tonywob

X-Plane at FlightSimExpo 2018

Recommended Posts

Tomorrow at the expo in Las Vegas, Laminar Research (Austin) are hopefully going to give a talk and show off/discuss Vulkan. 

Fingers crossed that this is going to be as hyped as many people think it will be, and we'll see some good improvements in performance and smoothness. Even if it still is early days.

Besides that, we also have demos from ORBX with their increasing X-Plane catalogue, and perhaps PMDG will surprise us 🙂

P.S. If anyone is going, please take pictures and report

  • Like 6

Share this post


Link to post
Share on other sites
Help AVSIM continue to serve you!
Please donate today!

Isn't it Vulkan? Anyway, here's nVidia's take on it:

https://developer.nvidia.com/Vulkan

Considering that Intel can barely improve CPU performance these days, any tech that minimizes the importance of the CPU will be a welcome relief.

But hyped technology has a habit of turning into background white noise once introduced. Consider Gsync, SLI and DX12 as examples.

 

Share this post


Link to post
Share on other sites

Yep thanks, I've fixed the typo 🙂

Share this post


Link to post
Share on other sites
2 hours ago, jabloomf1230 said:

Considering that Intel can barely improve CPU performance these days, any tech that minimizes the importance of the CPU will be a welcome relief.

But hyped technology has a habit of turning into background white noise once introduced. Consider Gsync, SLI and DX12 as examples.

As I understand it (and I'm not GPU geek), Vulkan is on a different level; basically the "next" version of OpenGL that X-Plane already uses.

It may not be a night-and-day difference when it arrives, so we should probably be cautious about getting our hopes up too high. Apparently it's mostly about making things more efficient for the developers. But we still may see some frame rate improvement (keeping fingers crossed). I don't think Ben and Co. would be doing all this work to roll out an additional API (along with Metal, and still supporting OpenGL for backwards compatibility), if it didn't mean good things for X-Plane.

Here's a dev blog post from 2016 explaining why they're doing this:

https://developer.x-plane.com/2016/03/what-vulkan-means-to-developers/

  • Upvote 1

Share this post


Link to post
Share on other sites

From Ben's blog:

Quote

The biggest single feature of Vulkan is its new multi-core friendly threading model. Vulkan is “externally” synchronized, which basically means applications can do whatever they want, but have to talk to different parts of the driver from different threads.

So in other words, if they code it better to use Vulkan then it means they can make better use of the CPU and stop it being such a bottleneck. I'm sure it's going to take time.. but exciting none-the-less. My understanding of it is very rudimentary, but when I see bad performance in the sim and check what is going on, it is almost always the CPU.

  • Like 1
  • Upvote 1

Share this post


Link to post
Share on other sites

Vulkan works amazing in the game doom very high frame rates and works great on low end hardware hopefully they take advantage of the tech.

  • Like 1
  • Upvote 1

Share this post


Link to post
Share on other sites
7 hours ago, jabloomf1230 said:

Isn't it Vulkan? Anyway, here's nVidia's take on it:

https://developer.nvidia.com/Vulkan

Considering that Intel can barely improve CPU performance these days, any tech that minimizes the importance of the CPU will be a welcome relief.

But hyped technology has a habit of turning into background white noise once introduced. Consider Gsync, SLI and DX12 as examples.

 

Gysnc is probably one of the best technologies to come around in the past decade.  It has no effect if your sim is running 30 fps, but for 40-120 fps it has completely eliminated screen tearing and latency.

  • Like 2

Share this post


Link to post
Share on other sites

Hi,

Vulkan/Metal are new APIs designed to remove a layer of 'abstraction' between the program and the video card.

Traditional APIs like OpenGL consists in a series of functions to tell the video card what and how to draw, then, the video card drivers takes the list of high level commands, transform these to internal low level rendering data, and process the data for rendering basically triangles with textures. Each of your CPU core has to take its turn to send its instructions to the video card.

Vulkan on the other hand consists in a series of functions to build the internal data the video card will process directly for rendering. In other words, instead of basically saying draw a triangle with a red color, you build the internal video card data needed to represent a red triangle into a 'command buffer', then send the entire buffer to the video card for nearly 'direct' rendering as-is.

The advantages are many: you directly talk to the video card in its native language, you can have each of your CPU core work in parallel to build up the command buffers. Also, since you do know what your data consists of, you can optimize the way you build up the command buffers in a way to 'pack' drawing commands using the same resources together more easily (2 triangles with the same texture). The drawback is that this puts more work to be done in the program itself.

  • Upvote 4

Share this post


Link to post
Share on other sites
Posted (edited)

I checked today on my linux box and find already that the workload is much better spread over the cpu, one core being at 100%, the other three at 50%, bringing the total cpu occupation at around 75%. The complaint used to be that the processor was used only 25%. I hope laminar is able to squeeze more out of my aging system. But I can not complain already , I have typically around 30fps and super smooth.

I am wondering what they're bringing to the table today.

Edited by jh71
Corrected Franglais

Share this post


Link to post
Share on other sites

Excellent! Can't watch it just now though  what are the main points anyone?

Chris

Share this post


Link to post
Share on other sites
Posted (edited)

Cool lots of great things coming

 

Edited by Silverbird
error

Share this post


Link to post
Share on other sites

Highlites.

 

Particles for more realistic . fires, smoke contrails, easy to edit on t he fly, Great demo including ligting of teh ground by fires and heat haze.

Improved fuselage effects by prop wash.

Improved down wash seems the fuselage effects they had where wrong and affecting this too. So hopefully less thumping onto the ground when developers implement it.

More stuff for lego bricks airports, more clutter fences etc.

11.25 is in beta and out there with new airports and a couple of better modeled us cities. Vegas and cat remember teh other one

New lights with different colours to light your buildings

Vulcan might make 11,3 but its a long road. I think thats what they said, but opengl will remain

 

Oteh rbits  I cant remeber but its getting so much better

 

  • Like 2
  • Upvote 1

Share this post


Link to post
Share on other sites

Great to see Las Vegas getting more love again from Laminar!

Share this post


Link to post
Share on other sites

Physics:

.) Better jet engine modeling (single spool and twin spools types modeled);

.) Plane Maker now estimates the Cd (coefficient of drag) of aircraft bodies. The aircraft designer can use the estimated value or input his own;

.) Modeling of prop wash ahead of propeller disc (i.e. suction);

.) Improved downwash modeling, crosschecked with NACA reports on the B707;

.) More accurate calculation of fuselage (and other bodies) side-force and lift. It is now smaller in magnitude and its application point varies dynamically with angle of attack/sideslip;

.) The new physics improvements must be opted-in with a check-box in XP settings;

 

Systems and Avionics:

.) Improved modeling of crew and passenger oxygen systems, based on physical units modeled after real equipment;

.) Improved anti-ice modeling with electric, bleed-air, chemical and pneumatic systems;

.) Improved modeling of propeller governors, with different failure modes simulated depending on the specific type of governor/engine;

.) Improved modeling of GA autopilots. Now there are position-based and rate-based autopilots, with or without electric trim;

.) Improved modeling of airliners autopilots: N1/EPR modes, CWS, REAL dual and triple channel checks (for autolandings), flare and rollout guidance (it was shown a video with the default B737 doing an autolanding and rollout with 17G25KT crosswind);

 

Vulkan:

.) The work being made to optimize the rendering engine for Vulkan, should already produce some improved performance for 11.30;

.) Vulkan will not improve multicore use per-se, but will allow them to fix multicore use;

.) Vulkan beta possible (not certain) before the end of 2018;

 

  • Like 4
  • Upvote 2

Share this post


Link to post
Share on other sites
10 minutes ago, Murmur said:

Physics:

.) Better jet engine modeling (single spool and twin spools types modeled);

.) Plane Maker now estimates the Cd (coefficient of drag) of aircraft bodies. The aircraft designer can use the estimated value or input his own;

.) Modeling of prop wash ahead of propeller disc (i.e. suction);

.) Improved downwash modeling, crosschecked with NACA reports on the B707;

.) More accurate calculation of fuselage (and other bodies) side-force and lift. It is now smaller in magnitude and its application point varies dynamically with angle of attack/sideslip;

.) The new physics improvements must be opted-in with a check-box in XP settings;

Thanks for the summary!

As an end-user, it's great to see all these improvements. But man, this has to be rough for the plane developers, especially payware that includes a commitment to support the product through the v11 cycle.

As it is, I know Carenado has been holding back the release of patch/updates on certain models like the v11 Pilatus PC-12, which has some issues that really need fixing (like an annoying sound click). They've been waiting for the current version to stabilize to release the update. Good luck with that!

I'm guessing the opt-in setting is designed to allow backwards compatibility for add-on aircraft, which gives the developers some flexibility. But of course as end users, we're all going to be clamoring for support of the new features. It won't be a tenable situation for the default XP11 aircraft to have better, more realistic flight modeling than v11 payware. It will be interesting to see how the various plane developers handle this.

  • Upvote 1

Share this post


Link to post
Share on other sites
42 minutes ago, Paraffin said:

I'm guessing the opt-in setting is designed to allow backwards compatibility for add-on aircraft, which gives the developers some flexibility. But of course as end users, we're all going to be clamoring for support of the new features. It won't be a tenable situation for the default XP11 aircraft to have better, more realistic flight modeling than v11 payware. It will be interesting to see how the various plane developers handle this.

As far as I understood, the purpose of the opt-in setting is to give time to gather data from the community (both users and developers) before making this option the default. This will ensure no existing aircraft flight model will break, while giving the necessary time for developers to adjust their models, and users to report any problem. I just hope, for the sake of helping everyone out, this option will be also plugin-controlled, otherwise it will be a hard time for users to remember to activate the option for aircraft A and deactivate for aircraft B.

  • Upvote 1

Share this post


Link to post
Share on other sites

I'll copy-and-paste this from another forum where I posted my thoughts on... might be of interest in this discussion:

 

Quote

Just got back from FSExpo 2018.  I gotta say, man, I'm excited!  

Since almost a decade ago already, I've tried to position myself as a sort of "bridge" between (then) FSX (now P3D, mainly), and X-Plane. I've done this through helping develop tools, making tutorials, converting FSX/P3D payware planes to X-Plane, and giving Laminar lots of feedback on their betas and their tools.

I've been betting for a long time already, that we'd see a movement towards X-plane, but there were things that spoke against it (such as Austin, being sidetracked by mobile app development, the subsequent Uniloc patent troll lawsuit, his movie project, etc, and on the other side, we saw the stagnation of FSX, the spectacular rise and fall of MS Flight, then the Lockheed-Martin P3D takeover, then Dovetail, Steam, and more recently, the rise and fall of FSW... all sort of indicative of an infrastructure with some pretty severe problems of their own), but at this expo, it really felt like Laminar is PRESENT, with some serious clout.  Their core team consists of 15 people, and their presentation was amazing... the room of around 1500 attendees was electric!

Even my impression of the booths and all the users that were out there was, that FSX/P3D developers are realizing that they better get a move-on to set up shop on the X-Plane sim. 

I was really curious to see PMDG's presentation, to see if they'd make some announcement of perhaps giving X-Plane 11 another shot... but I felt it wasn't only myself that was disappointed... and that's all I'll say about that.  

In speaking to P3D developers, I've had a strong impression that Lockheed Martin plays their cards close to their chest... don't give a lot of details on what's coming in future versions of the sim... but it IS obvious to me that they consider it worth trying to catch up with where X-Plane is, as X-Plane seems to always be a few steps ahead of P3D.  And what really struck me is, that P3D is NOT a community-driven sim. The sim community is something secondary to Lockheed Martin.  It definitely seems that the driving force behind the sim are more military training interests rather than the sim community's experience of the sim.  I'm not sure if I'm correct about this, but if the funding for the development of the sim comes from a 51-billion dollar military industrial complex, interested in the sim primarily for its value in training military personnel, it seems pretty detached from the market forces that could make the sim a better sim for end-users and hobbyists.  What they seem to be signalling lately, is that it's worth trying to retain the sim crowd, but in order to do that, they have to do what the industry leader (X-Plane) is doing... so it is obvious to me that P3D is not an industry leader, as the Flight Simulator franchise was under Microsoft... and that's only going to be a matter of time before that catches up to them.   But I also couldn't help but imagining Lockheed Martin not really caring that much... provided the sim is good at doing what they got it for, namely training military personnel.

In contrast, it seems that X-Plane development is really aimed at the customer's experience, and the Laminar team seemed extremely  receptive to suggestions for improvements, they were candid about their road map, and their potential road blocks, and it seems that they are continually working to improve bottlenecks to add-on development, such as documentation and tools.  (They're even previewing REAL-TIME particle tools for dev's to use!  I hope that's a glimpse of things to come, as I've been pressing Laminar for a long time to provide us with more real-time in-sim development tools!)  

A recent Google Analytics search I tried, comparing FSX/P3D to X-plane shows that X-Plane out-ranks Prepar3D, which outranks FSX by a large margin, and for the first time recently, X-Plane out-ranks P3D and FSX combined!  There's definitely a shift happening.

Now, as Thranda Design, it has been my ambition to be in a position to facilitate FSX/P3D developers to get a foothold in the X-Plane world, and I'd argue we've done a pretty good job at it so far, with Carenado, then Alabeo, then Just Flight relying on Thranda to port their products over to X-Plane... and the sense I got at the Flight Sim Expo is, that there are others who are very interested in figuring out THEIR road map to sell their products in X-Plane.  This could be a very good thing for X-plane, but it could also come with very unpredictable side-effects, such as very rapidly change the landscape of payware add-ons  in X-plane.  Hopefully for the better!

I also had a short one-on-one talk with Austin, because I was very interested in what he mentioned in his presentation, about working on a personal electric VTOL aircraft, and using X-Plane to validate the design.  I told him that I'm actually quite surprised that X-plane isn't being used much more for that sort of thing... and he agrees. It's a great tool! It's no CFD, but it has its place.  And the fact that Austin uses this sim to do real-life prototyping also forces Austin to keep improving the sim's physics, which should also give X-Plane users a sense of confidence in the underlying physics of the sim... i.e, that it is on a completely different level than the competition, which relies more on table data to get planes behave similar to the real one it's modelled after.  All in all, I'd argue that that is THE foundation that makes X-Plane the better sim, although all the recent additions (64-bit, built-in G1000 and other avionics, PBR, native VR support, particle effects, real-time development tools, etc. etc. etc.) will keep making it tougher and tougher for the competition to keep up... and there might be a lag, but I can't help but think it'll end up being a lonely place for P3D add-on developers who haven't thought hard about considering development for X-plane.  

I think time is running out for P3D dev's to make up their minds about not WHAT they should do, but HOW they want to go about doing this.

Unfortunately, X-Plane is still a severely moving platform, which is a constant source of headache for anyone trying to develop add-ons for this sim... but ya gotta choose your battles, and I must say, I felt extremely lucky in Vegas, that I've been making the "bets" I've been making in the past... with X-Plane, with my partners, with my team, and the decisions we've been making.  It's paying off, and it's becoming VERY interesting to see what'll unfold in the sim world over the next few years.  I even got the sense that Laminar is pouring significant resources into making the sim increasingly stable and even-keeled, to make it more 3rd-party developer friendly... but it takes many simultaneous layers of in-depth improvements to get there.  Judging by how their team has grown, though, I do believe they're stepping up to that challenge, and things seem to be improving quite rapidly.

So yeah... It's so exciting to be a part of this.

I'd be curious to get some feedback from end-users about what sorts of scenarios they'd consider desirable, how the market might be transformed in the coming years, etc.  What's your take?

 

  • Like 7
  • Upvote 4

Share this post


Link to post
Share on other sites

I didn't attend the conference, but I tried to watch several live streams of the event and some of the exhibitors. What I found very interesting was the live chat section as there were a lot of people asking constantly "Where is X-Plane support?" or "I want this in X-Plane" etc.. If developers aren't now embracing this, then it's probably time they should consider it.

Also... Aerofly news was very quiet.. did anyone hear any announcements apart from ORBX's Netherlands?

 

Share this post


Link to post
Share on other sites
Posted (edited)

I think this information may be beneficial for the discussion and end-user understanding of where the sim is at and where it's headed.  It relates to what was presented by Laminar at the 2018 FSExpo. 

Quote

One thing presented by the Laminar team was an "opt-in" check box for airplanes.  The idea here, is for newer versions of X-Plane (which will feature revamped and improved flight dynamics) to retain backward compatibility with older aircraft.  

This is an initiative that came out of a conversation I had with Austin last year, asking Laminar to take more decisive steps to avoid breaking add-on developers' planes with every version upgrade of the sim.  It's impossible for developers with numerous aircraft to keep up with the changes in the sim, especially as there's almost continually a beta version of the sim in the works.  Testing for release of our products sees us testing against the latest stable release, while preparing planes for future versions forces us to check them out in beta versions of the sim, although we'd never publish a final aircraft calibrated to a beta-version of the sim.  This two-track development process is more than exhausting, however, and I'd venture to say that it might be one of the major factors preventing existing FSX/P3D developers to successfully set up shop in X-plane.

Anyway, Laminar's proposed solution was, to try to retain older planes' flight dynamics in XP11 as closely as possible to the way the plane was originally authored.  So in essence, Laminar is trying to have two flight models in the same sim, one for old aircraft, and one for aircraft saved in the current version of the sim. Presumably, the new version of the sim will have a more detailed, nuanced, "high-definition" flight dynamics model that accounts for more variables... and older planes, saved without these settings explicitly authored, will have to run in compatibility mode.

Now, officially, Laminar has said that they wouldn't touch flight dynamics until XP11.30... but unofficially, we DID notice that in more recent versions of the 11.2x run, saving a plane in PlaneMaker 11.2x DID seem to cause the ground effect problems we were having in previous versions of the sim to be mitigated... but since there is absolutely no documentation on this, and since there isn't even an official admission from Laminar that any flight dynamics changes have been made to that version of the sim, we as developers are stuck in an awkward phase: we cannot revamp flight dynamics on existing planes, until XP11.30 is FINAL. 

In the mean time, we have unofficial and undocumented "opt-in" functionality into sim-wide flight physics changes that aren't even acknowledged officially. We don't even know from which point on (from which version of PlaneMaker), flight dynamics are currently locked into some sort of compatibility mode, and we have no idea what flight parameters would be affected if we even saved it in the newest version of PlaneMaker. (It's not like differences are easy to see with a file compare program either, because the way X-Plane handles these .acf files might actually be based on internal sim algorithms, as opposed to explicit settings that could be observed in the .acf files themselves... as in, X-Plane, by reading the .acf file's header, senses that a plane was saved in PM10.5, and will say, "OK, then forget about the new turbine engine model and the ground effect corrections... apply XP10's flight physics to this plane!").  This means, we will be faced with customer complaints that certain planes that we've released during this phase are not behaving as expected.

We, as payware authors, have received a lot of criticism for this... "Why do the planes slam into the runway upon landing?" "Why does competitor X or aircraft Y not have this problem?" And the truth is, we do not have a 100% clear picture on why this is... other than that there seem to be some changes that are happening to the sim, which won't be finalized or documented or fully dialed in or even acknowledged until XP11.30 has been finalized.

There are a lot of things we need to revamp on a lot of planes, to make them fly as intended in X-Plane... but as of right now, I don't think it would make a whole lot of sense to bring those planes onto our work bench and twiddle around on them, tweaking parameters which we have no clue if they'll be overridden or not, or if they will be changed in XP11.30 or not, stabbing in the dark due to lack of documentation or even a basic understanding of what X-Plane is doing.

It seems like a step in the right direction, to have an "opt-in" check box... as opposed to having the sim merely read the version number of PlaneMaker in the header of the .acf file and automatically opt in new flight physics, which are neither documented nor acknowledged officially, nor is there any sense of how permanent these would be, nor to what degree or extent these opt-in features would affect already-tweaked flight regimes.  But I do hope that Laminar will document these things sufficiently for us add-on developers to respond appropriately to these sim-wide flight dynamic developments.

Either way, there's absolutely nothing to be gained by trying to adjust any of our planes' flight dynamics during the v11.2x or even 11.30-beta version run of X-Plane, as the flight dynamics will likely have to be re-done from the ground up anyway, and they're quite likely to still be changed before then.  And it'd be nice to have a bit more light shed on all of this, before we go making such significant revamps.

It looks promising that we'll get more info and documentation as XP11.30 enters into its Beta phase... but again, this complicates matters a LOT for developers, as Laminar will be relying heavily on guys like us to give well-researched, concise feedback on these changes to the sim. In the past, if we DIDN'T react to the Beta phase (normally because we find ourselves developing on stable non-Beta versions of the sim), we get the blame for not having given Laminar the appropriate amount of feedback when the sim was still in Beta. 

This is a real problem, and it really cuts into our dev time, making things a lot more inefficient than we'd like them to be.  

At this point, I don't know what I'd like to see in the future... on the one hand, it's a good thing that Laminar is responding to our need for them not to constantly break already published planes with their updates. On the other hand, the piecemeal, undocumented implementation of this has resulted in a massive amount of chaos on our end, and a difficulty in conveying to end-users what's happening under the hood.  I hope all this stabilizes soon.

 

Edited by danklaue
  • Upvote 2

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now