How much more can FSX handle? (be warned, lots of opinion)

[BaldyB 20]

I often wondered, you may think I'm nuts! If MS wrote a version of FSX for the Amiga platform you wouldn't end up with a better sim. With it's 9 view ports and smooth scrolling, on the ILS? The darn thing did an amazing job with graphics. It was designed to be a flight simulator. The TV stations were having a field day with it with the video toaster program making it do some rather impressive wipes, roll ups and a slew of other effects. It was almost IBM compatible but not quite. No computer, ever, even now, could scroll as smoothly as the Amiga!

 

Give me a break, I'm only 81.     

[mtrainer 136]

There was a MS version of Flight Simulator for the Amiga, I owned it.  The day it came out it represented the best flight simulation available for the general public as it managed to pull ahead of the MS-DOS version which at that time was a little bit older and running on an inferior CPU.

The Amiga did have great graphics but to put all 4096 colors on the screen at once required utilizing a display mode called HAM, which stood for Hold And Modify, and involved assigning an interrupt to the raster for every single scan line, which took a lot of overhead and work for the CPU.

While the machine was overall a huge improvement over the IBM VGA graphics card, by today's standards the Amiga demonstrating that old version of Flight Simulator would be considered a slide show.  It was tough to land as you really had to think ahead of the aircraft because the frame rate was so bad.

The TV stations did use the Amiga but remember the toaster add-on was hardware-level graphics, and not part of the Amiga itself.

As for the Amiga being "almost IBM compatible" I don't know where to even start with that as the Amiga had a Motorola 68000 series CPU while the IBM machines at the time were running Intel 80286 CPUs, completely different architectures and instruction sets.  I know as I used to develop C software on both platforms at the time.

Since this thread really isn't about the Amiga let me pull us back on course.  In fact Mr. BaldyB re-reading your post again I'm not even sure if there was a point there...but I'm a sucker for all things Amiga as I loved that machine.

As Kevin has pointed out, one of the benefits of having a Microsoft Flight Simulator platform that hasn't changed in more than 10 years is the high quality add-ons we find today.  While I'm certainly no advocate of having a new FS re-write every 2 or even 4 years, I do hope that someday we'll be given a "ground up" version of a Flight Simulator that would negate the need to purchase many of the add-ons that currently exist.  By keeping the entire code-base under one roof I would expect module conflicts and bugs to be fewer overall.

As a software engineer for the 2nd largest firm in the United States I understand that undertaking a project like this would take many years, a huge team, and a budget that wouldn't be for the faint of heart.  Return on an investment that large might take decades, if ever, to recoup.  So I am not holding my breath.

Several years ago there was a open-source public domain flight simulator being worked on but I haven't heard of any updates in quite some time.  Something like this might be our best long-term hope for an all new flight model.  Meanwhile we can expect another 10 years minimum of building on the MSFS base package.

I also worry that the fragmentation of the MSFS code (Dovetail / Lockheed Martin) might not be a good for makers of add-ons as it increases their engineering efforts and support models.  This will get worse as the two platforms continue to diverge over the course of time.  But, the competition it sparks should keep both firms on their toes.

I'll make a 10 year prediction here:  In ten years the flight sim of choice for the masses (home enthusiasts) will still be based on MSFS, but we'll see ever larger portions of the code base executed outside the original sim.  3rd parties might even offer SDK packages for license that can be leveraged, for instance an all new GPU based set of flight modeling algorithms, or an optional add-on package that offers an entirely new rendering path.  Under the hood the overall product will look like a complete mess, but most software in place for over ten years already looks that way anyhow.  I'll also venture to say that due to the increased complexity the cost of the base simulator and the add-ons will likely be twice the price of today's offerings.

Either way, I'm extremely thankful for talent we're seeing at PDMG.  From everything I've read and watched about this company they seem to be firing on all cylinders.  I've waited for products like these since I first started flying Bruce Artwick's Flight Simulator for the Commodore 64 and began to realize the future potential of computers and advanced simulations.

It's a good time to be alive.

Mark Trainer

[Werner747 107]

High fidelity in relation to the original 2006 release of FSX, indeed.

 

Though by now PMDG do most of their simulation magic outside of FSXs code base, as do some other top-end developers, so I am not sure if platform stability is that much of a benefactor. Sure, most developers in the business know FSX inside out, as a platform, but are they not mostly dressing up an old lady for the prom, so to speak? 

 

What has been achieved with FSX is pretty incredible. It's been a success story. Though you have to wonder what ORBX and others would be able to create without the limits of the MSFS render engine, 64 bit support, DirectX 12/Vulkan support (keyword being GPU calculating physics in real time) etc.. There is a trend towards platform independent development for a reason. 

 

FSX is not dead yet. Not by some margin. Though I am wondering, when we will start to see features missing in FSX releases, functional ones and not just graphical. I hope we are not at this point yet. So far there has not been a response in regards to the 747v3 to address this. 

I think if there were any specific features missing from the P3D vs the FSX, we would have known by now :-).  As others have said, P3D, although being developed and improved, still really is FSX at the base, so I think it will be very long before reaching this point.   Could be many years in the future!

 

Regards

[Paul_Smith 92]

 

 

I'll make a 10 year prediction here: In ten years the flight sim of choice for the masses (home enthusiasts) will still be based on MSFS,

I dearly hope that you are wrong, but the pessimist in me is afraid that you are probably right.

 

I am also inclined to agree with your view of the investment model being that no one organisation can expect a return on the investment involved in starting again from scratch, which really only leaves one viable option for the future of our hobby, though I must say I think it is very unlikely to actually happen. That is if Austin Meyers and Laminar Research were to open source the core of x-plane 11, there is a possibility that it would be picked up and enhanced by a much larger community of developers and hobbyists. We might then see the physics based simulator we dream of, but the reality is that I just don't see it happening. There is still the outside chance of a full world simulator coming from another source, such as a game engine being retired, but as you might have guessed, I am not overly optimistic about the future of our hobby. And sorry, but one PMDG release every three years or so is not enough to keep a platform alive, not when it is just another Boeing and a re-release at that!

[BaldyB 20]

 

 

Either way, I'm extremely thankful for talent we're seeing at PDMG. From everything I've read and watched about this company they seem to be firing on all cylinders. I've waited for products like these since I first started flying Bruce Artwick's Flight Simulator for the Commodore 64 and began to realize the future potential of computers and advanced simulations.

It's a good time to be alive.

Mark Trainer

My fond memory's of the Amiga may warp my brain cells over time but, Man I loved that machine!

I retired 16 years ago and thought I would have plenty of time on my hands, never happened. As an Operating Engineer by trade I am OK. The kids needed some help sold some property to help them get started, not much time!  I now have sufficient time to fly. I agree 100% with respect to PMDG, waiting for the Boeing 747-400. If it's modeled as well or better than the 777 it has to be a winner. I have wondered if you could capture all of the talent and get them to focus, (That includes guy's like you!), WOW! Some don't like coding by committee but?  Just installed Aivlasoft on two computers, it's a very nice app. Don't get lost and using up all my fuel taxiing around airports. Trust me that's a plus.

 

Baldy

[kevinh 409]

I dearly hope that you are wrong, but the pessimist in me is afraid that you are probably right.

 

I am also inclined to agree with your view of the investment model being that no one organisation can expect a return on the investment involved in starting again from scratch, which really only leaves one viable option for the future of our hobby, though I must say I think it is very unlikely to actually happen. That is if Austin Meyers and Laminar Research were to open source the core of x-plane 11, there is a possibility that it would be picked up and enhanced by a much larger community of developers and hobbyists. We might then see the physics based simulator we dream of, but the reality is that I just don't see it happening. There is still the outside chance of a full world simulator coming from another source, such as a game engine being retired, but as you might have guessed, I am not overly optimistic about the future of our hobby. And sorry, but one PMDG release every three years or so is not enough to keep a platform alive, not when it is just another Boeing and a re-release at that!

I don't understand why you have so much invested in hoping the future is not based on MSFS in some way. P3D will be 64 bit eventually. The new DTG sim is 64 bit and also a development of MSFS. We already have physics based sims in the form of FSX, P3D and XP10. We don't need to dream of them.

 

It isn't only PMDG keeping FSX alive. FSL, Aerosoft, ORBX, HiFi Technology are just some of the other developers who do.

 

It makes mo sense to start from scratch with a full world simulator. Better to start with an existing structure and improve it.

[MorsAbAlto 19]

 

 

We already have physics based sims in the form of FSX, P3D and XP10. We don't need to dream of them.

 

Their approaches to, ahem, "physics" differ though and the MSFS approach is the most legacy of them - lookup tables are so 1980s and leads to all kinds of weird behaviour. Surely no-one is asking for Level D kinds of precision, but a true physics based engine, enabled by GPU-assisted computing, would go a long way to enhance fidelity for us simmers. Sure, there is a lot of fantasy and good will involved. But instead of constantly "hacking" the FS flight model, developers would be able to devote more time to other things, if the simulator would just calculate the proper values in real time. The newest generation of GPUs offers vast amounts of computational power when used with a modern API. Most of it goes to waste with the current simulators. From an enthusiast point of view it's a tragedy.

 

Have a look at DTGs Train Simulator. This products update policy is terrible. They are switching to Unreal Engine 4 for their next product release, because their current engine is hopelessly outdated. Why would you expect different from their new flight sim? Engine improvements? Call me a pessimist, but I will believe it when I see it. That leaves XP and P3D as serious contenders. One of which is exclusively offered to the professional market. 

 

So I'd have to agree with Paul mostly. The future is not looking too bright. FSX is a decent platform, though it falls behind current technology ever faster. We need a company with enough budget to start from scratch or at least adapt a suitable engine for flight sim purposes. The current example for re-writing an engine would be the development of Star Citizen, a big budget space game. 

 

There is one other prospect for us simmers: crowd funding. Maybe some day a seasoned developer will stand up to the challenge of creating a superior engine with a few millions. 

[kevinh 409]

Their approaches to, ahem, "physics" differ though and the MSFS approach is the most legacy of them - lookup tables are so 1980s and leads to all kinds of weird behaviour. Surely no-one is asking for Level D kinds of precision, but a true physics based engine, enabled by GPU-assisted computing, would go a long way to enhance fidelity for us simmers. Sure, there is a lot of fantasy and good will involved. But instead of constantly "hacking" the FS flight model, developers would be able to devote more time to other things, if the simulator would just calculate the proper values in real time. The newest generation of GPUs offers vast amounts of computational power when used with a modern API. Most of it goes to waste with the current simulators. From an enthusiast point of view it's a tragedy.

 

Look up tables form the basis of every full flight simulator I've worked on or know of. They certainly are physics based simulations. The tables compute aerodynamic coefficients as a function of the input variables, such as alpha, Mach, altitude, etc. It's a very data friendly way of calculating these variables which feed the equations of motion. You don't have to hack at the FSX flight model because you can enter data into the air file in a very straightforward way. The FSX model is limited in the variables it takes into account but the method is very definitely physics based. It's perfectly possible to produce a convincing and realistic simulation in FSX without hacking at it or extending it. You just need to understand how to use it.

 

FSX certainly does calculate the proper values in real time. I can only assume you have misunderstood what "look up tables" means in this context.

 

As far as I'm aware GPUs can't yet calculate aerodynamics so they wouldn't be much use for such a specialist task.

[Werner747 107]

 

 

There is one other prospect for us simmers: crowd funding. Maybe some day a seasoned developer will stand up to the challenge of creating a superior engine with a few millions.

 

although the idea is sound, I am not sure that all the funding for a new research project on the scale you suggest is practical.   If we are talking millions, how many millions?  To give you an idea, look at P3D and the funding behind that :-).   Just saying...

 

Regards

[Cann0nF0dd3r 21]

I think the problem with crowd funding is the same reason we haven't had a new version of Flight Simulator and Microsoft canned the development team......interest. To be honest FSX/P3D/XPX are niche games, IMO if there was enough interest for a crowd funded flight sim PMDG would have enough customers to not have to sell their products at $60-130!

[Werner747 107]

I think the problem with crowd funding is the same reason we haven't had a new version of Flight Simulator and Microsoft canned the development team......interest. To be honest FSX/P3D/XPX are niche games, IMO if there was enough interest for a crowd funded flight sim PMDG would have enough customers to not have to sell their products at $60-130!

True which is why I say that the development costs involved is just too much to realistically expect the community to carry.  It is not the Call of Duty franchise that is for sure.

 

Regards

[Cann0nF0dd3r 21]

I'm cautiously optimistic for DTG Flight Sim, I think they have a great starting point with Flight School and if they can pull it off it might breathe some new life into the genre.

[MorsAbAlto 19]

 

FSX certainly does calculate the proper values in real time. I can only assume you have misunderstood what "look up tables" means in this context.

 

As far as I'm aware GPUs can't yet calculate aerodynamics so they wouldn't be much use for such a specialist task.

 

You seem much more knowledgeable in this field of technology than me. So I concede that point. Though isn't that what is happening in every airframers development lab? Calculating aerodynamics? Maybe they are using CPU farms for that, but seeing how many applications GPUs and PPUs have nowadays it would really surprise me if it could not be done.

 

But let me say, that the limited set of variables in FSX do make for some weird physics indeed, especially with the default planes and the famous "flying on rails feeling". What I was thinking of, was an approach similar to X-Plane, where the aerodynamics of the airfoil are actually taken into account. As far as I know, you can model a supersonic flying trash lorry in FSX. In the sim that I am proposing, that would be impossible or at least very hard.

 

 

True which is why I say that the development costs involved is just too much to realistically expect the community to carry.  It is not the Call of Duty franchise that is for sure.

 

You might well be correct, but let someone put this assumption to the test. After all we have people around here that spent thousands of dollars for their sim setup, why not invest a few hundred more in a future sim?

[downscc 1,327]

 

 

"flying on rails feeling"

 

I've seen that accusation for years and I've never understood it.  In a literal sense it is absurd, there are no sensations or feelings when flying any PC based simulator.  As a real world pilot since 1974, one doesn't not compare the flight sensations of a simulator with real flight.  The lack of any movement suspends any comparison between the sensation of flight and the simulation.

 

Just what does that mean? Is this a claim that there is no slip or skid in the simulation?  I've never felt that my NGX in either FSX or P3D was exhibiting odd behavior within the normal flight envelop. True that at the edges of the envelop the simulation looses fidelity but that's true of all except million dollar simulators.

 

My earlier contribution to this thread had something nice to say about all three major platforms, but this snippet of FSX limitations just happened to rub me wrong today. Sorry.

[kevinh 409]

You seem much more knowledgeable in this field of technology than me. So I concede that point. Though isn't that what is happening in every airframers development lab? Calculating aerodynamics? Maybe they are using CPU farms for that, but seeing how many applications GPUs and PPUs have nowadays it would really surprise me if it could not be done.

 

A GPU can calculate a graphical effect for display. It could possibly use the shape of the model, plus additional data relating to airflow, to compute aerodynamic forces. But how does it feed that back to the sim? It might become possible in the future but for a simulation it's better to keep the core of it in the CPU.

 

But let me say, that the limited set of variables in FSX do make for some weird physics indeed, especially with the default planes and the famous "flying on rails feeling". What I was thinking of, was an approach similar to X-Plane, where the aerodynamics of the airfoil are actually taken into account. As far as I know, you can model a supersonic flying trash lorry in FSX. In the sim that I am proposing, that would be impossible or at least very hard.

 

The flying on rails feel is due to (1) default air file settings and (2) FSX's limited weather engine. But even the default aircraft experience sideslip in crosswind and propeller slipstream effects. So it is something of an exaggeration. Using good payware addons and a decent weather engine it simply does not exist. Do you think PMDG sims feel like they fly on rails? In contrast X-Plane is far too lively and skittish, especially when simulating large aircraft. Neither is perfect. The fact that the external model shape makes no difference to FSX is completely irrelevant to simulation. X-Planes aerofoils need look up tables to compute lift, drag and pitching moment. All it does is take things to a lower level of detail. It's good for "what if" experiments, but if you want an accurate simulation of a real aircraft the "whole aircraft" aerodynamic model, as used by FSX, is easier to deal with and adjust. Where X-Plane scores is it's more detailed equations of motion. You could develop FSX, include more aerodynamic variables, and it could be every bit as good as X-Plane. DTG's new flight sim could well do exactly that. As long as it's open to third party developers there is no danger of dumbed down physics taking a back seat to visual appearance.