Reflecting on the future of flight simulation

[oktorn777 0]

To put my reflections into perspective it might be useful to introduce myself. I have been out of the loop for many years now when it comes to flight simulation, but I still occasionally follow the news on simulation related stuff. I did a lot of flight simming years ago during my youth and I always was fascinated by the idea of recreating a realistic virtual world. For a long time, flight simulators where the only consumer programs out there which attempted to recreate such a believable virtual world. Professionally I am working as an atmospheric scientist and therefore got quite a weak spot when it comes to depicting that particular part of the simulated world. :-) Further, I used to have a PPL licence when I was a university student and I am still actively involved in paragliding.When I was a kid we had this apple computer and my dad one day brought a flight simulator home for it. I cannot remember the name, but it had black and white, wire frame graphics. The world was one big, black, flat square with an occasional white line representing a landing strip. The world was surrounded by four black walls with outlines of mountains drawn across them. Still, I was completely fascinated by that thing having never seen anything like it before. From a very young age I was fond of airplanes so that definitely helped keeping me interested in flight simulation. I spent too many hours in that sim trying to escape that box world by trying to escape by flying over the cardboard mountain tops, convinced there must be something behind them!Looking back from where we are today flight simulation has come quite a long way. Flight simulators are now multi gigabytes in size, have real world data driven graphics and and the airplanes have much more realistic flight characteristics.So the question arises if there can be still significant improvements be made to existing home use simulators like FSX.Which areas have still the biggest shortcomings?I will only focus on the world as a physical and visual entity for the purpose of this post. The visual aspect of a virtual world is much more connected to physics than most people probably realize. The light from the sun for example has a certain spectrum and this light hits the atmosphere in which different scattering (mainly layleigh and mie scattering), absorption and reflection processes are taking place . The scattering of sun light in the atmosphere depends on many factors such as the composition of the atmosphere, suspended particles, water vapor content, inclination of the sun to name a few. This physical processes have a direct effect on the sky coloration, the light spectrum and intensity actually reaching the earths surface illumination the surface and objects. So physics are not only critical to accurately calculate the flight path and movements of an aircraft, but are equally important when it comes to rendering a believable world.So for example, how does a flight simulator like FSX simulate the physical processes which determine how the sky looks? Well it turns out it actually doesn't simulate anything at all. FSX simply uses a couple of coloured bitmaps to draw the sky around you depending on the day of time.Looking at the MS Flight screenshots I have a hard time to hide my disappointment.And here is why:- The overal rendering (lighting, terrain, clouds, autogen, airplanes, instruments) looks nearly identical to FSX apart from the reworked shadows and new textures.- The cloud rendering looks identical to the system introduced with FS2004 in 2003 (see this paper by Niniane Wang for more details: http://ofb.net/~niniane/clouds-jgt.pdf). Basically clouds are a pile of 2d sprites arranged in 3d boxes which then are color shaded based on vertical color levels made up by artists. (see this paper by A. Bouthors et al. for a different approach on real-time cloud rendering: http://www-evasion.imag.fr/Publications/2006/BNL06/)- Looking at the terrain rendering the same issues as in FSX can be see: "soft rounded" look to mountains and blurries in the distance. (see this video showing the Outerra engine using fractal algorithms to refine and render real world elevation data (they got some a physics based atmospheric rendering in there aswell).There are many more things that are worth noting, but I think this will do for now. Clearly the MS Flight team is not putting in much effort to move the rendering away from an artist created to a physics driven system. For me this missed opportunity is a major disappointment. It feels like the team has either not enough resources, they have moved their focus to different topics or they just lost the "bite" to truely innovate and push the genre forward.

[daru1 0]

You're about to get a lot of nay-sayers to your post, as there are a lot of people who think that Microsoft can do no wrong when it comes to Flight. I, for one, am not one of those. You are dead on in your observation - and you back it up with sources, which is amazing. While you are not going to sway the opinions of these nay-sayers at least you have provided some with an open mind something to think about. Great post.

[bonchie 2,305]

Developers are married to their tried and true methods. Logically, they seek to improve upon them but the basis remains the same.It'll probably take another developer not yet in the game to go in a completely different direction with something like the Outerra engine. Until then the profit potential of flight simulators (and PC games in general) are likely not going to allow current developers to drop everything and start over, not just from scratch, but from a completely different starting point.

[mgh 89]

Clearly the MS Flight team is not putting in much effort to move the rendering away from an artist created to a physics driven system. For me this missed opportunity is a major disappointment. It feels like the team has either not enough resources, they have moved their focus to different topics or they just lost the "bite" to truely innovate and push the genre forward.

Your conclusion is wrong because the whole permise of your post is wrong.Microsoft is developing Flight to meet the needs of a much wider market which frankly doesn't care about such points as "light from the sun for example has a certain spectrum and this light hits the atmosphere in which different scattering (mainly layleigh and mie scattering), absorption and reflection processes are taking place . The scattering of sun light in the atmosphere depends on many factors such as the composition of the atmosphere, suspended particles, water vapor content, inclination of the sun to name a few"Yes, of course Microsoft could spend more money on including these, and the other features you mention, but it just wouldn't be worth while financially. Microsoft is a commercial company, like any other, and is only developing Flight to make money - not to satisfy the desires of enthusiasts.

Developers are married to their tried and true methods. Logically, they seek to improve upon them but the basis remains the same.It'll probably take another developer not yet in the game to go in a completely different direction with something like the Outerra engine. Until then the profit potential of flight simulators (and PC games in general) are likely not going to allow current developers to drop everything and start over, not just from scratch, but from a completely different starting point.

There's nothing to stop someone else developing a flight simulator for enthusiasts but no one has - and I doubt anyone will as long as Microsoft does - remember Aerosoft Flight Simulator 2012?.

[Glynn 405]

There's nothing to stop someone else developing a flight simulator for enthusiasts but no one has - and I doubt anyone will as long as Microsoft does - remember Aerosoft Flight Simulator 2012?.

HelloThere is onehttp://forum.avsim.net/topic/353630-tons-of-new-screenshots/

[mgh 89]

HelloThere is onehttp://forum.avsim.n...ew-screenshots/

That is true, but X-Plane has been around since 1992. I don't still don't see anyone else starting development now.

[RobbieHe 0]

Give me FSX with slight visual improvements but double the performance per given CPU and call it what you want (MS Flight), and I will be happy! What are the chances of that, I wonder?RH

[simmerhead 885]

It would be nice to have more accurate weather simulation, but I think the amount of compouter power it would require would make it impossible to run with current hardware. Meteorological calculations in the real world are extremely CPU intensive and are done by the worlds most powerful computers. I higly doubt we'll see a physics based model for that anytime soon in a consumer flight simulator. I much rather see that computer power spent on better and more realistic scenery.

[CrewF16 26]

The one thing that no one seems to be mentioning is the fact that there is still room in FSX. I do not understand the technology behind designing the simulation, but I do know that it takes a killer rig to honestly run FSX maxed out. Using real world weather and all. I showed my wife some video's of Flight and she just turned around and said that it looks like mine. I got to thinking about it and maybe they have figured out how to do it without using all of the resourses nesecary now to run FSX at Max. I am not saying that there are not improvments because you can see them. But I understand what (Oktorn) is saying about the engine behind the sim.

[oktorn777 0]

@mghI think I haven't made it really clear what I wanted to convey with my original post.First of all I wanted to show that Flight is not using a new rendering engine which becomes apparent when comparing the available screenshots to FSX or FS2004 or even FS2002 (which all basically use the same rendering engine for the terrain, apart from the added support for higher resolution textures and DEM data).I strongly believe that the visuals of "realistic" games can only become truly "photorealistic" for the lack of a better word if developers actually start to simulate the physics behind the visuals of the real world.The huge benefit of this approach is that you do not need artists anymore to generate all the visuals, but the physics and data driven system generates them automatically and in a realistic manner. Products like REX only exist because of the shortcomings of the artist generated visuals of flight simulator. Once you have the physics in the engine artists can move on to generate other items of the world rather than spending time on atmospheric effects, clouds and water.It would benefit all users because everyone, no matter of the actually understanding of the underlying processes, is very good at judging what looks realistic and behaves realistically, because everyone is seeing and interacting in the physical world every day.To summarize, visuals driven by physics make things look realistic without the need of artists spending countless hours trying to recreate effects and visuals artificially.Ultimately, it is the only way forward for the entire game industry, at least for the industry which tries to make real looking games. I am confident that 5 to 10 years from now all major rendering engines for games will have moved to a much more physics dependent approach to generate the visuals.I understand that companies like to make money. But I also do firmly believe that you can only be truly excellent at what you do when got a genuine passion for it. If you got passion you will be innovative and ultimately come up with a superior product which in the end will sell better too. Where I am coming from we have a very simple tenet, either you give something everything you have or you don't do it at all.If you are interested, watch this video of Outerra to see that such a system is actually feasible to run at decent framerates (20+) on todays hardware: http://vimeo.com/31560308PS: Here is the paper I actually wanted to linked about real-time cloud rendering by A. Bouthors et al. (2008): http://www-evasion.imag.fr/Publications/2008/BNMBC08/clouds.pdf

[simmerhead 885]

If you are interested, watch this video of Outerra to see that such a system is actually feasible to run at decent framerates (20+) on todays hardware:

Outerra is not a simulator. Add weather, clouds, flight dynamics, AI aircraft/boats/cars, ATC, more variations of ground textures and 1000s of scenery objects and compromises have to be made...

[Oliver Ooi 85]

I am basically ok with the FSX cloud rendering (with rex texture of course), I am not sure the same technology will be used in Flight, but I just hope they should at least add cloud shadow in the new sim.

[PebbleBeach 120]

As a stock holder of MS shares. My concern is my investment and dividends, as with most of us top 1%ers. I demand this "game" be for gameboxes as that is where my investment would tend to be the most lucrative. Millions compared to tens of thousands users. Sorry hobbiest.A scenerio of course

[CrewF16 26]

I take it that Outerra is an Add On. It looks great but I would like more info. What Rig is running it? What would happen if you started the Queen at Seattle Tacoma Int. Ect.Thanks,

[simmerhead 885]

Outerra is not an addon or even a flight simulator. It is a planet rendering software, kinda like MIcrosoft's ESP.