Author Topic: AW To get a new flight model  (Read 655 times)

Offline AKDejaVu

  • Platinum Member
  • ******
  • Posts: 5049
      • http://www.dbstaines.com
AW To get a new flight model
« on: April 14, 2001, 12:22:00 PM »
Posted by badboy in the wrong forum (but it is a good one.. so I brought it here):
---------------------------------------------

Hi,

EA have posted a news article:
 http://www.ea.com/worlds/games/pl_airwar00/nw_devnews1.jsp

It seems they would like us to think that they have something new and spectacular in store for us in the future. With outrageous statements like “No other flight sim had anything remotely approaching this, no matter how famous its brand name.” Their advertising spiel attempts to convince us that they have some “well-kept secret” that would be “dynamite” if their customers are all patient enough to continue paying for it while they work on it.  Unfortunately, those claims would have been funny if it wasn’t for the fact that those customers waiting for something good are destined for such a big disappointment.

In order to expose the weakness of their claims, let’s look at the details of what amounts to little more than EA propaganda.

 
Quote
During the 1980s and early 1990s, flight sim creators usually relied on programmers to create the flight model. Given the limitations of the PC, this was the only feasible method.

What flight sim creators are they referring to? At least one of the flight simulations published by EA during that period had an excellent flight model and was written by an aero graduate. At least three of the flight simulations I’ve been involved in, or written about, in the same time frame, have had aerodynamics engineers involved in the project. But they did say “usually” so I guess they covered their bases on that point.

 
Quote
But in the late 1990s, EA.com's Kesmai Studios began working on a revolutionary new approach to flight models that started with flight models for a Vietnam-era flight sim

It takes some balls to cite vapour-ware in your advertising   Air Warrior Vietnam used an approach so revolutionary the world wasn’t ready for it…lol. Oh but wait… They are going to tell us what the revolutionary part is:

 
Quote
In seeking a better answer, the Kesmai team turned to two respected experts in this field. The first was Robert Shaw, author of "Fighter Combat," widely read by military combat pilots as well as flight sim gamers. The second was Dr. Brian Stevens, author of the classic standard graduate-level aeronautical textbook, "Aircraft Control and Simulation," and currently a fellow at the Georgia Tech Research Institute.

This is a very common fallacy, and unfortunately there has always been a trend in advertising to take an inferior product and lift it onto the shoulders of celebrities. In a flight simulations context this amounts to giving a copy of your game to a test pilot or war hero and quoting them on the box saying how realistic the flight model is. EA are taking this a step farther, Respected fighter pilots and honourable academic figures make for good advertising, but do they make for a good flight model? Let’s allow the rest of the news flash to speak for itself.  

 
Quote
Stevens and Shaw developed a method that combined tabular aircraft data with some very sophisticated non-linear algorithms.

Hmm, could this be the new approach they have been referring to? I hope not because there is nothing new about that… “tabular aircraft data” simply refers to the use of look up tables, a technique that has been used in flight simulation from its inception. The “non linear algorithms” might be called sophisticated, but all it really means is that instead of modelling some of the aerodynamic relationships as straight lines, they use curves. Is that new? Well one of the first real time simulators to use non-linear functions was the Bell & Western flight simulator built in 1943. About 32 of these were built and they simulated the aerodynamic characteristics of at least seven different aircraft for the US navy. Almost 60 years old, it can hardly be called revolutionary. But wait they are talking about PC based flight simulations right? Well Spectrum Holobyte had a non-linear flight model in Falcon3 when it was released way back in 1989, and Falcon4 features a combination of tabular aircraft data and non-linear algorithms, so it has been done before. Perhaps what they meant was that the method was new to them, and revolutionary only to themselves? Sheesh, they covered their bases there too.

 
Quote
The results were amazing. The results were so accurate that the model even produced the correct "feel" of the plane as it flew through the air.

Amazing… Of course no other flight simulation has ever done this… I laughed so hard when I read that it almost hurt.

 
Quote
For many WWII era planes, this information simply wasn't available. We solved this problem by acquiring a commercial aerodynamics-modeling package.

While the developers who already have good flight models continue to scrape together every scrap of real world data, building their models so as to produce results as accurate as possible to how the real aircraft performed, we are informed that EA will be using a bog standard aero package to help guess how they performed. Anyone else think that EA are missing the point?

Personally, I would rather accept what meagre flight test data is available, backed up with anecdotal evidence from the pilots who flew it, over just another aero package, particularly one where they couldn’t even trust the “expensive” prop add on.

They also make the excuse that the prop has been left off because they were doing something even more accurate… Well, I guess they already knew that aircraft behave very differently under power, than when gliding without a prop, and the image they have posted shows very clearly an aircraft with streamlines entirely unaffected by propwash, which of course means that they will produce correspondingly erroneous aerodynamic coefficients. Naturally, they intend to fudge that later  

 
Quote
A highly detailed model of an airplane goes in, computer does a bunch of calculations on it through various angles of attack, yaw rotations, elevator deflections, etc., and out pops a *lot* of data. Good geeky data that engineers love. Good geeky data that makes for cool flight models.

So that’s how they are obtaining their tabular data… And their fancy aero package must be the non-linear part… Oh boy my sides are hurting! A "cool" flight model, EA are killing me!

 
Quote
Our model also includes specific engine performance attributes, real RPM values, changes in output with altitude and/or boost, and more.

Off course it does, the flight model wouldn’t be worth a damn if it didn’t! The only WWII simulation that has come out in the last few years that didn’t do this was… Hmm let me think… Yep Air Warrior!

 
Quote
Will these new models be harder to "fly" in the game? Well, yes, if we don't do anything. But we've noticed that relaxed realism is more popular than full realism. Flinn had the breakthrough idea of applying modern "fly by wire" concepts to the game.

So let me get this straight, they are building a “revolutionary” new flight model that will be “dynamite” and it is for relaxed realism? What’s wrong with that?   If they only want something for the relaxed realism player, why don’t they stick with the flight model they already have?

 
Quote
A lot of the drudgery is taken out of getting the plane to go where you want, but the performance and handling are correct, the edges of the envelope are where they belong, and you can watch the rudder move as you maneuver.

That’s ok then, they are just taking out the drudgery, I can still watch the rudder move… Phew for a minute there they had me worried! But wait a moment, they say the performance and handling will still be correct… Don’t forget, when they say correct, what they really mean is that it will match the “geeky data” (their words not mine) output from their dodgy aero package.

So, it looks to me as though we will be getting something different, but it won’t exactly be new or revolutionary. The good news is that for those in the community really interested in aircraft that match their real world counterparts, as closely as possible, and in high fidelity aerodynamic similarity, the simulations out there already delivering that aren’t going to be threatened by AW, that seems to be destined to maintain its status as the laughing stock of the WWII simulation community.

Badboy

Offline Revvin

  • Silver Member
  • ****
  • Posts: 1724
      • http://www.ch-hangar.com
AW To get a new flight model
« Reply #1 on: April 14, 2001, 12:58:00 PM »
Why knock them for trying to move the sim forward? Lets wait and see what they come out with first and then make an opinion. Mind you I have to laugh at the 'secrecy' about this kinda reminds me of how Hotpants over at iEN kept telling us how everything had to be kept secret because once we eventually saw it it would be like the second coming..well Warbirds III beta is out and apart from a few small touches its the same old sim.

Squidzilla

  • Guest
AW To get a new flight model
« Reply #2 on: April 14, 2001, 04:29:00 PM »
Knock them?

It's going to be geared to the relaxed/arcade masses. So I'll put it this way. It's like modelling a 1969 Dodge Charger down to every teeth on the gears in the tranny, the carb delay when the accelerator is floored and the effects ion traction from the limited slip differential  so that the thing performs exactly like it did in real life. Then tossing it into an arcade game like "Dukes of Hazard:The Boss Hog Hunt!" and neutering the hell out of it for <cough> playability.

 IMO, it's a bit of a slap in the face to "sims" like WB's and AH when they (EA) promote such an arcade (the future AW fm) as being so refined and complex.

 


Offline bloom25

  • Silver Member
  • ****
  • Posts: 1675
AW To get a new flight model
« Reply #3 on: April 14, 2001, 09:16:00 PM »
"Sandwiched between various Pentagon projects, these two luminaries worked with Dr. Kelton Flinn, original creator and programmer of Air Warrior. Stevens and Shaw developed a method that combined tabular aircraft data with some very sophisticated non-linear algorithms."

I like this part.  

I sure was nice of these 2 "luminaries" to find some time to develop these wonderful new FMs for them.  Those guys at the Pentagon can be SO demanding.  

" ...a method that combined tabular aircraft data with some very sophisticated non-linear algorithms ... "

Gee, I've never heard of GASP multiple linear regression, or HEAVEN FORBID, that nasty Taylor series approximation.    HT used to be an Electrical Engineer, I'd assume that AH probably uses something just  a little more complicated than "simple linear models."  

It's sure a good thing they came up with their own method for modeling the prop, since EA couldn't afford the proper software.  The statement that the models for the prop are not good enough made me spit a little Pepsi out of my nose too.  I personally use a program called Hspice to model the behavior of circuits.  One thing you learn is that if the model isn't good enough, you get another model. DUH!!  

The kicker is after all that work touting their new uber-FM they go and destroy it by saying they are adding an easy mode to take the challenge out.

I'm really wondering how they got the stick to fight back and subject the player to G forces.  I mean afterall, they said: "The results were amazing. The results were so accurate that the model even produced the correct "feel" of the plane as it flew through the air."

That article is good for a laugh at least.

 



------------------
bloom25
-MAW-
(Formerly of the)
THUNDERBIRDS

Offline Badboy

  • Silver Member
  • ****
  • Posts: 1226
AW To get a new flight model
« Reply #4 on: April 15, 2001, 05:16:00 AM »
 
Quote
Originally posted by Revvin:
Mind you I have to laugh at the 'secrecy' about this kinda reminds me of how Hotpants over at iEN kept telling us how everything had to be kept secret because once we eventually saw it it would be like the second coming..well Warbirds III beta is out and apart from a few small touches its the same old sim.

The ironic thing about the secrecy part is the fact that other developers are already years ahead of them. They are trying to do something that has been common practice for years, but don't tell them that, it can be our little secret  

Badboy
The Damned (est. 1988)
  • AH Training Corps - Retired
  • Air Warrior Trainer - Retired

Offline Seeker

  • Parolee
  • Gold Member
  • *****
  • Posts: 2653
AW To get a new flight model
« Reply #5 on: April 15, 2001, 06:59:00 AM »
One thing puzzles me, Badz.

I've always collected your PDF's avidly; and in them, you often state that AW "hit the numbers", the inference being that while AW was undoubtedly behind in terms of features, eye candy and development, the basic sim Kelton put together in the begining was still modelling A/C performance accurately.

Do you still hold to this? and if not, what's the basis of your latter conclusion?

Offline Revvin

  • Silver Member
  • ****
  • Posts: 1724
      • http://www.ch-hangar.com
AW To get a new flight model
« Reply #6 on: April 15, 2001, 07:01:00 AM »
I'd wager Maddox' MMPOG game based on IL2 will give plenty a run for their money, eye candy is'nt everything though so we'lljust have to wait and see. I'll be giving WW2OL a try but don't hold out much hope of it being much fun when some 14yr old armchair general tries to bark orders at me   Its an interesting concept but I think they are concentrating so much on the technology side and not put much thought into the game and the way it will run for example instead of a in-game mission planner ala Aces High they will have just a forum??!?! For anyone who's tried to play online team based games such as Counterstrike and Tribes 1&2 they will know that even with the relatively small numbers of players playing the game (up to around 16 for CS and 64 for Tribes 2) its VERY hard to find players who will play co-operative and not lone-wolf it and break down the whole idea of teamplay, I wonder how CRS are going to change this. Targetware have an exciting product and it will be a refreshing change to play a different era.

I still don't understand the animosity towards another sim, it happens here and on AGW and it amazes me how plays can turn being a customer to an online sim into a religion..a fanatical religion where they cannot bear to see mention of a competitor. If AW delivers what it promises in this press release the great for them and their community, even if its what we would consider a dumbed down flight model then what the hell! we don't play it and for those that want a little more they are spoilt for choice in the more realistic FM sims. AW has enticed alot of players who might never try a sim with its relaxed FM and alot ofthem go on to try sims like Warbirds and Aces High.

Offline Badboy

  • Silver Member
  • ****
  • Posts: 1226
AW To get a new flight model
« Reply #7 on: April 15, 2001, 10:21:00 AM »
Badboy
Quote
Originally posted by Seeker:


One thing puzzles me, Badz.


That's because you don't appear to have read the message... It is directed at false and misleading marketing claims and is not related to my previous work.

Badboy
The Damned (est. 1988)
  • AH Training Corps - Retired
  • Air Warrior Trainer - Retired

Offline Seeker

  • Parolee
  • Gold Member
  • *****
  • Posts: 2653
AW To get a new flight model
« Reply #8 on: April 15, 2001, 10:13:00 PM »
On reflection, you're right, I was asking the wrong question.

The article mentions look up tables. I thought the current trend was away from these?

Offline paintmaw

  • Copper Member
  • **
  • Posts: 216
AW To get a new flight model
« Reply #9 on: April 15, 2001, 11:48:00 PM »
AW3 is dead , bury it already , lol

Offline Badboy

  • Silver Member
  • ****
  • Posts: 1226
AW To get a new flight model
« Reply #10 on: April 16, 2001, 08:43:00 AM »
 
Quote
Originally posted by Seeker:
On reflection, you're right, I was asking the wrong question.

The article mentions look up tables. I thought the current trend was away from these?

I don’t think so. Some of the most recent flight sims are using them, both Falcon4 and MSCFS for example, though I hesitate to cite MSCFS, because it has some really weak models. Even so, there are some strong advantages to using look up tables. For example, if you want to model drag through the transonic region (for compressibility say) it isn’t good enough to figure drag as a function of angle attack alone, it needs to be considered as a function of both AoA and mach number. Now, supposing you actually have data from a manufacture’s wind tunnel, or flight tests that give you curves of wave drag coefficients, or other data that allows you to deduce them. The most sensible thing to do is to put the raw data into a look up table and allow your flight model to use it directly. Why is that sensible… Because a single relationship that accurately represents drag as a function of AoA and Mach number doesn’t exist (unless you can show me one). For a more tangible example, consider prop thrust. EA have already stated that the expensive prop add on wasn’t good enough for them (that still makes me laugh) but what is the most accurate way to predict the thrust produced by a real propeller? You mount it in front of an aircraft engine and nacelle complete with fuselage, wing roots and canopy and put it in a wind tunnel and test it to death. You end up with a great deal of tabular data that is most often presented in graphical form. Of course no sim developer would resort to that, so it is fortunate that it has already been done for almost every propeller, they are just difficult to find. Typically, the graphs will represent power and efficiency curves, but there are a lot of variations. Some give thrust and torque coefficients and others get even more exotic. But the point is this, if you really want to match the real world numbers, you need to work directly from the test data and that is most easily done with look up tables. Why, simply because you can take a whole bunch of curves of power coefficients against advance ratio for blade pitch variations from 0 to 60 degrees in 5 degree increments all on a single table… A simple, accurate and efficient representation of real propeller performance. There are a lot of simple analytical methods for predicting prop thrust, and due to the difficulty of finding real world data, many developers resort to them. But they are just an idealisation of the real thing, real world data in a look up table is the real thing  

Now, because look up tables are so efficient, developers who don’t have any real world data, might still use an aerodynamic package to do calculations for drag and lift coefficients, for example, and then put the results into look up tables that will be used by the sim. It is also fairly easy to fool the punters by matching some of the more commonly known values for top speed say. The point is that just because a flight sim uses look up tables doesn’t mean that it is good, and just because a flight sim doesn’t use them, doesn’t necessarily mean that it is bad. If you are estimating your data it is still possible, if you work hard at it, to match the performance of your model quite closely to real thing.

I have used all of the above techniques, in my own flight models, in some cases I use tables exclusively, in others I don’t use them at all, sometimes I combine those methods. Needless to say, none of this, or almost anything else related to flight modelling is a secret  

Badboy


The Damned (est. 1988)
  • AH Training Corps - Retired
  • Air Warrior Trainer - Retired

Offline Seeker

  • Parolee
  • Gold Member
  • *****
  • Posts: 2653
AW To get a new flight model
« Reply #11 on: April 16, 2001, 01:50:00 PM »
Thanks for the explanation.

I see your point about real data being the final ptoof; but that leads us further.

No one is doing any research on P-51's any more, all the data must be avilable by now.

Is there any reason why any modern P.C. flight sim shouldn't be able to match that real world data precisely, if it's merely a matter of entering a published and accepted value of X mph at Y alt, quite disregarding game programming such as reality settings, graphics or pilot aids?

If not, why do so few of them agree?