X-Plane V11.30 Komt eraan deze week!, Tijdens de FlightSimExpo afgelopen weekend kwamen de mensen van Laminar Research, oftewel de makers van X-Plane V11 met de volgende mededeling:
- Nothing new on this front, except the fact that 11.30 should be out this week.
With it, some minor adjustments to helicopters are coming. Well, one at least, according to the leaked changelog: “Now we find ground effect for each BIT of the helo rotor! This phases in the ground effect from front to back as we approach a tall helipad!”
What this means is that when you are approaching a helipad, for example, on the side of a mountain, you will get IGE (In-Ground-Effect) partially as you get the helicopter on top of the said pad.
This will present a bigger challenge for all of us when using these places to land. It’s on 11.30 and, again, it should be out this week.
New Autogen and Library
Alex also showed Petr’s new industrial autogen work, which will ship in X-Plane 11.30. The new autogen is an art upgrade that makes the existing scenery tiles with industrial areas look a lot better.
As part of his autogen work, Petr has been restructuring and organizing the scenery library, which should be a welcome improvement for gateway artists and third party developers. Over the years, the library has become disorganized and cluttered; Petr is making sure that useful art assets are in the library for reuse and are labeled in a sane way that makes them straight forward to find.
New York and Washington DC
Alex also previewed our new landmarks for New York City and Washington DC. Austin and I are particularly excited about New York City – having the real bridges modeled is going to look fantastic.
One note about the presentation: Alex incorrectly stated that NYC and DC would ship in X-Plane 11.30. They will almost certainly not ship in X-Plane 11.30. We’ve been shipping the highest profile landmark scenery packs with our smaller updates between the major patches. Since 11.30 is a major patch with a lot of code changes in it, these landmark packs will wait and go into a smaller patch.
X-Plane 11.30 and Experimental Physics
Austin showed some of the new physics tech he was working on – he has been working on prop wash and body lift forces in a lot of detail.
Blade theory does a great job of predicting how an aircraft will fly, but at its heart it is a decompositional technique – we break the aircraft into parts and then let you fly the sum of the parts. Where this becomes difficult is where there are interactions between the parts. So it’s no surprise that after almost thirty years, the area where Austin can still add better fidelity to the flight model is all in multi-part interactions:
- Prop wash – as air is accelerated through the prop, what else does it interact with? The more accurate this air mass acceleration is, the more accurate the effect on other parts of the aircraft.
- Down-wash and ground effect. When the down-wash from the wings interacts with the ground, what happens?
Austin also showed his custom test rig to measure lifting body forces. Lifting bodies work reasonably well in X-Plane now, but a better model means more accurate flight over a wider variety of regimes.
Flight Model Research Mode
X-Plane 11.30 will also introduce research mode. Research mode enables these new physics; with it disabled you get the 11.20 flight model. The basic idea is: it takes a lot of experimenting and testing to validate new physics algorithms. Until now, we’ve had to try to do this during open betas. But a beta program is relatively short (maybe 8-10 weeks) and during the beta, other parts of the simulator are broken. Aircraft authors often don’t have time to try the new physics at all, or can’t use the beta due to other bugs. As a result, we get almost no feedback during betas about flight model changes.
Research mode lets us have the next-gen physics code optionally available to everyone over a much longer time. You can think of all of 11.30 as an “open beta” of the next-gen physics. Developers who are doing real aircraft R&D can get our very best math, and authors who are trying to sort out flight model problems with Austin can use his latest tech.
Research mode is not something you set on an aircraft, it’s something you set in X-Plane. The reason for this is that research mode isn’t meant to be a stable physics model for shipping payware aircraft; it’s meant to be in-flux and evolving. We’ll make real versions out of the code that was “research mode” once we reach a point of stability.
At some point once the new physics are heavily tested, debugged, and stable, we’ll make an ‘official’ version of the flight model, and all aircraft saved in Plane-Maker will get the new physics. Aircraft saved in prior versions of X-Plane will use the compatibility code path.
Systems and Avionics
Philipp presented a number of new systems enhancements, all of which are scheduled to be part of X-Plane 11.30.
- Crew and Passenger oxygen systems.
- Anti-ice (with several different actuation mechanisms).
- Prop governors (including start locks).
- New autopilots.
One common thread of all of these new features is an increased level of real world accuracy. Philipp based this work on experience with specific real-world aircraft, and the setup for these features is based on the real world system you are simulating. Units are real world (e.g. liters of anti-ice fluid) and behaviors match real-world devices. One thing this means for third party authors: you should use these features by describing your aircraft, rather than trying to coax X-Plane into giving you an expected result.
Besides new GA autopilots (capable of accurately simulating the less sophisticated behavior that less expensive non-GPS GA planes have), the airliner autopilot has several upgrades, including full coupled auto-land. Philipp showed the 737 auto-landing under pretty serious cross-wind gusting conditions, and I can safely say that the new autopilot flies a lot better than I do.
Particle Effects System
I showed a few videos of the new particle effects system, which we are targeting to ship in X-Plane 11.30. Particle effects create smoke, fire, mist and other ephemeral effects in the sim; until recently it has not been possible to customize them.
The new particle system effects editor provides an in-app graphical editor to let authors create and customize effects in real-time. The particle effects system is entirely artist-driven; artists can make any effect and attach it anywhere on the aircraft. Effects are driven by graphical key-frame tables and datarefs (which can come from plugins). The whole system is extremely flexible.
Besides standard billboarded particles the effects system can generate “heat blur” particles that distort what is behind them and light emitting particles. The demo showed an engine fire that lit up the ground and blurred the area above it.
To answer a question that came up a lot after the show: particle effects can be attached to objects on aircraft and created by the XPLMInstancing API. We are working to allow particle effects in scenery objects, but it’s possible that will come after 11.30 ships.
I presented a progress report on our port to Vulkan. For the last year, Sidney and I have been rewriting the rendering engine to run on Vulkan and Metal as well as OpenGL. This is going to get us faster performance, reduce stuttering, and enable us to run the engine on multiple cores for better hardware utilization.
After a year of banging on the code, we’re at what is roughly perhaps a half-way point. We have several parts of the underlying code ported to Vulkan, and we can run Airfoil-Maker natively on Vulkan on Windows. While Airfoil-Maker on Vulkan isn’t useful to anyone and it’s a far cry from all of X-Plane, it’s a useful check-point for us, demonstrating that the code we’ve written so far is on the right track. It’s still too soon to say what performance will be like under Vulkan, but the Vulkan implementation of Airfoil-Maker makes 69% less driver calls than the OpenGL one; we thought that was a good sign.
We don’t have a release date or version for Vulkan other than to say that we will ship Vulkan in the v11 run and it won’t be in 11.30. We are hoping to have a beta this year, but that’s a hope, not a hard date.
To answer a question we’ve received since the talk, we are also working on a Metal port for OS X. The Vulkan and Metal port are not separate efforts; we are making a single engine that can use either graphics API as a back-end. We don’t have Airfoil-Maker running on Metal yet but I am hopeful that we get there soon.