Third time’s a charm? I hope so. We cut a new release candidate this week.

The main reason for the new RC: the very latest AMD drivers (2020.8.3) crash X-Plane in Vulkan. We are working with AMD on this; in the meantime you can run the 2020.8.2 drivers. We wanted to get an error message into the app guiding users – hard crashes were spiking in our auto reporting.

Since we were recutting, we also included fixes for non-working manipulators, and some GL clipping of plugins and the map by airplane wings.

I expect this RC to go final, although I realize I’m tempting fate by putting it in writing. We have a few more fixes we are sitting on for an 11.51 bug fix patch to avoid introducing last minute changes in RC.

About Ben Supnik

Ben is a software engineer who works on X-Plane; he spends most of his days drinking coffee and swearing at the computer -- sometimes at the same time.

72 comments on “X-Plane 11.50 Release Candidate Three

  1. Random questions not related to RC.

    1. Does x-plane have the concept of “retro reflectors” for airport signage and markings? (Think normal roads and how signs and other markings reflect more light than you’d expect than with two /r^2 trips)

    2. Are their any thoughts to letting “AI flies airplane” to use an FMS route a user adds in? I know it’s a flight simulator, not a sit simulator, but it’d be cool.

    3. Any thoughts to using the existing engine heat shader as a generic “heat off the runway/distortion at horizon” shader?

    4. Similarly, shaders for runways/roads that would ease in some reflections to give appearance of being wet?

    As always I know this is off topic and I won’t be butthurt if you don’t answer 🙂

    1. 1. We don’t have retro-reflectors – interestingly I haven’t seen them treated in the major PBR game engines either. With a deferred renderer we’d need some way of marking them as a “special” material.

      2. No – we have command line tricks to tell the AI to fly places for testing, but it’s really not meant to be a fancy user facing feature.

      3. Yeah – we may do that some day. You can make heat particle effects _now_ using the particle system, so this can be done in 11.50 with a plugin and a lot of hand-waiving trickery.

      4. I’ve seen people do statically wet textures – we have work in progress to make this dynamic.

        1. Both 1 and 4 made me think about a different game, FS19, that uses GLSL based shaders and an addon can include their own shader code, as the base shaders come with ‘injection points’ for that, to completely manipulate vertex and fragment shader behavior. It is used for reflectors, and wet surfaces there among other things too.

          Afaik, for Vulkan, a similar technique could work allowing addon devs to include SPIR-V code (bytecode that can be compiled from GLSL) to modify shader behavior in a much more low level, but supported way down the line, as opposed as through some ‘keep your hands off me’ internal art datarefs alone.

          I hope LR will consider such an extension point down the line. Even though increasing any public surface also restricts how much room for change you have internally, I hope that when the shaders codebase stabilises over time, such a thing would be possible.

    2. What i personally do with runway markings for example..just give it a “smoother” roughness value so reflection will be less scattered, hence feels more reflective. Roughnesa in x-plane is a normal map map so it can be stored within the same texture.

      1. Right – this is a good idea even if we don’t have proper retro-reflectors – in practice the layer of paint may make an effectively less rough surface.

  2. I don’t know what happened to the reflection, but since RC1 it looks like blurry shadow, a shimmering black thing with little resemble to reflection.

    I wonder if I’m the only one who has this issue or it is a known issue.

    1. Is RC1 your first beta? The reflection shaders are completely redone in 11.50 and especially on Vulkan. The reflections slider now does notably more and with a less of the FPS hit, so it’s safe to set it higher now.

      1. Wait, let’s be clear about what happened: the tech in the engine is TOTALLY unchanged – what changed is _what the reflection slider controlled_.

        We have two kinds of reflections: planar water reflections (introduced in v10 I think) and cube map environment reflections (introduced in v11). The cube maps are needed for PBR materials.

        IN 11.40 and earlier, the cube maps would update in slow real time if the reflection slider wasn’t at minimum – this is real low and therefore the only way to use the sim with high fPS was reflections off. Therefore no one could get planar water reflections without eating it on FPS.

        In 11.50, cube map update is always only-when-lighting-changes, so raising the slider never gets you live updating cube maps and it never eats FPS, letting you access the always-available planar reflections, just without the hit.

        1. Can you describe the difference between cube map reflections and planar reflections? Is the cube map stuff… windshields and airplane bodies, and planar is water, and reflection of scenery geometry?

          1. Planar reflections are only for water. Cube map reflections are used for all solid non-water surfaces (aircraft, scenery, you name it) as part of the PBR lighting model.

          1. Speaking of sun angle, anyway to still make shadows work in lower angles? The exact momment when shadows are turned off is a killer a bit when it comes to immersion:-(

    2. I meant the 11.50 r1 to r3 versions.
      The water reflections are bad in VR… blocky and… well not as good as I remember the beta ones.

      From what I read it means the planar water reflection.

      I need to test this in 2D mode too, but I wonder if others have same issue.

      It remind me the blocky shadows in the cockpit.

  3. Thanks for getting that manipulator fix in for 11.50r3! It seems to be working well now in our planes.

  4. Sorry but this is the worst RC (Vulkan) so far for me in VR (Vive Pro and Oculus Rift cv1). Blurry images and shimmering (like AA/AF are missing) and some minor stuttering. Both RC1 and RC2 were much clearer and stutter-free. Stock scenery and no plugins btw. I have filed a bug report with my log files. I sure hope this improves before a final release!

    1. I don’t think we changed anything between r2 and r3 that could possibly be related to this. If you have blurriness we can collect some telemetry and look at it, but nothing we did from r2 to r3 (add a clear of the depth buffer, use double precision for one manipulator transform, and check for 20.8.3 drivers) would explain this.

      1. Ya, I wouldn’t have thought it should effect me either, but it has. Maybe check with Andy to see if he tweaked something while you weren’t looking, lol!

  5. For me it is 50% better than RC2, it worked on Vanilla version, but still crashes on my main version that have custom scenery and some plugins.

    Already sent the but on the automatic bug reporter.

    Also, the bug that I have created for RC1 about the lost of FPS, I’m starting to believe that it is related to the new NVidia driver… can’t be sure about that, but I think that you are already checking on that.

    Also, the Log.txt doesn’t shows anything to me that might suggest why did it crash, is there somewhere else where I could look into?

    Thanks

    1. Ben, Just a quick update here.
      About the crash, it was related to my CPU Overclock, after disabling the overclock the crashes have stopped.
      About my belief that the performance drop was related to the new driver, I just discarded that possibility, have uninstalled 452.06 and installed 445.87 again, had the exact same performance.
      Moved back to 452.06, made a few tests running on Vulkan and OpenGL, on Vulkan I’m getting 22 FPS on NY Area, and on OpenGL I’m getting 20 FPS on the same area.
      On both my CPU stays with 25% of utilization and 45ms on frame timing and my GPU with 53% utilization and an average of 20ms on frame timing.
      Not sure if this helps you on something, but this it what I was able to check so far.

  6. You may have tempted fate Ben… Switching between r2 and r3 at YSSY with real-time weather enabled from ActiveSkyXP at 0600 4/9/2020…

    11.5r2 FPS = 35-36
    11.5r3 FPS = 15-16

    Not connected to VATSIM. Just sitting at the Gate A12. Although I am running a 1080ti eGPU performance went south with the latest update.

    I know I should file a bug but would anything have changed? I am running OpenGL as performance using Vulkan (Windows 10) is very poor.

      1. Not to be taken seriously:
        you should release a rc where nothing changed at all, just create some false changelog. And then wait for the “OMG, worst version evar, totally destroyed my fps. And all textures are now a blurry mess.” comments. And just chuckle…

          1. OK, I think I managed to sort this out now.  I rolled back to 11.41 and got this running properly.  I then copied this folder to a different drive and renamed it.  I then updated 11.41 to 11.50r3.  Now all seems fine and I get exactly the same VR image quality and performance I got with 11.51r1/r2.  I’m not sure why this fixed my problem, maybe the shaders got corrupt somehow and reinstalling them fixed.  Who knows?  Anyway, I thought I’d let you all know.

        1. Sorry mate but your comments are not very helpful. All any of us can do when going through beta/rc version updates is report what we see. In my case, in VR, it’s very easy to both see and monitor results immediately with tools like my SteamVR performance graphs with my Vive Pro. I can also monitor this with my Oculus Rift via OTT performance graphs. WRT to R3 I found that this visually was worse than r1/r2 and I got a lot of stuttering, esp. when turning around the yaw axis. With bot headsets I found I needed to lower my graphics setting (including super sampling) quite a bit in order to maintain a decent 45fps and reduce stuttering. Of course that only makes visuals even worse. Since I flew in VR with r2 and then updated to r3 I was able to see immediate changes. Whether or not the developer says this should not be the case, or whether you think that this is BS, does not change things for me. If you and Ben wants everyone to be complacent and only post atta-boys I guess my bug reporting days are over. Rant Over and Out.

  7. Short note: On the AMD driver crash: I had no crash in r2 until I upgraded the driver. Now in r3 I don’t have crashes even though using the current driver…

  8. I was so pleased to see the visual ground and cockpit shadows were almost smooth in RC2 – now I am back to the flickering jaggies of many past releases. I do find it dissapointing as I thought you were almost there in fixing a long outstanding issue.

    All else seems good!

      1. And yet visually they have with me! Not perfect in previous version, but close to it. Now they are jagged and thin ones are broken And flickering like way back.

        I have the other new flight sim, but still prefer X-plane with my community and purchased mods. I use visual GA flights and I think my X-plane looks better flying in the UK. Oh – and my house is not correct in the other sim anyway!

        I haven’t changed anything of late in my Steam version, but have decided to now try the non beta version to see if it looks better.

        1. A you sure? That sounds for me awfully alike someone or something have messed with art controls, like “enhancing” lua scripts. I had some terrible flickering with xVision for Lua.

  9. In this release, OS Ubuntu 20.04, AMD gpu, privative drivers (amdgpu-pro) all working OK (openGL and vulkan). Very good FPS (>60 fps in vulkan vs >50 fps in openGL).

    Good job!!! Thanks

  10. @ Laminar

    I wanted to submit a bug report but received the following message in the Firefox (80.0.1, Windows 10) developer console when using for bug submission form:

    “The character encoding of the HTML document was not declared. The document will render with garbled text in some browser configurations if the document contains characters from outside the US-ASCII range. The character encoding of the page must be declared in the document or in the transfer protocol.”

    1. I’ve addressed the issue causing that error message to appear, but didn’t find it was preventing bug report submissions under Firefox 80 on Windows 10. Were you able to submit your report?

      1. The error message seems to be gone, thank you. I tried submitting again and nothing seemed to happen after clicking the “submit” button. Then I tried again and it did work. Maybe the issues was that uploading the screen shot took so long that I THOUGHT nothing was happening (where I live, the upload speed is very low). A progress bar for the file uploads would be nice.

          1. This is also a possibility. However, I used the form before with the ad blocker enabled and it worked.

        1. Thanks for your persistence. You might be right about the file upload, and the lack of messaging for it, being the issue.

  11. Hello, when I fly over the ocean at high altitude, I observe some geometric shapes on the water, they are like patches of a different color than the one forming a kind of rectangle. The rest looks great and the performance excellent, very happy with the work you are doing, regards

    1. Same here. It also depends on camera angles, I have noticed. I can sometimes make it disappear by looking around.

      1. I have them as well, I filed a bug report with pictures. Both with the grid lines on the water and the pixelation in the distance that Greg mentioned.

        I am not sure if this may be an ortho scenery thing as I have the entire US territory loaded.

        All else runs great and averaging 45+ frames in VR. Nice work!!

  12. Hi,

    I still get better frames and less stutters in windowed mode than in full screen mode.

    I filled this bug XPD-10928 early July, it seemed to be fixed in b17 but it’s back since RC1

    rx480

    b17 was smoothest for me… I can file a new bug if it’s any help.

    Thanks.

  13. I know there wasn’t supposed to be any change related to performance/stuttering, but me too I’m experiencing a bit more stuttering in VR compared to the RC1 (I skipped RC2). Maybe installing the “other simulator” is causing this lol.

    I’m not familiar with what exactly going final means, but I still have the low CPU/GPU usage behavior in VR when flying over big cities with lots of buildings (mentioned in the RC1 comments), will this be addressed later?

    The one other issue that bothers me is that trees look brown when the sun is behind them (even if the sun is not directly hitting them). This looks really strange in VR, I tried to find info about this but could not find anything, is this something known? (here’s what I’m talking about: //i.imgur.com/O14syKe.png )

    Cheers

  14. Airplane movement in and out of cockpit, STUTTETRING. Camera movements, inside, outside, excruciating suttering. The changes in color and brightness of the water do not matter to me, the stuttering is unbearable … I see that you are spoiling it more and more with stuttering …

  15. Hi Ben,

    Just submitted another bug report about flickering or blank displays. Just did a clean install with no addons and tested using the LR B737. I’m using Pro Vega 48 card on Metal. Please could you comment on this?

    Thanks,

    Jeff

  16. Wonderful experience with 1150rc3, great frame rates, the sim runs smooth and reliable (iMac i9, VEGA 48 graphics). Not a single crash since rc2 although I’ve done a huge amount of flights.
    Thanks to the team, keep up your great work!
    Benedikt
    x737project

  17. While developing a plugin, I noticed a bug with XPLMInstanceSetPosition where dataref animations driven by values in the “data” array do not update while replay is in motion, but will update while replay is paused. Verified broken for MacOS and Linux (and I don’t know about Windows). I filed a bug report.

  18. Xp-11.50 RC3 works better when Vulkan is disabled in graphics settings. In RC2 I had no problem. (i7 9700KF, RTX 2060S 8GB, 32GB RAM, Windows 10).

  19. Hi. I wanted to play xplane again after a while. Vulkan mode doesn’t work for me on Linux.
    Linux version 5.8.7-arch1-1, AMD RADV POLARIS10 (LLVM 10.0.1)
    Submitted data through crash autoreporter.

    1. With privative drivers (amdgpu-pro) all worked ok for me (Ubuntu 20.04). There are problems with open drivers.

      1. I have tried Mesa RADV, AMDVLK open and propertiary AMDVLK from AMDGPU-PRO, all with same result, crash on startup.

  20. Since the beta 14 or later, I am enjoying pretty stable and faster frame rate. Thank dev team for that.
    A couple of very minor issues I have are (1) regenerating plane thumnails causes CTD at least with zibo 737 and rotate MD-80, (2) the sight of bottom of the splash screen during the load up in VR is blocked by the backdrop ground image. If I have to see the loading progress bar, I have to lower my head down to see the obstructed progress bar.
    Otherwise, I love and really appreciate the progress and performance improvement made by the team so far.

  21. I now find

    “XYZ.obj uses blend glass but is not set to glass lighting”.

    for old freeware scenery objects in the logs. What edit or attribute addtion in the .obj files would fix the object errors and keep the log clean? No information to find about this anywhere, at least not for people not familiar with GFX lingo.

      1. Thank you very much!

        I figured that it’s safe to ignore, but it’s likely better to change that in a future scenery update so people don’t complain that “it’s not Vulkan compatible”. 🙂

  22. ouch… thanks bed for the update, sadly another dip in the wrong direction for VR users.. stutters are back 🙁

  23. I don’t know if others have this problem too, but with RC3 the joystick and keyboard profiles are not loaded correctly. X-Plane forgets the assignments I made

  24. I seem to be having VR dropping out since installing RC3. I can get the vision back by unplugging and plugging into the graphics card again (GTX 1060) but can’t get sound back although Active Sky still can give me the weather on 122.00. Could this be an issue with the aircraft rather than X-plane? This is happening in the TBM900 although as this is happening after an hour or so flying, the fact that I generally fly the TBM that might not be the case.

  25. I don’t understand why it keeps high fps and the stuttering is more pronounced in each new update. When you move the stage, inside the cockpit, out of it, or move your view, it’s a mess and the loading on a single thread of the processor is blatant and outdated. You go from bad to worse with Vulkan.

  26. RC2 and some beta`s before was very fine for me, smooth no stutters Nothing , now with RC3 i have after 15 minutes flight stutters with pan view or on plane turnings. Cache and preferences always deleted. no Plugins running. (maby memeory leak in RC3/R3 now ?

  27. If you end the unbearable stuttering, you will gain many followers. If you continue with that problem you will lose many clients. As you don’t use the rendering engine to use the GPU, you will end up with X Plane … Either you can’t or you don’t know, because I don’t want to think of something darker …

Comments are closed.