This is what our flight sims could look like

[gizmosellsbunnys 106]

Building the next flight simulator will cost more than building fly over ramps for freeway in major city. A core engeine would cost like 100 million at least. These concepts are nice but they are not practical. The cost to build FSX was into the millions and if development of FSX happened in 2008 instead of 2005 the last flight simulator for Microsoft would be flight simulator 2004. They tried to kill flight simulator in after flight simulator 95, but it was Bill Gates as CEO that prevented it from occurring. These videos look nice as concepts but technically and financially unfeasible.

[HiFlyer 9,343]

These tech demos are great, but what happens to all that visual fidelity when you have to also calculate physics, fluid dynamics, avionics logic, and all those other under-the-hood details involved in simulating flight?

 

If we demand the fidelity of a multimillion dollar simulator that requires multiple linked computers to run on our desktops, we are going to have problems. Period.

 

To the extent that we do succeed in cramming some facsimile of of that level of detail into our systems, yes we usually have to give up some visual complexity, which unfortunately leads to the loss of interest of a large portion of potential users with no ambition to be real pilots.......

 

Thus robbing the project of even the potential of the financial support necessary to even initiate it...........

 

And here we are. :unsure:

[tttocs 301]

To the extent that we do succeed in cramming some facsimile of of that level of detail into our systems, yes we usually have to give up some visual complexity, which unfortunately leads to the loss of interest of a large portion of potential users with no ambition to be real pilots.......

 

Uhh...

 

OK. First off, I think there's some exaggeration going on here to make your case. What "level of detail" are you referring to and why would it take "multiple linked computers..."? What we have already does a credible (not perfect, but credible) job with aviation simulator details, but doesn't offload enough to the GPU for pure graphics, leaving us fighting for CPU resources. What I see in this demo is impressive in that it DOES do great graphics leaving a lot of CPU headroom. That's precisely what DOES impress me about the demo, and is precisely what we need.

 

I know what I want, but I'm not sure what it is that you want. If it's a game for the masses, those exist in multitude.

 

 

What's interesting about the Valley demo is that with my modest card, it averages about 24-25 fps and peaks at about 37 fps in some locations.

 

Exactly! Efficient graphics that take advantage of modern GPU capabilities, even on modest cards. My current system is quite CPU limited, but good on the graphics end and the Valley demo absolutely rocks turned all the way up, even with fairly aggressive anti-aliasing.

 

Scott

[HiFlyer 9,343]

Uhh...

 

OK. First off, I think there's some exaggeration going on here to make your case. What "level of detail" are you referring to and why would it take "multiple linked computers..."? What we have already does a credible (not perfect, but credible) job with aviation simulator details, but doesn't offload enough to the GPU for pure graphics, leaving us fighting for CPU resources. What I see in this demo is impressive in that it DOES do great graphics leaving a lot of CPU headroom. That's precisely what DOES impress me about the demo, and is precisely what we need.

 

I know what I want, but I'm not sure what it is that you want. If it's a game for the masses, those exist in multitude.

 

Scott

 

There have been lots of threads in this forum and in others that spell out the level of detail expected, and especially when at the edge of the envelope, that kind of realism has a high computational cost. A cost that its clear from statements such as "Its not a highway simulator/Its not a tree simulator, what about the FM" etc that some are reluctant or unwilling to pay.

 

Given that current GPU'S are multipurpose computationally, it seems clear that while they can provide high-end graphics, they can also be diverted to even more detailed calculation of things like wind and weather modeling. In other words, alternative ways of using extra resources, and historically the sim community tends to opt for calculation and "realism" over "eye candy"

 

But that "eye candy" tends to attract the "masses" whose financial resources fund the development of a detailed sim in the first place. If we are happy being not for the masses, then by all means we should continue on the current path, but then Its hard to see from whence the resources to fund the ultimate simulator often mused about comes from.

 

As for multiple linked computers, I am thinking of the tendency to expect flight models comparable to that of a full fledged Boeing simulator. And those don't run on single PC's. (and tend to have stinky graphics)

[goates 576]

Given that current GPU'S are multipurpose computationally, it seems clear that while they can provide high-end graphics, they can also be diverted to even more detailed calculation of things like wind and weather modeling. In other words, alternative ways of using extra resources, and historically the sim community tends to opt for calculation and "realism" over "eye candy"

 

Toss two GPUs into a system and use one for graphics and one for CUDA/OpneCL/DirectCompute. Also, the number of CPU cores keeps increasing, essentially creating a system today that would be the equivalent of multiple systems linked together 5-10 years ago (or even less in some cases). Not to mention things like Intel including GPUs into their CPUs, providing another potential source of processing power. Basically there is lots of processing capability available today to provide both great looking graphics and a solid sim underneath. The real limitation is money.

[HiFlyer 9,343]

Whoever pulls together the funding for the next big civil aviation simulator is going to have to be very creative.......

 

It strains the brain a bit to try and imagine just what could simultaneously be financially feasible, commercially viable and also acceptable to the core community.

 

A thread asking for what we specifically (and can realistically and not mutually exclusively expect) in next gen tech from a potential new sim might be interesting........

[HiFlyer 9,343]

Toss two GPUs into a system and use one for graphics and one for CUDA/OpneCL/DirectCompute. Also, the number of CPU cores keeps increasing, essentially creating a system today that would be the equivalent of multiple systems linked together 5-10 years ago (or even less in some cases). Not to mention things like Intel including GPUs into their CPUs, providing another potential source of processing power. Basically there is lots of processing capability available today to provide both great looking graphics and a solid sim underneath. The real limitation is money.

 

Money in more ways than one. Already an apparently huge percentage of us appear to be pushing at the edges of technology to draw just a few extra frames from current sims, and the computers we are packing are far from the average. In fact they probably tend to put us even further outside the "norm"

 

A sim aimed at specifications that lofty seems unlikely (to me) at the very least until the type of system you describe becomes more common. Even current "AAA" titles don't aim that high, its just too small a target market. I can't think of much that does take full advantage of even dual cores yet. Quad cores and multiple Gpu's? Hexacores?

 

I would push the arrival date for that out several years before even the most successful games go there with any regularity, and civilian flight sims are several years behind today's best so they will get there even later if ever; unless they find some way to become more commercially viable and get a stronger revenue stream to fund development.......

 

Which means finding a way to attract the masses.

[goates 576]

 

Money in more ways than one. Already an apparently huge percentage of us appear to be pushing at the edges of technology to draw just a few extra frames from current sims' date=' and the computers we are packing are far from the average. In fact they probably tend to put us even further outside the "norm"[/quote']

 

I would argue it's more that we're pushing the limits of current sims, not current computer technology. FSX leaves plenty of computing power on the table because it simply can't utilize it. Leaving us to tweak the sim beyond what was probably expected.

 

As for games and modern hardware, they are getting better all the time. Even consoles with their multiple CPU or core designs are helping developers figure out ways to make better use of it all.

 

[LeeL 272]

Didn't Rob say in his last interview that prices he had seen being talked about to create a new flightsim weren't accurate? Or am I mistaken? I doubt it would take 100 million to make a flightsim. If it cost that much even Bill Gates wouldn't have continued making it.

[John_Cillis 2,406]

I think looking at apps like Valley and Outerra, one can see the potential for future sims if someone with the funding can expand on what they do. But it's a challenge to find the talent out there that understands how to model the physics and flight dynamics of aircraft. And there's the added challenge of finding skilled 3-d aircraft modellers, painters, and gauge designers.

 

Back to this thread, Outerra runs so amazingly fast and smooth on my system and Valley runs really well too. I'm so hooked on looking at Valley that I did not run MSFS at all until late this afternoon. Outerra doesn't move me to the extent that Valley does but it does "wow" me with its performance.

 

On an unrelated note, since a just week ago I was flying home from Orlando, I did not get time to touch MSFS until today. I run FSX with Tileproxy, but after flying "for real" again, I realize that simming still has a ways to go to catch up with the look of real flight. So here's hoping that we'll see another sim take hold. Or perhaps the next release of X-Plane will take simming to a new level.

 

John

[HiFlyer 9,343]

I would argue it's more that we're pushing the limits of current sims, not current computer technology. FSX leaves plenty of computing power on the table because it simply can't utilize it. Leaving us to tweak the sim beyond what was probably expected.

 

As for games and modern hardware, they are getting better all the time. Even consoles with their multiple CPU or core designs are helping developers figure out ways to make better use of it all.

 

All true enough. ^_^

 

Thinking about it, I wonder how much of the amazement at this demo comes from how far graphics have come........ and how much from how far we have not.

 

As has been pointed out, very near this level of graphics has been out "in the wild" for quite some time, and is only a very small step removed from Skyrim and others........

 

As for Outerra.........

 

[HansRoaming 73]

That still doesn't explain who's going to build the core simulator - the graphics are only a part of it. The companies mentioned develop addons for the existing simulations.

 

From the Unigine site:

 

 

Physics

 

Collision detection

Comprehensive set of geometric primitives (box, sphere, capsule, cylinder, convex hull)

Rigid body physics

Various joints, motors and springs

Dynamic destruction of objects

Vehicle physics

Particle system physics

Cloth physics

Rope physics

Rag doll physics

Inverse kinematics

Force fields

Fluid buoyancy and interaction

Time reverse support

Multi-threaded core

 

A2A also sell their own flight sim, Wings of Victory for example. Other companies mentioned replace everything from avionics through to flight dynamics modelling through to ATC, route planning and aircraft AI. From memory Scaled Composites use X-Plane as a visualiser for their own sim core.

 

Building the next flight simulator will cost more than building fly over ramps for freeway in major city. A core engeine would cost like 100 million at least. These concepts are nice but they are not practical. The cost to build FSX was into the millions and if development of FSX happened in 2008 instead of 2005 the last flight simulator for Microsoft would be flight simulator 2004. They tried to kill flight simulator in after flight simulator 95, but it was Bill Gates as CEO that prevented it from occurring. These videos look nice as concepts but technically and financially unfeasible.

 

FSX was developed over like how many years, you make it sound like the whole multi year personnel budget was spent before they released a shippable product. Look at AeroFlyFS and Kerbal Space programme for examples of simulations that are being developed for way less than you mention.

 

Ultimately your both arguing from a position that the market players and products for flight simulation have reached the pinnacle for whatever is possible for the future history of mankind, even if I personally can't answer your questions by saying how I would do something it would be a strawman to assume that there is or will be no opportunity for a new product to enter this growing market.

[tonywob 4,768]

Problem with Outerra is that it seems to be going nowhere and will be outdated before it amounts to anything

 

This is exactly how I feel about Outerra. It looks fantastic, but development is so painfully slow that it feels like it will never get anywhere. Each update they release offers changes which are mostly unnoticeable, and they've been promising features like real trees and different land textures for a very long time now. Of course, I can imagine it must be incredibly difficult to program something like this, so I have the up-most respect for the guy (It seems to have one main developer) doing it

[simmerhead 885]

This is exactly how I feel about Outerra. It looks fantastic, but development is so painfully slow that it feels like it will never get anywhere. Each update they release offers changes which are mostly unnoticeable, and they've been promising features like real trees and different land textures for a very long time now. Of course, I can imagine it must be incredibly difficult to program something like this, so I have the up-most respect for the guy (It seems to have one main developer) doing it

 

The folks behind Outerra have been vety clear about them not making a flight simulator, so while it looks good it is going nowhere fast for flight simmers at least.

 

 

 

Well, there's still fluid dynamics and physics to consider, both of which require a good chunk of processing power to do well. It's one thing to see the visual effects of weather, for example, but it's what you don't see-- i.e. the computing cycles needed to simulate the way the weather affects a body that's already moving through the air under aerodynamic forces-- that affects the performance.

 

BTW, interesting choice of username if the flying bits aren't that important to you.

 

Sure, the physics will take it's toll, but nothing compared to full systems simulators like the PMDG NGX which wouldn't need such a complex graphics engine.

 

Well, my name isn't "flightsimmerhead". I am more interested in the simulation of planet earth than the actual flying bit since it is pretty tame compared to real flying. I use FSX to go places that I can't go to in real life. Besides, it is a pretty nice way to relax to buzz treetops...

[tonywob 4,768]

I have to admit Valley was really impressive, but only at ground level. In fact, I spent nearly an hour walking round, and I found it quite relaxing. However, from the air the scenery wasn't very impressive, and wasn't meant to be. Although these effects are nice, we don't need them when flying 10,000 ft above the ground. As nice as Orbx's airports are, I tend to just taxi and take off, without really looking around. Enroute scenery however is super important to me.

 

Outerra is, on the other hand, well suited for flight-simulation, and it's a shame nobody is pushing it forward as an engine for a flight-sim. The only flight-sim at the moment that has impressive graphics, great performance and a real feeling of flight for me is AeroflyFS (and this is what I am holding out for to be the next big flight sim)