMR OpenXR path) Hence on a WMR device there is no DCS —> OpenXR —> SteamVR (OpenXR) —> SteamVR motion reprojection The up-scaling technologies FSR and NIS have nothing to do with motion reprojection - they can be used according to your own taste and you can always decide to use motion reprojection on top of that.
  11. I have no doubts about your specific experience - I’m very well aware of the fact that some people observe severe problems here. But that has nothing to do with the usage of a ‘good working version’ of motion reprojection versus a ‘not so good working version’. There is a big misconception that motion reprojection is something which can be changed by the user or applied as an alternative. As a G2 user you only have the opportunity to use the WMR version of motion reprojection. SteamVR motion reprojection works only for Vive and Index. Asynchronous Space Warp only applies to Oculus/Meta. And OpenXR Toolkit has no motion reprojection solution to offer at all. Selecting the option motion reprojection for a G2 in SteamVR or OpenXR Toolkit does not - surprise, surprise - select a specific motion reprojection variant. It just tells WMR to switch on its motion reprojection. SteamVR and OpenXR Toolkit only act as a frontend to the WMR Runtime. I have different VR headsets and therefore always edit the windows registry for the desired VR runtime. This can also be done by a tool: https://github.com/WaGi-Coding/OpenXR-Runtime-Switcher It does nothing else than setting the registry to the selected VR runtime. Never experienced any ‘wobbling mess’ with my G2 *). Never experienced differences in motion reprojection quality no matter in which tool I switch it on. I want to apologize for my wording ‘placebo effect’. I did not want to offend anyone. I know that VR can be a pain in the <profanity>. But it is of no help when problems are broken down to the wrong causes. That just creates esoteric and snake oil solutions. *) I tested a 6900 XT some time ago. That created a wobbling mess for all of my VR headsets when using motion reprojection. But that is due to the fact that AMD cards are not so suitable for that specific VR functionality. No problems with my 3090 at all.
  12. Motion reprojection (asynchronous space warp, etc.) is respectively the proprietary implementation of the different VR device manufacturers. It runs fully within the VR software. The motion reprojection of SteamVR cannot be used as an alternative to WMR reprojection on a WMR device. If someone thinks otherwise, I’m looking forward with great expectation to the technical explanation for that. Also I’m not aware of the fact that OpenXR Toolkit delivers its ‘own’ motion reprojection implementation as an alternative to the native motion reprojection of WMR. Even the dev of OpenXR Toolkit (mbucchia) was perplexed by the user reports about different behaviours (ghosting, stuttering, ...). OpenXR Toolkit just sets a registry key that enables the setting from OpenXR Tools for motion reprojection in WMR. If anyone experiences differences in motion reprojection depending on his settings back and forth (SteamVR - WMR - OpenXR - whatever) that is a pure coincidence or something else (probably some other mess on his machine).
  13. To be even more pedantic: SteamVR motion reprojection cannot be used on anything else than Vive and Index devices. WMR VR headsets cannot be run with SteamVR in terms of a WMR replacement - they can only use SteamVR as a frontend to make use of WMR functionalities and to allow a per-application control setting. Motion reprojection itself will just be handed over to WMR. There is an unbelievable amount of placebo effect user reports on the forum like ‘when I use SteamVR motion reprojection on my WMR device it runs better/smoother’.
  14. In the resent past sniper pods on the C-model enhance the air-to-air capabilities. This creates the ability to identify targets without using radar. Especially when the target cannot be tracked by the radar because it is a stealthy aircraft. We are talking about AIM-9X employment by the way. The sniper pod acts as a capability boost here.
  15. !?!?!? This sentence denies the fundamental principles on which our whole universe is build on - the physical laws. I spend 30 years in the industry as a development engineer and I never stumbled over a single piece of technology in which a gain of capabilities could be achieved without any trade offs. There is no such thing as a win-win situation in technical systems. The simple rule for every piece of technology in the The Real World is: The sum of all inconveniences in a technical system always remains constant. The only freedom you have is to decide which kind of inconvenience you are willing to live with. Let me therefore correct your sentence for you: It's not like in The Real World where there is always a sacrifice of one thing for the gain of another; more often than not in games there is a gain of capabilities for no real loss. Your welcome!
  16. I think it was JFK who said: "We choose to master the F-104 radar system in DCS not because that is easy, but because that is hard."
  17. Are you sure you're not unintentonally back to OpenXR again? Did you check wether the environment variable XR_RUNTIME_JSON is set? IIRC that variable has precedence over the registry setting.
  18. This is an interesting one. I played around with Process Lasso some time ago to find out, how much cores the Oculus runtime likes to have to run without any performance issues. I did this test with the Rift CV and the Rift S. My finding was that Oculus needed 2 (physical!) cores to run properly. Forcing Oculus onto 1 core created performance issues. From what I see on my 9900K the MT version of DCS dedicates 8 threads (4 cores) to the rendering process and uses the remaining 8 threads (4 cores) for the rest. I don’t have Process Lasso installed at the moment - perhaps someone could test a scenario with Oculus getting 2 physical cores (4 threads) assigned without being forced to 'compete' with DCS MT.
  19. Same experience with my Rift S. As an additional confirmation: On the Caucasus and Nevada maps I am able to get 80 fps (ST and MT version) with my settings. I can definitely rule out ASW as an influencing factor since the micro stutters still occur with the MT version with native 80 fps on the Rift S. Setting Turbo Mode On in OpenXR Toolkit leads to an Oculus crash on my side. That specific stutter phenomenon doesn't seem to be connected to frame times or fps or overall system performance. It must be something else.
  20. On my side OpenXR @100% will show a per eye render resolution of 1648 x 1776 pixels or 1504 x 1616 pixels for the Rift S depending on the setting on the Oculus side of either ‘Quality’ or ‘Performance’. Those resolutions are Rift S standard for 100%. OpenXR does not touch my Oculus render resolution by default. It also does not override OTT or Oculus Debug Tool settings, nor DCS settings. I have to chance the 100% value in OpenXR explicitly to force OpenXR to manage my resolution.
  21. At the moment the MT version of DCS forces the use of OpenXR with Oculus. Start with native Oculus is not possible with MT. So if the use of OpenXR causes stuttering on the Rift S nothing can be done against that right now. The standard version of DCS has none of those restrictions and works as before.
  22. When I’m not flying the Tomcat I’m flying the Yak with this livery. I like well worn stuff…
  23. What I would like to know on the topic of the new forge elements: Were falling off buttons a thing from the very beginning when the Tomcats went into service or did that start to happen when the jets started to age and got rattled apart more and more? My wife is great fan of the Top Gun movie. When I once showed her the DCS Tomcat cockpit in VR, she said it would look like a scrap heap and said it would totally destroy the whole Top Gun legacy and she refused to believe Tom Cruise sat in such a cockpit. She was especially disgusted by the ‘shabby textile stuff' strapped over the instrument panels and said ‘I don’t want to know whats living underneath of it’. And now when ever I fly the Tomcat and my wife is around she asks ‘Are you flying with that scrap heap again?’.
×
×
  • Create New...