Gaming FSX ... Oh Fsx...

[Hielor]

Flightgear? OpenSource and free.

And also not really a serious contender, compared to FSX or X-Plane. As they say--you get what you pay for.

 

[SolarLiner]

Once I performed a short haul flight on the a small passenger twinjet (Wilco E-Jet or Aerosoft Airbus X, I believe), and made a mistake during the fuel requirement "calculations": after touching down I had a single-digit amout of lbs remaining.

That's not a mistake, that's an incredibly good money-saver ! :lol:

I once did a Toulouse - Dublin flight, native A321, with FS Passenger. I don't know why, but apparently Nantes had DCA firing. Did not finish the flight

 

[Screamer7]

Is FSX's executable using Large Address Aware flag? If it isn't, it can only address 2 GiB of virtual memory. You can try enabling this flag if it isn't already set to get additional ~1.5-1.9 GiB usable by the game.

More info here ORB:

http://kostasfsworld.wordpress.com/fsx-oom-and-addon-vas-usage/

Cheers

Johan

---------- Post added at 06:01 PM ---------- Previous post was at 05:59 PM ----------

I use this tool everytime I reïnstall FSX or after hardware upgrades.

I used it too, and it does make some improvements to FSX.
The most important one is the HIMEMFIX option.
It's essential for FSX.

Cheers

Johan

---------- Post added at 06:19 PM ---------- Previous post was at 06:01 PM ----------

You can save your flight and resume with the saved state, but what if you are on your final approach and you are on VATSIM?
Then you have problems.
And the OOM's tend to happen when you are near your destination.

Do this if you have FSUIPC running:
Click on FSUIPC,
Click the LOGGING tab,
Enter these values in the following boxes.
Leave BASE as IPC,
In the OFFSET box enter 024C
In the TYPE box enter U32
Leave the HEX box unchecked.
Now you can see how much VAS is left if you run FSX in windowed mode.
It is displayed at the top middle.
Please note that there is a difference between Virtual memory and Virtual Address Space.
The programmers will know about this.
You can also see it in a program called processexplorer.

 

[Thorsten]

And also not really a serious contender, compared to FSX or X-Plane.

Out of professional curiosity - why would you think so? (Disclaimer: I'm part of the Flightgear development team)

Flightgear is OpenSource, which means that not every aircraft and airport comes with the same quality, but if you use the high-quality planes available for FG and look at well-developed corners of the world, then...

I have looked a bit into FSX and compared, and
* Flightgear has definitely superior flight dynamics, I found FSX really unrealistic
* Terrain texturing is a bit better on FSX, not much though these days
* FSX still has a few eye candy features FG doesn't have though
* FG does weather and clouds much, much better, likewise lighting in low sun and interplay between lighting and weather is better handled by Flightgear
* I guess user-friendliness in terms of configuration etc. is up to force of habit

All in all, my impression was that FSX goes for WOW! effects which are sort of cute, but don't actually happen in real life that often (like plane reflections in the ocean) - FG doesn't do this, reality more often than not looks a bit dull and hazy...

Granted, I'm hardly impartial, and fond of tinkering and customizing, so FG was a natural choice for me because you can tinker with basically anything without much ado (it's made to be tinkered with). But I think also in other aspects it's quite competitive.

As they say--you get what you pay for

Which would make Orbiter crap then? Linux is something that happened to other people? Mozilla ain't working? Wikipedia is a pile of nonsense?

I don't know who says these things - but they're demonstrably wrong...

 

[boogabooga]

Out of professional curiosity - why would you think so? (Disclaimer: I'm part of the Flightgear development team)

* Flightgear has definitely superior flight dynamics, I found FSX really unrealistic

Are you an engineer? (serious question)

 

[Thorsten]

No, theoretical physicist. With a pilot's license during some time in my life (not active right now though).

 

[Hielor]

So, first off, these posts might should be moved to the Flight Sim Megathread, where discussions like this typically take place.

Out of professional curiosity - why would you think so?

Starting with my "credentials"--I'm also a private pilot, so I have a little experience with how light planes perform in the real world. I've been using flight sims "heavy casually" for over a decade now. I also ran an FSX server for several years that I spent a good deal of time using and supporting, so I'm not entirely unbiased either.

My experience with FlightGear is also limited to a single attempt several years ago, when despite having a strong computer that could run FSX at good framerates, FlightGear ran at ~10fps. Maybe I should give it another go, to be fair.

Flightgear is OpenSource,

Being OpenSource does not automatically make a software product better than a non-OpenSource product, barring religious fervor, but that's a completely different argument.

which means that not every aircraft and airport comes with the same quality, but if you use the high-quality planes available for FG and look at well-developed corners of the world, then...

You basically just started your points with "if you completely ignore all the bad parts and look only at the good parts..." By that metric, Microsoft Flight was actually pretty good (ignoring the absurd pricing scheme).

* Flightgear has definitely superior flight dynamics, I found FSX really unrealistic

Ah, the X-Plane fallacy. Having marginally better flight dynamics doesn't (or at least, shouldn't--they perhaps fall for the hype) actually matter to the majority of desktop simmers. FSX's flight dynamics are fine for the average flight simmer, and it all depends on the plane--sure, some planes don't have the most realistic flight model, but then again, I bet some of the planes in FlightGear aren't perfect either.

From my perspective, FSX's flight model was good enough for the majority of flying I wanted to do--some of which was training for the real thing. The closest analogue to desktop simming in the air is instrument flight (since you specifically need to ignore bodily cues and only pay attention to instruments), and my instructor thought we'd met the required practice time for that a whole lot sooner than we actually had since I was so good at it from the FSX experience.

* Terrain texturing is a bit better on FSX, not much though these days
* FSX still has a few eye candy features FG doesn't have though
* FG does weather and clouds much, much better, likewise lighting in low sun and interplay between lighting and weather is better handled by Flightgear
* I guess user-friendliness in terms of configuration etc. is up to force of habit

FSX is eight years old. There is no excuse at all for FG being behind it in anything at all. FG (and X-Plane) have had eight years to catch up to FSX, and it's completely inexcusable for FSX to be better than either of them in anything.

All in all, my impression was that FSX goes for WOW! effects which are sort of cute, but don't actually happen in real life that often (like plane reflections in the ocean) - FG doesn't do this, reality more often than not looks a bit dull and hazy...

Average desktop simmer probably isn't looking for something 100% true to life. And on a calm day, you can certainly see reflections in bodies of water.

Granted, I'm hardly impartial, and fond of tinkering and customizing, so FG was a natural choice for me because you can tinker with basically anything without much ado (it's made to be tinkered with). But I think also in other aspects it's quite competitive.

It's competitive if you really want to tinker with the core program, and/or if you can't afford the $15 price tag for FSX.

The situation leans even further towards FSX when you look at the huge library of available addons, both free and payware. It's not just planes and airports, either--look at things like FSPassengers, which breathe new life into FSX simming. The addon options for FlightGear pale by comparison.

Which would make Orbiter crap then? Linux is something that happened to other people? Mozilla ain't working? Wikipedia is a pile of nonsense?

When compared to existing non-free options. Orbiter isn't crap, because there aren't any non-free options in the same genre to compare it to. Linux does some things great, other things notsomuch (my extensive game library mostly won't work on Linux, for example). Mozilla is irrelevant, since there aren't any payware browsers anymore. Wikipedia is a great resource for learning a little about a lot, but if you want to learn a lot about something you're going to have to go elsewhere.

I don't know who says these things - but they're demonstrably wrong...

Not in the case of FlightGear. I'll do the same for FG that I do for X-Plane --Let me know when it does everything better than FSX, without excuses, then I'll consider switching my civilian flight simming over to it. Eight years is an eternity in the software world. Microsoft canning FSX has completely destroyed the desktop flight sim market, since the competitors have stopped competing.

Personally, I have higher hopes for DCS World to become an all-inclusive flight sim, but their terrain model doesn't lend itself to worldwide coverage.

 

[Evil_Onyx]

I'm flying with Ideal Flight 10 and its making things more fun for me, better weather, better traffic and decent flight planning. So far its given me lots of fun flights to places I would not usually visit. I would never have visited Salzburg (LOWS). The feature I like the most is that it can save where you parked all the aircraft so you can pick it up at a later date.

My most recent incident in FSX happened whilst ferrying a F-14D from Nellis to Andrews, with random failures as usual. Over Illinois I get a master warning and 2 secs later I lose all power and start to freak out because the weather below me is stormy with low vis and its night. So I look at the fuel gages and determine I could fly on for about 2 hours at altitude, so i start flying by the stars. An hour later the weather clears and I decide to find a runway that I can land at, and finally land at Pittsburgh International Airport (KPIT) after a heavy but safe landing (Yay for a aircraft designed for carrier ops).

 

[Urwumpe]

I think in terms of realistic flight dynamics, DCS with a AFM-based aircraft is the reference. The dynamic behavior is really well done (by changes in AOA, slip) and the effects of wind and payload are really as you should expect. Ground effect is there, effect as it should in theory - I am no pilot to compare.

Can't speak much about the effects of damage on the aircraft. It has an effect, but who can tell if the effect is really realistic.

FSX is good enough for people who want to fly multiple different aircraft without having to certify first for the class. DCS is better suited for people who want to specialize.

 

[Screamer7]

Check out this screenshots from X-Plane 10 64 bit.
All the scenery and aircraft in the screenshots except the second one, is freeware which is pretty damn good if you ask me!
The second one is an conversion to X-Plane from Tampa Dubai reloaded. (FSX)

I think P3D just got some serious competition!

I can see myself sitting in a X-Plane cockpit sooner than later

 

[Thorsten]

@Hielor:

Remember what you claimed:

And also not really a serious contender, compared to FSX or X-Plane.

I'm not arguing that Flightgear works better for you. I'm not arguing it is superior in everything. But I am arguing that it is a serious contender these days and you can't dismiss it.

My experience with FlightGear is also limited to a single attempt several years ago, when despite having a strong computer that could run FSX at good framerates, FlightGear ran at ~10fps. Maybe I should give it another go, to be fair.

So basically you don't really know?

Ah, the X-Plane fallacy. Having marginally better flight dynamics doesn't (or at least, shouldn't--they perhaps fall for the hype) actually matter to the majority of desktop simmers.

Well, the base planes I've tried from FSX were really bad, to the point that I really got annoyed about the lack of obvious dynamical effects which have to be there. Probably if I invest in a few addon packages, than I get something that works better in FSX - point being, why should I shell out money to get FSX to a state which FG can deliver for free?

It may not be that important for you or the average simmer (whoever that might be) - but it does matter to me quite a bit.

You basically just started your points with "if you completely ignore all the bad parts and look only at the good parts..." By that metric, Microsoft Flight was actually pretty good (ignoring the absurd pricing scheme).

You have to acknowledge the essential differences in structure between a commercial and an all-volunteer project. If you create an airplane for FG, it will be tested whether it's running bug-free, but it won't be reviewed for accurateness of flight dynamics or for beauty of cockpit. The reason is that different people like to do different things with FG and some prefer a simplified helicopter over the real simulation which they can't get off the ground. For the project, both uses are equally legitimate.

The implication is that FG has a huge range of options. There's three different rendering pipelines to begin with (for legacy support, for good scenery and weather visuals, for dynamical shadows), three different native flight dynamics codes,... it's basically a toolkit (it's even been used as a rendering backend in a post-apocalyptic game).

Which means that if you want to do anything in particular, for instance experiencing realistic flight, then you have to configure the toolkit for your purpose before using it. Selecting the group of planes that has the features you want. Setting the options like you need them.

FSX is eight years old. There is no excuse at all for FG being behind it in anything at all. FG (and X-Plane) have had eight years to catch up to FSX, and it's completely inexcusable for FSX to be better than either of them in anything.

I don't have the numbers ready for FSX, but I do know for the abandoned Flight! project. In terms of man-hours per year of people coding, they had about 100 times more than FG. If I had a full-time graphics artist at my disposal, I could do amazing things much faster. Alas, if I have to do it myself, it's not nearly that fast.

I think it's quite an achievement to be competitive with a multi-million $ enterprise as an OpenSource all volunteers project as such.

The situation leans even further towards FSX when you look at the huge library of available addons, both free and payware. It's not just planes and airports, either--look at things like FSPassengers, which breathe new life into FSX simming. The addon options for FlightGear pale by comparison.

In fact, FG hardly has any addons because the whole point is that everything which works gets distributed with the next edition of the FG core - we aim to create a state where there are none. You don't ever need to manage mutually exclusive addons, you don't need to install additional scenery, because by and large it all comes from one central scenery server, and you get the addons delivered with the base product. Different philosophy. So if you're looking for addons, no wonder you don't find any.

Addons are really a suboptimal way to improve a product - the better way is to improve the product itself

Not in the case of FlightGear. I'll do the same for FG that I do for X-Plane --Let me know when it does everything better than FSX, without excuses, then I'll consider switching my civilian flight simming over to it.

I don't want you to switch anything - quite evidently you're getting from FSX what you like in a flightsim. I do want you to acknowledge that FG is a serious contender - better in some areas, worse in others.


@Urwumpe:

I think in terms of realistic flight dynamics, DCS with a AFM-based aircraft is the reference. The dynamic behavior is really well done (by changes in AOA, slip) and the effects of wind and payload are really as you should expect. Ground effect is there, effect as it should in theory - I am no pilot to compare.

Since we can't do real-time computational fluid dynamics, flight dynamics is cast by almost everyone in the professional business into coefficient schemes - a wing is characterized by lift and drag functions which are multidimensional lookup tables as functions of AOA, Reynolds number, ...

Computational cost of these schemes is low by modern computer standards, they're just differential equation solvers. They can be benchmarked (JSBSim from FG for instance is tested against some NASA flight dynamics tools), but the accuracy for any given airplane lives and dies with the data tables. Which means either real wind-tunnel data or offline computational fluid dynamics on a really accurate 3d model of the wing.

It's largely an exercise in data-gathering or waiting for 1000 hours for a fluid dynamics code to go over your modelled airfoil. Flightgear has some good planes here since it's often used by university researchers for some project - and then the data flows back. And there's tenacious perfectionists who spend years digging up data and refining tables...

---------- Post added at 07:58 AM ---------- Previous post was at 07:37 AM ----------

And while we're doing screenshots - less WOW!, more muted colors just as reality has:


How to do clouds:

How to do water, and its interactions with wind, and how it changes with the light:

How to light and shade clouds in low sun:

Atmosphere visuals to the edge of space:

...and back:

More clouds:

Scenery closeup - that's on Iceland:

And that's in the central Karakoram range:

And there's also nice cockpits - A-10 approaching Srinagar.

 

[Hielor]

Addons are better than adding everything to the base product because good addons can be really large, once you count high-res textures, effects, etc. Maybe I don't care about having xyz airplane, so I don't want to have to download it...

I was severely unimpressed with the "amazing" X-Plane 9 dynamic flight model. Flying a 172 at night over my local area, at night, with clear skies, at normal cruising speed, the plane was bouncing all over the place. That's not what happens at night with clear skies...

But, I'll keep in mind the improvements you mention, and maybe I'll give FG a try next time I want civilian flight simming...

 

[Urwumpe]

Since we can't do real-time computational fluid dynamics, flight dynamics is cast by almost everyone in the professional business into coefficient schemes - a wing is characterized by lift and drag functions which are multidimensional lookup tables as functions of AOA, Reynolds number, ...

By the same argumentation, you can call a Ferrari just a fancy way to build a Fiat. It is just a car, because we can't build flying saucers yet.

Or you can argue, that a Titan is just a better pocket calculator.

You are right - it is a simplified model of the reality. That also applies to CFD BTW, even if you do Direct Numerical Simulation. It is also only an approximation. A much closer one, but still not the real behavior.

What you wipe away with your seen-it-all-know-it-all attitude, is the simple fact, that you still have two different dimensions of quality in the approximations:

You can have more data in the coefficient and look-up tables.
You can have more and more accurate functions considering more inputs putting the coefficients and look-up samples together into lift, torque and drag.

And real-time CFD, as much as it gets possible with time, is not the better solution to a "coefficient scheme" - just as much as it makes no sense to simulate the motion of a baseball atom by atom. Even if it would be more accurate.

 

[Thorsten]

Urwumpe, before you go on here and say funny things about my attitude, you should perhaps know that computational fluid dynamics is something I teach to students. I know exceedingly well what I am talking about here, and I can give you the whole lecture if you really like...

Proper (= considered realistic by NASA or professional aircraft constructors) CFD for a wing is something that takes of the order of 10.000 to 20.000 CPU hours. A frame duration is 20 to 50 ms. That's right now about a billion (!) times slower than you can do in real time.

So doing it offline to create the coefficient tables makes sense. More so once you realize it doesn't need to run in real time, because the 'steady state' approximation made per frame by the use of the coefficient tables is really good for almost any situation (flying your aircraft through a sharp shockwave is about the only thing where I can see it fail).

You can have more data in the coefficient and look-up tables.
You can have more and more accurate functions considering more inputs putting the coefficients and look-up samples together into lift, torque and drag.

I fail to see how that is not the same thing. If you know your coefficients for all airfoils and elements as a function of AOA, Reynolds numbers, etc., you end up with a set of forces based on current AOA, Reynolds number etc.

From there, it's Newtonian mechanics - you apply the forces to the aircraft, you end up with a set of torques, you apply that to the intertial tensor of the aircraft, you get the proper accelerations. That problem has been solved 20 years ago.

Realism is down to having more accurate forces, they are down to more accurate coefficient tables and more dimensions to the coefficient tables.

So I don't get the point of your comment. Well, frankly I don't get the point of your whole post - I wasn't advocating real time fluid dynamics for anything, I was trying to explain to you where realism comes from and how to achieve it.

But it seems you already think you know all about it, so it's probably a pointless exercise

Addons are better than adding everything to the base product because good addons can be really large, once you count high-res textures, effects, etc. Maybe I don't care about having xyz airplane, so I don't want to have to download it...

The FG scenery rests on an SVN server. All addons are checked into this server, and there's an in-sim tool to sync to the server, so it automatically loads all the latest scenery for the place you currently are. So if you fly from an airport, and somebody has added models last night, you would get it automatically when you start up. Likewise, you don't need to pre-install any scenery - you just select an airport anywhere in the world, scenery will automatically be loaded and you're good to go.

There's also offline distribution channels available where you can download from the server what you need. But that's basically how it works.

Airplanes are likewise checked into the FGData GIT repository, and for each release (every 6 months) the whole list is put up for separate download.

I was severely unimpressed with the "amazing" X-Plane 9 dynamic flight model.

I have no experience with X-plane, so I can't really comment either way.

I will say that I am very sceptical about claims that they'd do something like a decent approximation to real time fluid dynamics with this blade approach. Where it matters in industry, coefficient schemes are used.

 

[Urwumpe]

Urwumpe, before you go on here and say funny things about my attitude, you should perhaps know that computational fluid dynamics is something I teach to students.

I do CFD (better: Software for pre- and post-processing and user functions) for a living. I don't teach it students, my customers are people who already know CFD, who are working on the edge of the technical and mathematical state of the art, and who are the first to know, when the real aerodynamics differ from the simulation (automotive design).

So please: Stick your argument by authority where the sun doesn't shine. It doesn't work here.

BTW, some weeks ago, you had been theoretical physicist in your presentation. Is that a translation error or do theoretical physicists in Finland teach aerodynamics for a living?

And PS: You need FVM CFD for having first data - before you can abstract it into a vehicle scale model. Think about it. When NASA does a simulator model to test how an aircraft/spacecraft will fly, it sure won't use a CFD system that needs a weekend to compute every iteration. Still you can be sure that the real-time simulator models today used in the design of high performance aircraft are better than the CFD models of the 1990s.

 

[garyw]

O-F Staff Note: Thread locked.