Free Falcon 5.5

[element1108]

I changed my nvidia settings in control panel to "performance over quality" (slid the bar full left) that seemed to fix the nvidia black scren of death. That was for AF, not sure how ff5 will work, will download it when I get home.

 

Anybody play it yet? Impressions?

[RvEFuSiOn]

Woohoo!, Gj my friend Jaeger and the entire Free Falcon team!

Appreciate your hard work as always. I plan to enjoy many countless hours of fun thanks to you fellas!

 

 

For those who want to see a few screenshots of Freefalcon5 check this out

 

http://www.freefalcon.com/index.php?page=screenshot

 

Peace

[Fahhh]

it work under win 7 64 bit and ATI?

last time I tried I had nvidia and was a no no... have never got into Falcon but would like to give it a whirl.... hang on a min ... another Flight Sim to learn? ..... ggggggoooo on then :)

 

It works quite well here - Win 7 x64 on an ATi HD4830 using Catalyst 10.4 drivers. For the half an hour I played, I saw no glitches or problems.

[asparagin]

^^ then it's mostly about nvidia: ati should pay those guys for making an ati exclusive title.

[Pilotasso]

Looking at screenshots, aside for new lighting, 3D models are still blocky, washed out textures (look at the carrier shots, they digitized the textures directly from photos to the game with maintainer personel pasted onto the walls and all! :D ) it all looks and should still feel DOS like just as the other iterations did. Without advanced fight model-- the real novelty to come for the series-- as in BMS 5 (or is it 4?) this one is just more of the same Im afraid.

 

yes thats right, "the avionics, the avionics!" I cant have fun just by pressing buttons, I need more to live. And this aint it. Been there donne that. :)

Edited May 11, 2010 by Pilotasso

[asparagin]

Looking at screenshots, aside for new lighting, 3D models are still blocky, washed out textures (look at the carrier shots, they digitized the textures directly from photos to the game with maintainer personel pasted onto the walls and all! :D ) it all looks and should still feel DOS like just as the other iterations did. Without advanced fight model-- the real novelty to come for the series-- as in BMS 5 (or is it 4?) this one is just more of the same Im afraid.

 

yes thats right, "the avionics, the avionics!" I cant have fun just by pressing buttons, I need more to live. And this aint it. Been there donne that. :)

 

Well the graphics has never been an advantage for the falcon series (except maybe in the 90's). But you say the flight model isn't accurate? You mean FF, or also for others in the series?

[element1108]

I have to say this is my first FF5 experience and I'm really happy with the final version! I played AF a bit, but was hindered graphically by the NVIDIA problems so I couldn't choose my ORD I had a hard time configuring my hotas because of screen glitches. This one is crisp and clean and I can read/view everything to it's highest fidelity NO problems!

 

Well done to the team! Now to start the MEGA process of learning (pretty much from square one)...I should be ready to fly online in a few years or so ;).

[Pilotasso]

Well the graphics has never been an advantage for the falcon series (except maybe in the 90's). But you say the flight model isn't accurate? You mean FF, or also for others in the series?

 

 

Yes flight model is pretty innacurate (and sterile) in some regimes, specialy T/W ratio, slow speed manuevers, landings and any other manuevers where wing laminar flow lift is not applicable. My brother is a fighter pilot and he confirmed it to me (infact he told me some military SIM's he tried pretty much suffer the same, as they are mostly avionics and procedures orientated). Thats why I deposit great hopes for BMS-whatever-its-name-is.

 

They can argue "but the perfomance charts this and that!" all day long, as much as they want, but to me-- and based on my engeneering point of view -- its impossible to reproduce accuratly aircraft flight by running mere interpolations from known points (and this is key: constant motion values, never transient) instead of full blown aerodynamics, inertial momentum, and force balance equations. All this is missing from the infamous data tables, that people stuburnly keep mentioning to justify the paper plane model on rails thing, many of them dont grasp any engeneering principles at all.

Edited May 11, 2010 by Pilotasso

[asparagin]

Ah yes the famous flight on rails.

[jspec3]

element1108, would you mind letting me know what OS/Graphics card you have?

[tflash]

Well, while FF5 obviously didn't work on my current PC (a very standard Win7/NVidia/widescreen system); I guess this "new reference in flightsims" might actually work on my former former PC (now in the cellar) which had XP (remember?).

 

I guess FreeFalcon is there for all the good memories. Nevertheless the people working hard to keep this project alive deserve much credit.

[element1108]

element1108, would you mind letting me know what OS/Graphics card you have?

 

 

 

Nvidia 9600GTM... Upgraded the latest drivers. I tried FF5.3 before updated drivers but it was even more effed up than AF.

 

I should also add I'm in XP no Vista oR 7

Edited May 11, 2010 by element1108

[jspec3]

ok I see - thanks.

[SUBS17]

All this is missing from the infamous data tables, that people stuburnly keep mentioning to justify the paper plane model on rails thing, many of them dont grasp any engeneering principles at all.

 

No what you don't understand is Falcon 4 is quite an old sim and thats why it behaves the way it does. But with F4AF and OF thankfully people have improved the sim in some areas at least it is fun to fly. I think perhaps you should give the VRS Superhornet a spin especially the pro version as that will have the most complex FM ever made for a FBW jet for a PC.

[element1108]

I am having issues configuring my X52 joystick for FF5. Seems to not want to stay centered and constantly drifting to the right. Tried a number of configs as well. Anybody have a profile I could use?

[Pilotasso]

No what you don't understand is Falcon 4 is quite an old sim and thats why it behaves the way it does. But with F4AF and OF thankfully people have improved the sim in some areas at least it is fun to fly. I think perhaps you should give the VRS Superhornet a spin especially the pro version as that will have the most complex FM ever made for a FBW jet for a PC.

 

 

I do understant, just step onto their forums and say that, they'll swear by it TODAY.

They dont realize or ignore by bias that interpolating flight model from points of constant motion (snapshots of constant motion) into a dynamic motion modeling violates a fundamental law of classic mechanics: F=Mxa. :)

 

Thus why there is a big difference betwee data tables and a dynamic flight model. Its not just because the later is more complicated. ;)

Edited May 12, 2010 by Pilotasso

[asparagin]

interpolating flight model from points of constant motion (snapshots of constant motion) into a dynamic motion modeling violates a fundamental law of classic mechanics: F=Mxa. :)

 

Ok, wait a minute, now letting letting the "flight simulation" example by side, there are different ways to describe a phenomenon:

 

1) the best is by finding the right formula(s) (this would be in this case the dynamic motion modeling): this is often in science impossible, or possible but comes with an huge amount of computation that no one needs, when:

2) Approximations are used in a much simpler formula than 1. This simplified formulas, maybe don't work on the entire spectrum but are accurate in their range and easy to use.

3) would be taking the reference points which you get by measurement: but in science this wouldn't count because it doesn't make any affirmation about the phenomenon.

 

So saying that their points violate a fundamental law, I would not consider as a good argument, because unless you have the "ultimate" formula, which I doubt that any simulation on the market uses, all the flight models used violate some laws at a certain point.

 

From the options 1-3 you will nevertheless notice that I totally agree on the accuracy with you though.

 

A.

[Pilotasso]

This is not a question of aproximations (because they are inevitable), its a question of starting off with the wrong leg.

 

Let me give you an example as an anolgy:

 

Spring force/elongation modeling (should be simple).

 

Imagine you have an equation F=Kxd where F is force, K, the springs constant and d the elongation. Then you make a data table showing elongation for each different force aplied.

 

You obtain snapshots of static load.

 

 

Next imagine you want validate the model by subjecting the springs under load. You load the spring and tadaaaaa it oscilates, not only up and down past the static equilibrium, but to the sides and with hystersis phenomenae should you take it to the edges of its capacity for long periods. You suddently realize theres alot to be said further than the intial snapshot data file into a makeshift simulation.

 

 

in panic you try insert scripts to cover up the black holes doing it easy and fast but it wont be convincing. This is what happens with falcon 4 (hope to see it done dynamicaly in BMS4).

 

I made a simulation in excell for car suspension and I did a static snapshot versus real time dynamic simulation (by aproximations) and there are acute differences. I took 8 months developing the dynamic simulation and Im sensitive to this subject.

 

As a matter of curiosity I simulated taking time units of a milisecond and had a few compromises in the matter of resolving more variables than equations but i found that it resembled a real system by less than 10% difference in the extremes of the envelope.

Edited May 12, 2010 by Pilotasso

[RvEYoda]

This is not a question of aproximations (because they are inevitable), its a question of starting off with the wrong leg.

 

Let me give you an example as an anolgy:

 

Spring force/elongation modeling (should be simple).

 

Imagine you have an equation F=Kxd where F is force, K, the springs constant and d the elongation. Then you make a data table showing elongation for each different force aplied.

 

You obtain snapshots of static load.

 

 

Next imagine you want validate the model by subjecting the springs under load. You load the spring and tadaaaaa it oscilates, not only up and down past the static equilibrium, but to the sides and with hystersis phenomenae should you take it to the edges of its capacity for long periods. You suddently realize theres alot to be said further than the intial snapshot data file into a makeshift simulation.

 

 

in panic you try insert scripts to cover up the black holes doing it easy and fast but it wont be convincing. This is what happens with falcon 4 (hope to see it done dynamicaly in BMS4).

 

I made a simulation in excell for car suspension and I did a static snapshot versus real time dynamic simulation (by aproximations) and there are acute differences. I took 8 months developing the dynamic simulation and Im sensitive to this subject.

 

As a matter of curiosity I simulated taking time units of a milisecond and had a few compromises in the matter of resolving more variables than equations but i found that it resembled a real system by less than 10% difference in the extremes of the envelope.

 

You hit the nail there I believe :)

 

My opinion on this has gone from one to the other about 4 times so far after reading this :p.

I'm thinking along the lines of non-conservative systems, in the meaning that the

"consumption" to reach a state is not only defined by the state itself, but also the

path taken to get there. Use a different path and you may get to the same end state

of a certain variable, while leaving others out.

 

However some situations I believe are well represented by snapshots, especially when

dealing with non-deformed objects, such as simple undamaged aircraft control surfaces.

 

I am not sure if the drag-lift tables are good or bad for a simple flight model. We need to

find some specific examples of where it fails in the case of the aircraft.

 

I believe that IF the model can be well expressed as a function

of current state F = F(x_0, x_1....x_n) (Function F of state variables x_0 to x_n),

then it can be fully expressed as a multidimensional interpolated matrix, but, hehe, I think

it may very quickly get out of hand when dealing with more than 2-3 variables :)

(Falcon uses for example Cd,Cl = F(AoA,Mach))

 

Like you say some situations (like your spring example with memory, nice!) cannot be

explained well by the current state at all (hence the if above), where

dynamic is the only real option.

 

Maybe we can find a third reason why it is a problem with these lookup-tables, but so far

this is a very good argument. Though we must wait for AFM to really be able to get

something significantly better out of fighters.

 

BMS5 is dynamic, yes. (close to AFM)

Edited May 12, 2010 by =RvE=Yoda

[Steel Jaw]

FF IMHO is an unstable mess.

 

I have much higher hopes (and expectations) for BMS.

[Case]

Interesting discussion.

 

I'm curious though about the nomenclature, cause I haven't heard anyone talk about solving differential equations, which is what I think should be happening, and the assumptions or simplifications that are made to do the simulations in realtime.

[Pilotasso]

You hit the nail there I believe :)

 

My opinion on this has gone from one to the other about 4 times so far after reading this :p.

I'm thinking along the lines of non-conservative systems, in the meaning that the

"consumption" to reach a state is not only defined by the state itself, but also the

path taken to get there. Use a different path and you may get to the same end state

of a certain variable, while leaving others out.

 

However some situations I believe are well represented by snapshots, especially when

dealing with non-deformed objects, such as simple undamaged aircraft control surfaces.

 

I am not sure if the drag-lift tables are good or bad for a simple flight model. We need to

find some specific examples of where it fails in the case of the aircraft.

 

I believe that IF the model can be well expressed as a function

of current state F = F(x_0, x_1....x_n) (Function F of state variables x_0 to x_n),

then it can be fully expressed as a multidimensional interpolated matrix, but, hehe, I think

it may very quickly get out of hand when dealing with more than 2-3 variables :)

(Falcon uses for example Cd,Cl = F(AoA,Mach))

 

Like you say some situations (like your spring example with memory, nice!) cannot be

explained well by the current state at all (hence the if above), where

dynamic is the only real option.

 

Maybe we can find a third reason why it is a problem with these lookup-tables, but so far

this is a very good argument. Though we must wait for AFM to really be able to get

something significantly better out of fighters.

 

BMS5 is dynamic, yes. (close to AFM)

 

Those data tables factors can be aplied to drag and lift equations but only for constant motion. In any given time you change the sticks position falcon 4 rapes mr. Newtons phisics fundamental law. :D

 

And let me add that I suspect it also doesnt apply when you do certain maneuvers such as knife passes. The perfomace charts are typicaly applicable when yout lift vector is perpendicular to the wing surface (i.e. pitch induced manuevers or simple straight flight).

Same for boundary layer for divergent conditions (stall and high AOA) etc etc etc.

 

While this can be used to see basic turning perfomace differences in aircraft it fails to translate to actual flight in every thru its entire envelope in a credible way.

Edited May 12, 2010 by Pilotasso

[sobek]

I'm curious though about the nomenclature, cause I haven't heard anyone talk about solving differential equations, which is what I think should be happening, and the assumptions or simplifications that are made to do the simulations in realtime.

 

The dynamic method automatically implies solving the differential equations or at least approximating them where they become too complicated. The difference is, one approach works with a lookup table, the other calculates exactly what the abstracted model does at a given moment (edit: and can be made to also work in transient conditions, like pilotasso mentioned). The table implementation was necessary at times when real time calculation would have cashed too many FPS ,sacrificing accuracy of behaviour on the way. Nowadays the table has turned, in terms of processing power.

[RvEYoda]

Interesting discussion.

 

I'm curious though about the nomenclature, cause I haven't heard anyone talk about solving differential equations, which is what I think should be happening, and the assumptions or simplifications that are made to do the simulations in realtime.

 

For a flight simulation I think you rarely need to analyticly "solve them" (perhaps for

computer control reasons such as FBW or intercept vector/inertial guidance), only

formulate them, and see where they lead. Just use some step model to see

"what happened next".

 

Perhaps I'm wrong, I never studied fluid/aerodynamic. :P

 

 

Those data tables factors can be aplied to drag and lift equations but only for constant motion. In any given time you change the sticks position falcon 4 rapes mr. Newtons phisics fundamental law. :D

 

Yes but FC's SFM basically works the same way with a few tricks there as well to help.

The hard coded SFM has other issues such as the auto trim keeping flight direction constant

instead of nose direction, creating a excessive bouncing nose effect for F-15 etc.

They both have issues, and are at about the same level of complexity.

Edited May 12, 2010 by =RvE=Yoda

[Pilotasso]

Interesting discussion.

 

I'm curious though about the nomenclature, cause I haven't heard anyone talk about solving differential equations, which is what I think should be happening, and the assumptions or simplifications that are made to do the simulations in realtime.

 

I confess I was weak with differential equations, Not only that, it also destroyed CPU resources when running them for my projects (remember MATLAB? yuck). They actualy are harder to compute despite their simpler apearance.

 

So I avoided them as much as I could. I preferred other methods such as instantaneuous motion analysis, balances and others. I often used excell instead of matlab and solidworks and such. I could see results and correct mistakes right away instead losing time with brute force debugging of differential equations and collision errors in those programs.

Edited May 12, 2010 by Pilotasso