Ultimate Game Engines for Flight Simulators Discussion Thread

Flight Sims
1

The dream flight sim to me would look amazing at terrain/landscape level and in the air over the entire globe.

Here are some amazing “Game Engines” to drool for flight sims…

  1. Outerra - http://www.outerra.com/
  2. Unigine - https://unigine.com/
  3. Unity 3D - https://unity3d.com/
  4. Unreal Engine 4 - Unreal Engine 5 - Unreal Engine
  5. Cryengine - https://www.cryengine.com/
  6. Havok - https://www.havok.com/
  7. Frostbite - Frostbite Engine - The most adopted platform for game development - EA

Are but some of my favorites my dream is a full world and outer-space all encompassing simulator (tired of the hop around :money_mouth_face:)

As to why flight simulator developers use a custom game engine instead of adapting an engine such as Unreal Engine 4? Here are some thoughts… https://www.quora.com/Why-do-flight-simulator-developers-use-a-custom-game-engine-instead-of-adapting-an-engine-such-as-Unreal-Engine-4

Please offer any valuable constructive discussion points as they are all welcomed?

2

A couple already do have pseudo-simulators on them; Project Wingman, for example, runs on Unreal (ignoring that it’s more of an arcade game, but a lot of the same principles apply).

In the past, the usual limitations were typically the physics engines and world space. Some engines also had difficulty with joystick support (I think it was Unity). When you put them all together, and coupled with the fact that sims are niche and expensive, you end up with folks like ED taking their core engines and just transitioning chunks and pieces over time. If the physics core works well enough, there’s not much reason to ditch it all just because the graphics engine is 5 years out of date.

Another thing that is seldom considered is the passage of time and how it affects lighting. A lot of the above engines are designed for one time of day at loading of the map and can’t be changed dynamically (though there are workarounds, they are not incorporated by default). I’m not sure if the latest Unreal engines work around this yet, but a couple years back there were several hacks to get time of day and the resulting lighting and shading changes into the engine that were imperfect.

Lastly, physics are often the real troublesome bit as most of the above engines use simplified physics for a lot of things (which makes sense). When you try to incorporate a flight engine – which is pretty demanding – it has a tendency to bog the engine down. I recall reading a couple threads about that on Unreal development where they ran into some problems of getting the flight engine updating as quickly as it needed to be.

Being fair, a lot of these issues are typical of most game development, but they take time and money to solve, and games aren’t real cheap to develop to begin with. Flight sims are already a very niche market and if a new developer is trying to compete with ED, 777, etc. they’re not liable to get dollars if the rivet counters aren’t made absolutely happy. Likewise, if you can’t capture the average casual gamer, you’re not going to do to well there either. It’s a very rough tightrope to walk and that’s why you get things like ED pushing the bug module – a well-known and popular aircraft is a great way to get some sales numbers up. Flip side of that is you have to get a team of lawyers in place to make sure your license is good to go.

At the end of the day, it’s not impossible (I still keep an eye on Unreal every now and again for a potential Longbow 3 project), but it’s not an easy undertaking either. Hence, most folks with the skills and ability to make games will do simple stuff to pay the bills.

P.S. Check out Empyrion or No Man’s Sky if you’re looking for the all encompassing simulator.

2 Likes
3

This.

1 Like
4

Yeah of course, flight physics and atmospheric pressure so on are components to the code-base. But from all that I have read there are many programs available as 3rd party licensed and open sourced plug-in’s that are allowed to work independently or written in to the file structure.

For example - JBSSIM here - http://jsbsim.sourceforge.net/

I’m no dev but seems many guys have tried to solve problems where there are needs with places like SourceForge and GitHub available solutions are found eiher way it would be awesome for future dream sim.:yum:

5

Because it’s not so simple to just pipe in a flight physics engine into the game engine; there’s usually some compatibility stuff that has to get sorted out, per my example above where there were issues getting the game to update frequently enough with the physics engine. Small nagging issues like that become difficult to solve, especially in a efficient and timely matter.

Once again, that’s not the exclusive problem – sims are very, very, very niche nowadays and people that buy more than one are a rarity. You’ll get some that will buy DCS, some that will by XPlane, some that will buy FSX, and some that will buy IL2. The number that will buy all of those is even narrower – like me, as I don’t have much interest in general aviation, so XPlane and FSX don’t tickle my fancy. So if I was to go out and make a general aviation sim, I’ve got to compete with the old standbys that most of the community is going to hold my product up against. What can I do to make those better and can I afford to spend the time and resources doing that? That’s the real question.

6

I keep telling everybody that I’m a rare breed! :wink:

7

TLDR; The amount of work going into modifying those engines negates a sizable enough portion of the advantage of using one of these.

Heck a certain space sim tried using a legal defense that they had modified the licensed engine so heavily that there was essentially nothing left of the original :smiley:

The point of most of these is quick proto-typing and while you might have something to look at far earlier, in the end it might not save any work, and possibly create more in the long run.

All that said, it would be great though if there were some sort of open source flight sim specific engine out there.

8

Yeah, that’s looking grim…

9

One must consider the scale of the environment that is to be recreated.

For instance, here is a screen capture of me flying a Super Cub from Big Bear down to San Diego. I’m using orthophoto scenery derived from satellite images of the area, and the weather is being represented as what were the actual meteorological conditions at the time of my flight. I’m pretty sure that all of of those geographic features below are accurate at least to the scale of the ortho images and elevation data provided. And it has to do this for the majority of the earth, your hard drive space being the primary limitation for that task. How well can any of those game engines present the earth as it actually exists and to what scale?

13
131092×457 352 KB

2 Likes
10

@BeachAV8R pointed out in another thread how actual flight visuals are, more often than not, dull. And its a good thing too because it is that hazy, washed-out norm that makes the extraordinary, extraordinary. Warm, saturated colors with ultra-sharp features and skies full of god-rays look amazing. But they compare with my reality about 4 times a year. For 30 years, I wished against hope for flight sims to match reality. I first began to see that dream come true in Falcon BMS (of all places.) Now, with X-plane and to a lesser extent, DCS, the visual dream had been met. From here on I will continue to appreciate any improvement. But my suspension of disbelief is already complete.

3 Likes
11

None of the FPS engine for sure. Except for one old exception.
Project I.G.I.!

12

Some really good points so far, good discussion.

However my opinion and dream would be to look towards future modern graphic engine implementation whether to choose a market type or custom build one using more modern tools, techniques etcetera.

Now to make the point even clearer “PC Flight Simulators” of course are not your general AAA video game genre that are developed annually with releases every year and with updates to that release fixing bugs and platform issues. They are obviously more complex and resource demanding that is a given if you know about Flight sims.

Eventually though someone will have to break through with improved modern graphic engines to draw a new crowd of flight sim enthusiast you can only keep doing the same thing and making adjustments for so long.

So back on point, as an example FSX-P3D-XPlane normally look the same once you move into the 4K scenery my point was to have improved modern graphic engines and why do they still look so outdated.

As an example to “Compare and Contrast” here is three videos one is FSX in 4K the other UNIGINE and a P3D example

FSX 4K

P3D

Compare and Contrast to UNIGINE

Another awesome UNIGINE example

To me FSX-P3D- XPlane all look so outdated that leaves many like myself wanting more in comparison to what we could have. We could also have better improved crash collision model physics like this … Revolutionary soft-body physics in CryEngine3 - YouTube but a dream it remains I guess.

Another interesting read a link I found researching the topic was this online post… https://www.gamedev.net/forums/topic/664893-flight-sim-game-what-engine-or-renderer/

1 Like
13

That UNIGINE video of the aircraft kind of shows the problem. Beautiful graphics and an aircraft defying the laws of physics.

But, wouldn’t it be possible to have a physics and systems modelling engine, simulating the aircraft, and one of these beautiful graphics engines to show just the graphics.
The physics engine just updating the position in the graphics engine?

14

That’s actually what happens.

1 Like
15

Yes, I guess it is. But wouldn’t it be possible for almost any flightsim producer to just make the aircraft simulator, and then buy off the shelf world engines, so to speak?

16

A video game has a Master Source Code-base where from a tree file structure forms, hence the “Graphics Engine”.

As an example the Master File structure would incorporate the following -

• Master Source Code programming – (or intent purpose of the program instruction that reads and writes to the below items……

• Graphics Engine
• Sound Engine
• Physics Modeling
• Weather Engine
• Particle Effects Engine
• Multiplayer Code
• Feature Code
• Mission Editor or Campaign Engine
• 3rd Party Apps and Tools Plug-ins coding access

Etcetera are but some of the components in the file structure that are developed and upgraded by developers.

Modern programming languages are:-

  1. C#, C+ or C++ …… the most common
  2. Python
  3. JavaScript
  4. Ruby
  5. PHP
  6. Objective-C
  7. SQL

A program file structure will look similar to this as an example -

git-model%402x
git-model%402x614×814 86.9 KB

Where when the time comes a programmer will update the “Graphics Engine”, then all the modelers and artist have to go in like crazy to update their air craft models and scenery so on. I hope that helps the discussion somewhat.

@TROLL, YES they can buy a licenced Copy of the Graphics Engine SDK (Software Developer Kit) and then build it further from the original basic shell to an advanced monster a world globe and dynamic campaign engine with many theaters.:wink:

As an example here : https://store.unity.com/

17

Obivously it could.
I think the answer you are looking for i better if furnished with an explanation of how I’ve seen sims developed.

I’d love to chat about it any night, Discord or Teamspeak?

18

To be honest, P3D looks more realistic to me than UNIGINE. UNIGINE looks more like a game while P3D looks more like what I’d expect from the real world. In general, most game engines tend to use a lot more color saturation than we get in the real world, simply because showing the contrast is necessary for most of what they’re doing. Once again, note that a license for UNIGINE is $1500, which is just for the entertainment license alone. That’s a pretty large chunk of change for an already limited development – and it doesn’t include things like licensing real-world equipment. How many people are going to rush to play, say, a simulator that uses completely fictional designs? TKOH tried and look how it turned out.

The physics modeling is where a lot of work has to go and anyone starting out from scratch is going to be up against existing products. It’s really hard to describe just how much energy and time has to go into this part for just a flight sim. If you’re willing to compromise on some of that, you might be able to get somewhere but then you get into contested territory of game vs sim and the potential to alienate both audiences. Not good when you’re sinking a lot of money and time into a product.

19

I had written a long detailed post about engine implementation in the gaming industry (I worked on the console side with a AAA dev house that did it’s own production). Then I decided to keep it simple, and to the point.

Most of the "visually amazing’ engines, do that and only that well. Graphics and rendering, period. You pay through the nose for them, and then you still have put your own physics, flight system modeling, etc on top of it. You have to hire coders and designers with a certain skill set (ie in the engine your using), which usually comes at a premium if you’re using something that is well known. Flight sim companies aren’t big money, and the big name engines are big money. I’ll guarantee ED’s total commercial sales (and I’m talking sales of all the modules together with sales to the 3rd party companies add in) are less then what EA brings in every year from Madden, or Activision with Call of Duty. I’d bet Angry Birds has more profit than all DCS products combined.

Until someone bothers to create an engine that has bomb graphics, AND the incredibly detailed physics that flight SIMULATION demands, it’s probably a non starter. Whats the markert to have an incredible visual engine and super realistic physics engine? Basically racing and flight. That’s it. Super small market compared to the incredible visuals and moderate physics market that is 99.9% of the gaming market.

5 Likes
20

Yes, that is the problem in a nutshell the numbers are dwindling in comparison to AAA games and why the Flight Sim world needs to act now by actually bothering as you state and do something to capture the wider video game market again or else when everyone over 40 dies out so will sims on a cost base issue. Don’t kid yourselves we are not getting pro flight sims but some what close to the pro training sims used to train pilots.

With some clever game developer tactics to bridge the market it can be achieved but I’m not going to post a business plan proposal, I’m just a another voice, sadly addressing the issue. DCS has a good base for this but cost is another and it continues.

Eventually someone will do it, just the right time and when the financial market is up again.