For those that missed it, here’s the recording of today’s live video Q&A.

 

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.

76 comments on “X-Plane Developer Live Q & A

  1. Sounds like a good thing to view… but where was the announcement that this was going to take place?

  2. Thanks for the Q&A!

    I was wondering, re: clouds and sky (and take this as an anti-complaint, you guys are kicking butt);

    Is the model now that there is a sky texture (based on altitude, time if day, etc), clouds that have some shading, and atmospheric scattering, all as kind of separate things that layer over eachother? If so, that sounds really hard to balance and might be the cause of some of the issues, like too dark clouds, and scattering being too white instead of starting closer as a blueing/decontrasting effect (because you are avoiding cyan clipping when the scattering sits atop the sky colors).

    Would it simplify things to remove the sky color mechanism and replace it with in/out scattering, and make clouds basically just white blobs where the shading comes purely from the sun, and they fade into low contrast blue shadow type things in the distance?

    I’ve been looking at the sky a lot lately thinking about this and noticing how clouds look at different sun angles, humidity levels, etc and it really seems like it all comes down to scattering (Raleigh and mie?), and volumetric lighting/shade.

    Or am I asking you do so something conceptually simple, but crazy pants to compute?

    1. The sky color is related to but not the same as scattering right now – pure mathematical scattering models can make the sky (the sky is just the scattering code run out to space), but when I experimented with them in v10 I found that they weren’t accurate enough to replace an artist-drawn texture, due to the simplifications in the simulation of the scattering physics.

      Some day we may be able to get them closer, e.g. rely more on scattering.

      1. Ben,

        Thank you for your response. To give you some context for this, its not that I’m just begging for “more pretty”. I think myself and a bunch of other flight simmers are in the following position:

        1. Not enough money for flight lessons/being a pilot and/or have little kids and we don’t them to lose us to a crash.
        2. Are drawn to flight not just for procedures, flight planning and IFR (though, thats all cool and I want to learn that too) but also because the sky is beautiful and calls out to us.
        3. X-Plane, by being a super solid simulator provides the best foundation for such a sim.
        4. You guys have made movements in the direction of photorealism.

        So while some people might be being “less than polite” customers about it, for me its pure interest, having some limited knowledge of the complexity involved, and hoping someday for an experience otherwise beyond my reach (I’d settle for even 1 jump seat ride in an airliner, but terrorists had to go mess that up, right?). Laminar Research sits in a unique place for that.

        So, i know it seems silly that some of us are asking for Atmosphere Simulator 2017, but it comes from a good place.

        I sometimes wish I was a meteorologist and a shader programmer so I could just contribute code 🙂

  3. Have you considered adding a way to *smoothly* transform your airplane into a different one while flying with the proviso that the transformation time depends on airplane similarity and intermediate results may not be powered and/or stable?

  4. LR really have such a great team…and it shows in XP. thanks for sharing a great QA session

  5. Hi Ben.
    Thanks for sharing. One more question that seems haven’t been covered (Or I didn’t catch). As far as I remember, half a year ago you mentioned that increasing distance for lights at night is definitely in your todo list but it’s too early to talk about priorities. Could we expect this feature in near future? Thanks.

    1. We did talk about it. You will not see it in the _near_ future because it depends upon a change in the architecture of the engine. We do at least have a design idea that I think will help.

  6. I may have asked this before or missed it in the video, but can you confirm if VR will work on my 2017 iMac if I have a headset and steam app?

  7. This was great, thanks for doing these. Chat was a little rowdy but that’s the internet for you, I guess. I hope you guys plan on doing more of these during your regular meetings. Thanks for making our favorite sim even better!

    One thing I wanted to ask, that was not touched on during the Q&A session.. Are the first landmarks going to be in 11.10 as well?

    p.s. props to Randy for trying to keep Austin’s speaking time in check (if that’s even possible) and trying to pick questions

    1. Ha ha, Austin’s like a tornado – you can talk to it and encourage it to change directions but it may not work. 🙂

      I think we may do another one this winter when a bunch of us are in one place again.

      I think the first landmarks will go into 11.05.

  8. Thanks to everyone at LR for the presentation. I feel like you guys covered a whole lot more ground in this Q&A than at the FlightSimCon. It’s good to hear about things going on now and what will happen in the future, coming straight from everyone’s mouth.

  9. can you guys also make the moving map in xp-11 to be able to zoom further out, it would be nice to see where you are state wise like on a moving map display on an actually seat back screen

  10. Hi Ben. You didn’t sound confident at all about fixing ground lights showing through clouds at night. I think it totally kills X-plane’s absolutely stunning night lighting, to the point I don’t bother with night flying anymore. Please please give us some confidence you are going to fix this.

  11. Hi Ben

    I am trying to write an email to Phillip regarding the GNS530 but I am not sure which email he uses.

    Problem: If you i.e. have flown from KLAS to KSEA the GNS530 have this route stored. But if you were to fly then KSEA to KLAS, you would need to select airport and dial the GPS nob for every four digit ICAO code again. Should be enough in my mind to just turn the last three digits. Or in many cases only the two last ICAO identifying codes. Example: ENTO to ENGM then ENGM to ENTO… Now we need to enter ENGM and ENTO not only TO for the returned flight.

    Hard to describe really.

    Also even with latest NAV DATA from i.e. Navigraph copied into Custom Data folder with correct structure, we do not get all airports in the GNS530, lots of airports are missing, Why?

    Feel free to send me an email of this blog is the wrong forum for this discussion.

  12. Hi,

    After hearing and enjoying this Live cast (I did not hear is live though) I would like to send my comment through this board regarding the simplicity of “graphical setup sliders”, I will say out front that in 90% they do their job, but in one case I think it distracts and do wrong to the scenery, and by that I’m referring to “car density” slider (which is missing).

    Here is the case that I first start to understand other people comments on “Car Density Slider”:
    I was “flying” in Australia, well the center area and was trying to do a night landing on a dirt strip with few guiding lights created using WED. Now I know that this is a remote place, but the road around this field was swarming with cars as if I were flying above Manhattan. This was not plausible for this area.
    The solution was: dial down the Object slider, and less cars are now driving on the road.

    So I half solve the problem and created a new one, I had to lose 3D Object fidelity to make the scenery plausible to that specific area.
    It comes down to the fact that the DSF probably does not hold information about density of population and traffic on roads, thus even if merging a feature like: “Car density” into the “3D Object” slider might gain LR more control on performance, it certainly missed in plausibility of sceneries. I mean, I can fly in a remote site in Alaska, that has one road, but if I’ll dial the object slider to maximum, will I see traffic density as in New York ?

    I prefer not to hack files in order to achieve what I want, I think that some features can’t be merge to others without fully realizing the impact on the overall scenery feel.

    My 2 cents
    Cheers

    1. I agree that one-size-fits-all for car density is not good! We have data that we could potentially use to make this more accurate.

      1. I’ve had thoughts on this too. Not only could car density be managed with performance in mind, but also a sense of immersion.

        Could traffic density also be a function of:
        1. Road Density (total length of roads per unit area). A single remote road in Australia may yield a density of 0.01 roadkm/sqkm, while a city may yield a density that exceeds 100 road km/sqkm
        2. Time of day (almost empty at night, and busy during morning/evening commutes).

  13. Hello Laminar team,

    Thank you for sharing this insight on te great road ahead for X-plane.

    Regarding te coming native VR I have the following question.
    Are there plans or is it considerd to implement Leap Motion support?
    I am using FlyInside-XP and fly with the Oculus Rift CV1 december 2016 version.
    I therefore do not have touch controllers and invested just before flightsim con with the big VR news in a Leap Motion sensor.

    With regards,
    Albert van der Heide
    A virtual flying dutchman

    1. We’re not looking at Leap Motion for the initial release of VR suppose – we’re focusing on the controllers made by the headset manufacturers.

    2. Leap Motion also has a limitation where it cannot interact with a “panel” type manipulator, which depends on detection of the mouse cursor over panel.png texture pixels. Hardware pointers may or may not have this issue.

      1. Ben and Steve, thanks for the response.
        Its a good choice to at least support the standard hardware.

        I just found out that Steam VR has a possible hack by which your leap motion detected hands can be emulated into the hardware vive or touch controllers.

        I wil be testing this out
        See this URL link
        //www.vrspies.com/realitycheckvr/leap-motion-controllers-are-working-in-steamvr-how-to-guide/
        Would it be possible to beta test this for you with the coming VR addition.
        I am already using all your beta versions already.

  14. I was surprised to hear Ben wonder where the bottom layer of cloud would be for city-light illumination. Never would have guessed that would be a problem given the hundred-mile-wide cloud slabs that still show up in big weather systems every now and then. Of all the cool modeling you folks put in this software, flying directly into an example problem from a high school algebra textbook on inclined planes is a surprising and shamelessly sub-par experience. I get the translucency quirks and the random blinking reflections in the water. I get the texture thrashing and the light render distance. But a perfectly-smooth-climbing-wall-in-the-sky?? Half the time, it’s not even a 5.5 on the Yosemite scale 😉

    1. I’m not exactly sure which problem you’re referring to, but if you think it’s high school algebra, you’ve definitely totally misunderstood it. Both the problems of seeing the city lights _illuminate_ clouds above them and NOT seeing the lights themselves _through_ the clouds are problems of the 3-d graphics pipeline and computer data structures, not math.

      1. I might have totally misunderstood xplanefan, but I suspect he’s referring to the fact that sometimes, cloud decks in X-Plane assume the shape of a big inclined slab. Instead, real cumulus clouds are usually big cauliflowers with a horizontal stratiform base.

        Speaking of that Ben, I understand that you’re not revisiting the weather engine in X-Plane 11, but please consider “low cost” (from a coding point of view) improvements. For example, improving the algorithm of cloud shape generation (I think you’re using perlin noise or similar) to have more realistic cumulus shapes would be a big step for realism.

        Improvements in weather is still one of the most popular discussion topics around X-Plane forums.

      2. This was more a rant on the ugly cloud-slabs in the sky and the fact that it *seems* as though the engine knows more about the large-scale geometry definition of cloud formations than you led on.

        City-lit clouds would be nice and all, but please don’t get me wrong when I said “cool modeling” above: XP11 in general is pretty incredible.

  15. In the “ask anything” spirit, are their any plans for allowing developers to control ground vehicles like we can aircraft? It sounds like most/all of the components are in the system. For example, Austin commented on fixing the tire force physics so that it’s good enough for a racing simulator. I’m sure I’m not the first to wish we could use X-Plane for driving simulation.

    1. There is not a plan for plugin control of ground vehicles but there is a plan for a better API to put objects in the sim, which would make the creation of plugin ground vehicles a lot easier.

  16. Are there plans to increase the size of the weather “tile” or move the layers according to wind? Why not wrap it around the tile, so that what disappears beyond the edge of the tile spawns at its opposite side every nth frame? Or is there a better idea?

  17. On X-plane’s ATC system, I actually use it often and quite like it, quirks and all, but it sounds great that its planned to get an update. The main issue I have with it at the moment is when using Xenviro for weather, the ATC keeps trying to switch landing runways on me when near the destination airport. It just can’t seem to make up it’s mind, and I’m vectored in constant circles. This makes me wonder what data X-plane’s ATC is currently using to choose a runway for landing, and is this a known issue ?

    1. It uses the sim’s own weather – is X-Enviro shifting the winds frequently between very strong winds in different directions?

      1. X-enviro’s winds at ground level are not shifting, it uses metar data which doesn’t change much in a short time period. I suspect the issue could be that X-enviro uses 9 upper wind levels and as you descend (under ATC instructions) through the 9 levels the winds will change to a certain extent. Maybe that is enough to confuse ATC, if its constantly monitoring current sim wind (rather than airport metar wind) during descent ?

          1. Aha, there’s our problem ! X-enviro sets the datarefs for all 3 of x-plane’s wind layers to the same values, being equal to the wind speed and direction at the plane’s current altitude. I’m betting they do this to nullify x-plane’s internal intepolation algorithm so it doesn’t interfere with their own (because they have surface and nine upper wind layers to simulate with smoothing etc).
            So as suspected ATC is getting duped into incorrect landing runways as the plane descends through wind layers that often vary 90 degrees or more from surface winds. Once ATC realises it has you on the wrong runway, it asks you to climb again and head for the correct runway, but the climb puts you back into the original higher wind direction, and so the process repeats until you run out of fuel ! I haven’t tested but I suspect this would also affect the AI planes. Perhaps you guys could take this into account when looking at improving the ATC.

          2. I’m not sure how we would ‘fix’ that. A plugin can avoid this by setting the top two wind layers to the ‘current’ wind (and bracketing the airplane’s altitude) and using the bottom one to manage the airport.

        1. It makes no sense to ask LR to fix issues caused by 3rd party plugins. If you want better realism in ATC I would invite you to consider IVAO or VATSIM 🙂

  18. Howdy Ben,

    You have asked what interests us in regard to ATC/AI, so here it is.
    I like flying like Cowboy Austin and many other cowboys/-girls (VFR pilots). Flying as a cowboy is really great fun and adventurous, not dangerous, it isn’t the wild west. There are also rules (visual flight rules). We are aware of the airspaces (un-/controlled or special use airspaces; it would be good to see them on the map or even to plan our flights with the map), the meteorological conditions, know how to avoid collisions, …, yes we like to fly self-responsibe in the most of the time, we love that kind of freedom. We know that there are airspaces, we have to follow the commands of the air traffic controller and have to respect them. I don’t know all the details of the next step of ATC/AI, but I think, as you mentioned, it is a step forward but still far from being completed. So for further steps (if not already implemented in 11.10), we would love to see, that the air traffic controllers, VFR/IFR-pilots are aware of the airspaces and know how to behave with/in those areas. So e.g. it is possible to request a VFR departure (direction of departure, altitude climbing to), or we can request to fly through a towered airspace, …, or get (on request) flight information (e.g. a cowboy is flying towards us on 11 o’clock or a glider (this is another big topic) is climbing in front of us, or what is if we got a serious failure and have to declare an emergency landing). Yeah flying is so much more. So I’m really exited and curious how the next step of ATC/AI (and other things) will changing our flying experience.

    Questions to WED:
    -I saw under Airflow, there is an opportunity for Pattern Direction (Airport Flow), but I can’t find a field for the “pattern altitude” or do they fly a default traffic pattern or is there a database for them?
    -So I think Taxi Routes are good for non-towered airports, too?
    -Many towered airports close their towers overnight, keeping the airport itself open (different airspace). So open/close tower times for the whole airport would be useful?

  19. Dear Ben, I’m sorry, I don’t know where I can ask this (I apology):

    Existing a way for a plugin to receive an event from X-Plane when a new METAR file is loaded by X-Plane in saved to the X-Plane main folder ?

    Thanks for your help,
    Cheers
    Stef

    1. Hi Stef,

      Yes! The sim sends _itself_ the command “sim/operation/load_real_weather” when a rescan is needed. So intercept (but do not OVERRIDE) this command and you can know it happened, whether from a plugin or the sim itself.

      1. Dear Ben,

        I’m sure the trouble comes from me but X-Plane do not call this function every time and not if I click the “Load Weather” button.

        Here is my code :

        int XPMETARLOADED( XPLMCommandRef inCommand, XPLMCommandPhase inPhase, void *inRefcon)
        {
        if( xplm_CommandEnd == inPhase)
        {
        GoodWay_LoadMETAR();
        }

        return 1;
        }

        and the callback installation is ensured on Plug in init like that :
        stCommand_XPMETARLOADED = XPLMFindCommand( “sim/operation/load_real_weather”);
        XPLMRegisterCommandHandler( stCommand_XPMETARLOADED, XPMETARLOADED, true, NULL);

        of course, stCommand_XPMETARLOADED is not NULL.

        Could you help me please ?
        THANKS A LOT !!!
        Cheers
        Stef

  20. Hello Ben.,
    First I want to say thank you for the amazing job you guys are doing. during the Livestream I’m afraid there was a misunderstanding. As a result, a few of us came across in a rather bad light. I am speaking specifically of myself and a few fellow formation pilots. We tried to ask questions regarding networking. Specifically because of a few current issues. Please, Allow me to preface,. There is a very large group of close formation virtual airshow pilots, it’s a community of people who take flying very seriously and approach it with a very professional attitude. Up until this point the program of choice for this has been DCS. Now, a lot of us have considered using Xplane 11. But here is where we run into problems. Currently, there is no centralized server system. And the current peer to peer network code creates so much lag that it is virtually impossible to maintain any type of formation in any aircraft. As an example, using only 2 players in single-engine aircraft, 1 aircraft could not catch up to the other even though the lead aircraft was flying purposely slow. This happened regardless of which aircraft We chose. We even tried it with the StinsonL5 . Under no circumstance could we maintain formation or, for that matter even join up in one. So we wanted to know what you guys may be planning to do about the network code, which is sorely lacking. I don’t believe the question was understood properly. Or perhaps we didn’t ask it properly. I know this is not your area. But perhaps you can pass this along. If you would just take a look at the community you would find that it is well-established. And ready for a change. DCS is a good simulator but we feel Xplane 11 could be better.. I would also make it known that a lot of us are custom aircraft designers and scenery designers and doing these for DCS is a fiasco. Which is another reason we want to move to Xplane. I have been a very, very long time Xplane user, (since the very beginning actually) and even though I remain largely in the shadows I have been promoting it to the community profusely. So any answers we get on this issue would be greatly appreciated. if you have a moment I would beg you please take a look at my page on Facebook and see what we are about. //www.facebook.com/groups/DCSFormationPilots and my own site //www.facebook.com/JazAero

    always a pleasure chatting with you.
    Yours truly
    Jim

    1. Hi Jim,

      Is your community doing formation flying across the internet, by punching holes in local NAT layers, or by flying on a LAN? What kind of ping times do you see between machines, and what UDP update rate are you using?

    2. Yes, internet flying would be one point to build in X-Plane. I used to fly online Falcon AF, Falcon BMS, Lock on, and so on with a 3 mbps/512kbps connection… And I really don’t understand why it’s so hard to make X-Plane internet compatible and friendly. And a working copilot system would be really really really appreciated. ( I had to buy a 3rd party plugin to achieve copilot work
      – and not always works since it needs a configuration file).
      This in my opinion is the biggest and only lack of X-Plane. Internet friendly.

    3. I appreciate a post like this because honestly, the way you guys were conducting yourselves in chat would make stay as far away from that community as humanly possible.

  21. HI Ben,
    I hope you can help guide me here. As i am sure you have seen, AMD has made a “come back” with their new Ryzen CPU. Benchmark tests and reviews are calling it the bees knees. However, through some research I believe xplane utilizes a single core performance rather than multicore. I have read on your recent blogs that you are making incremental changes to that. I am wondering if maybe it is time to upgrade my i7 4790 to the ryzen 1800x. I enjoy VR and flyinside makes me nauseas despite being a great software.
    Do you think Ryzen and multicore is the way to go for xplane?
    Can you please list your ideal specs for xplane please?
    Thank you Ben.

    1. I would wait until taking a risk with AMD. Their CPUs are generally either lower performance… or very prone to overheating. I’ve made that mistake once… only very hardly I’ll go back to AMD.

  22. Hi Ben, not being critical at all, XP11 is coming along nicely, however, why do the land textures have a overall dull look to them in XP11 ?, the color saturation seems off a bit, in P3Dv4 there’s a slider in HDR that adjusts color saturation, it really makes colors come alive, maybe you guys could add a slider to control how vibrant colors are ?

      1. Yes !, Reshade is exactly what I’m talking about, XP11 looks great with it.

  23. Hi Ben,

    I lost the opportunity to see if one of my curiosity could have been answered, possibly even with addition coming directly from Austin.
    Usually in both games and simulators the AI use a simplified physics/mechanic that drives them to not overload the system also because visually speaking the result can be the same. However, X-Plane decided to go in the opposite direction with its biggest cons (more CPU required). Is there any possibility to devs reconsider this choice and begin a more cpu friendly approach that basically, if tweaked properly, can provide the same visual result.

    Thanks in advance

    1. We’re not reconsidering this choice because it’s not actually a problem. I’ve looked at the perf of lots of AI planes and it’s all in the _drawing_ of the AI planes. So the TODO for improving AI perf that we have on our road map is _not_ dumbing down the physics (which will get us almost no improvement in FPS) – it’s optimizing the drawing paths for the AI planes.

      This has been covered in detail before – the major cost of the flight model code running on the AI is ground collision tests, and they are inescapable even for a ‘simple’ flight model. The flight model has perfect multi-core threading and a plane’s cost is mostly drawing. So there’s no win here.

      1. Very good to know that since having full physics is always better than dumb model. I hope you can archive better overall performance that is not dragged down as you add new AI as soon as possible. Thanks again

  24. Dear Ben,

    Could you say me if you know a trouble from X-Plane (v11, last version) do not call “sim/operation/load_real_weather” function every time and not if I click the “Load Weather” button inside the weather panel.

    Here is my code :

    int XPMETARLOADED( XPLMCommandRef inCommand, XPLMCommandPhase inPhase, void *inRefcon)
    {
    if( xplm_CommandEnd == inPhase)
    {
    GoodWay_LoadMETAR();
    }

    return 1;
    }

    and the callback installation is ensured on Plug in init like that :
    stCommand_XPMETARLOADED = XPLMFindCommand( “sim/operation/load_real_weather”);
    XPLMRegisterCommandHandler( stCommand_XPMETARLOADED, XPMETARLOADED, true, NULL);

    of course, stCommand_XPMETARLOADED is not NULL.

    Could you help me please ?
    THANKS A LOT !!!
    Cheers
    Stef

  25. I did some testing in cold weather where altimeter error is of most concern, and there doesn’t see to be any error modeling on the indicated altitude due to temperature change, was wondering if there was a plan to implement this at some point, I know Austin is aware of it //austinmeyer.com/2012/05/26/your-altitude-is-wrong/

    I’m sure there are other things on your collective radars right now, apologies for being “that guy” who had to bring up one more thing…..

    1. I asked Austin about that a few years ago (probably it was during the v10 run). He was aware of the issue, but said that revisiting the atmospheric model was complex and something to do in the future. You should email him directly and try to convince him to implement the feature.

  26. Some thought on the live stream: You seem to have a fancy microphone, but maybe also invest into some “microphone spider” or tripod to get that thing off the table, or make everyone who knocks on the table pay one dollar for the Scotch fuel.
    Also for the non-native listeners: Try to speak not faster than three words per second.

Comments are closed.