Hey Peter. Just getting around to installing FMs

[+Tailspin 3]

I'll admit I've been putting off installing your latest FM updates. Nothing to do with your work. Its my personal additions like engine armor, new German pilot, changing exaust emitters to fit the planes (I think the InlineExhaustEmitter makes a better Exhaust Emitter for radial engines and I don't care for the stock CleanExhaustEmitter effect for any aircraft ) that I have to add/change with every update. Anyway, when I was flying the SPAD 13 I noticed the aileron animation was "locked" in a right turn (just visually, not affecting control) and not responding to stick input. I searched the data.ini and found the following:

 

[LeftAileron]

 

(scroll down to)

 

ModelNodeName=Aileron_L

ReverseModelOrientation=TRUE

AnimationID=

 

[RightAileron]

 

(scroll down to)

 

ModelNodeName=Aileron_R

AnimationID=

 

 

There is no AnimationID assignment and the ReverseModelOrientation=True line should not be there. The fix should read like this:

 

[LeftAileron]

 

**********

 

ModelNodeName=Aileron_L

//ReverseModelOrientation=TRUE<------- Either comment out or delete this line.

AnimationID=1<----------------------------Add the number 1 here.

 

[RightAileron]

 

**********

 

ModelNodeName=Aileron_R

AnimationID=2<---------------------------Add the number 2 here.

 

Now I can get around to flying. Once again, THANKS for all your hard work.

[+peter01 2]

Thanks for pointing this out Tailspin - let me know if you find any others, trivial, major. or in-between.

 

I'll update some time down the track, as you point out its not easy for anyone if I keep updating, so won't be doing small fixes unless critical.

[Bandy 3]

This thread reminded me of an incident I had flying an SE5 a while back.

Since I have never seen any aileron lock-up like this before in the game, I took a screenie.

I thought perhaps it was some new damage modeling Peter introduced, but now I'm wondering if it was an animation bug as mentioned... so I'll also look into the data.ini...

PS: It wasn't a good day, but at least I made it home :blush2:

Edited September 17, 2007 by B Bandy RFC

[Bandy 3]

Didn't want to start a new thread, but thought I would include these pics of "zombie" aircraft [NB: no pilot in these AI aircraft] since it may be related to the new FM's as well. I think this was brought up before, but am unsure of whether it was discussed much, if at all...

This is not a case of mis-labeled pilot files (as in the Alb DVA thread), these AI-pilots were "killed" by other AI pilots, yet the airplanes kept flying, and flying, and flying... Not sure whether they could be "targeted" or not, I can't remember.

Edited September 17, 2007 by B Bandy RFC

[+peter01 2]

Don't know.

 

But are you using my latest version, in d/l section? I ended up standardising damage modelling in that for all planes.

 

Doesn't mean that nothing odd will be occurring, just that it is part of how planes are modelled and the game works, I guess.

[+Tailspin 3]

How did you standardize damage modeling? I think some planes (like some of the A-Team's) are set to certain codes so certain damaged sections will be displayed. I suspect thats what may be causing this:

 

 

Careful with that axe, Eugene.

[Guest capun]

Word of advise. You just cannot do a standard damage modeling referencing a visual mesh. I call that Visual modeling vs game physics modeling

 

First there are two levels of visual damage modeling. Let's use the AviatikC2 as an example

 

In the AviatikC2.ini you find this

 

[DamageTexture]

DamagedPostFix=_holes.tga

DestroyedPostFix=_Shredded.tga

 

These reference to the main textures damage versions. One is the bullet riddled textures, the other the shredded textures. They are just textures covering the meshes

 

Now in the AviatikC2_Data.ini you find this

 

[TopWingMidLeft]

ParentComponentName=TopWingLeft

SupportComponentName=BottomWingMidLeft

ModelNodeName=WingTopO_L

DestroyedNodeName=D_WingTopO_L

ShowFromCockpit=TRUE

DetachWhenDestroyed=TRUE

DamageRating=DISABLED

.

 

This section uses the D_WingTopO_L mesh which is a 3D Model of a destroyed Left Outer Wing Top section

 

Not all meshes have a 3D model that is why you see sometimes blank statements but if you have them in the 3D model you have to declare them otherwise they will show up during normal "good" flight. The statement acts like a "Hide Node until Main Node Destroyed"

 

It is critical to have the proper 3DS Max hierarchy set-up to have the Destroyed meshes replace the good ones, normally at the same hierarchy level as the "good" mesh. The OUT file should tell you how's the hierarchy set-up. Like this example. If the parent (WingTop_L) gets destroyed, all of the children meshes should get destroyed.

 

Level 1- WingTop_L (159 polys, 477 verts) 'Aviatik'

Level 2- WingTopO_L (62 polys, 186 verts) 'Aviatik'

Level 3- Wires01 (6 polys, 18 verts) 'Aviatik'

.

Level 3- WingStrut_L (132 polys, 396 verts) 'Aviatik'

.

Level 2- D_WingTopO_L (93 polys, 279 verts) 'Aviatik'

 

Shoot, I don't know how to do tabs in this forum.

[+peter01 2]

How did you standardize damage modeling?

 

As in, they are as they were originally. Not that I had intentionally changed things, but when you redo FMs, things creep in - so just tired to make sure it was as done in all FMs. For example, "destroyed" "disabled" "heavy" damage to different wing sections are as in all other FMs, including ones done by TK - if the inner portion of the wing is destroyed, the plane should be destroyed - its not evidently the case, but it should be if "destroyed" actually means destroyed.

 

And I asked Bandy which version he was using cos if I look at it, I want to know which one I should be looking at.... and of course, I'm not looking at previous versions I have done.

 

Also asked a while back - was this happening before?

Edited September 18, 2007 by peter01

[+Tailspin 3]

Hmmmmmm.....OK, thats weird. I was just thinking that some of the ATeam planes were done a little differently than the others.

 

ps..gotta run. Got a 9-ball game to go to.

[+p10ppy 2]

pretty sure the the "destroyed" "disabled" and "heavy" parametres are for the AI only and they dont really affect the FM

they are to let AI Attackers know the damage state of their target (i think, read it some where a while back... )

[Bandy 3]

But are you using my latest version, in d/l section? I ended up standardising damage modelling in that for all planes.

 

Sorry Peter for missing your question, been busy with our kitchen renovation project the last few days.

I'll look into what version of FM they are. Unless I've missed something the last couple of weeks, I think they are the latest but I could be wrong.

EDIT: your FM's "updated Sept 11" oops, may have missed that update, I'll double check...

 

As far as my memory serves, I've never seen "unattached" wing tips before, going back to Nov '06.

 

RE: the "zombie" pilotless AI A/C, I like them occurring in game, in fact, it adds that bit of the unexpected which I crave. The zombies aren't flying actively, they just drift along in a lazy circle, much like I imagine A/C with a dead pilot might behave on occasion when he hasn't slumped over onto the stick.

Edited September 19, 2007 by B Bandy RFC

[+peter01 2]

Spent some time looking at damage modelling, and for those interested, this is how I think most of it works. And its all fixable, but as all things in this game, there are some decisions/compromises.I may be wrong, its certainly a bit complex, you will need to read this carefully. Its a bit of an essay - sorry about that, but thought I should get these thoughts down.

 

So......

 

Firstly, damage modelling is completely different for the AI and the player. Its also done differently for different planes. And visual damage and expected "real" damage effects can be quite different for the player plane too (thanks for post on this capun). These differences are the basis for understanding how damage modelling works in the game, and sometimes the combination of these differences can be seen in one damaged plane at the same time - with some funny results.

 

The important difference is the AI versus Player difference.

 

On the AI side it should all be okay, "reasonably". The reason it should be okay is simply that a "DESTROYED" rating for a section does indeed mean destroyed. Its just a matter of ensuring that critical bits - inner wings, fuselage etc have a destroyed rating, and this is how they are done, by TK and all others. Without going into why this works, just want to point out it does - so the AI should be fine. It will be destroyed when it should be destroyed, if not, there is an error in FM that is fixable.

 

I say "reasonably" because of 3 qualifications:

 

Firstly, I did introduce errors by mistake or via experimentation in some planes, this may have confused things. All of my d/ls AFAIK should be now okay - tho they are beta still :)

 

Secondly, with improved AI (parameters), the AI will not be affected as much as the player when control surfaces are damaged or destroyed - ailerons, elevators and the like. This was always the case - the AI flies a simpler flight model and I think an easier damage model is applied too. It may just be a bit more obvious now, due to the improved AI. They will chase you around more aggressively, destroyed control surfaces or not.

 

Thirdly, how a destroyed AI (or player plane for that matter), plunges down depends largely on how the FM is done. Most should and do go straight down, some won't, may circle slowly etc. Changing this could involve a lot of redoing or it may be someting quite simple. As this isn't common, isn't such an issue, I'm uncertain what exactly causes this behavior, and requires an enormous amount of testing (for different damaged bits), I'm not going to worry about it. Its pretty trivial IMO.

 

The player side of damage modelling is different. There are some issues here with the game and other things. Fundamentally the player plane is not destroyed if a section rated as "DESTROYED" is actually destroyed, with the exception of a pilot or fuel tank hit (or the component these are housed in eg nose), and the engine (in a different way of course). This is definitely a problem, though may be done for good reasons - eg, to simulate the difficulty of controlling damaged planes, rather than have them just destroyed etc.

 

But damage modelling for the player plane can be improved/fixed, with some compromises.

 

As an example - the Dr1 Tailspin raised before, but its true of many planes - not all lose wings when damaged, as I said, different planes are done differently. In this plane, the player can still fly having say lost the entire top wing - the AI can't, it will be destroyed (a difference between AI and player damage modelling). The reason the player can still fly is that although some lift is lost - should be 1/3 approx - the plane will fly reasonably with 2/3 lift (or even a lot less). So how the FM is done is largely immaterial (the allocation of forces and lift to different sections etc) - actually, i don't think you can do much with it in the FM. Its the game mainly - its based on a jet sim (one wing gone - all lift lost!), and no coding is in place to mimic the effect of losing one of several wings in the plane. That is, there is no way of mimicking the visual damage effect physically, that I know of.

 

So you will get the situation of losing one wing, even in a biplane, and flying anyway. Obviously if you are in the plane and you lose a wing you'd expect it to be difficult to control - but thats not the case.

 

So how does TK do it? Because i haven't flown one of TKs planes with one wing missing - have you? Well, its the visual damage modelling.... you don't lose a wing with TKs planes, you may lose parts, but not one wing. The physical damage modelling is the same in TKs planes as all the others - when a part of a plane is "destroyed" even tho the part may or may not fall off (the latter in TKs planes), the plane is actually impaired (I think!), you lose lift and the like. What is different is the visual damage modelling.

 

So, the answer could be that you define parts in the FM of some planes so that wings do not become detached when destroyed. This would need to be done in this way, because as capun implies, if not done that way, there could be graphical glitches. But I didn't experiment with this, because in the end which ever way you do it, the aim of not have wings fall off when damaged is achieved, with no difference to anything else.

 

The compromises? Although you are doing this for player planes, AI planes are affected in the sense that their wings won't fall off either (they will still be impacted, flying will be impacted for both player and AI). And, well, to state the obvious, you don't get wings being shot off - maybe some/most like this visual effect.

 

The benefit - seems more realistic, not so annoying (it happens to me in some planes as the player all the time). Just as importantly you will not get "unattached wing sections" - for player or AI. The real cause of unattached wing sections is a bit more complex, and I won't go into that...mainly because I don't think its fixable from a FM perspective.

 

One last point - a plane that has a bit destroyed that is rated as "DESTROYED" will not be targeted by the AI - as in the AI flying the planes or the player's plane. This was always the case, But if as player wings don't fall off, even though your performance may be somewhat be effected only, the AI will think you are gone. This is true whether visually the bit/wing comes off or not - its just that if you don't actually lose the wing, its a bit more perplexing if this happens.

 

That last point also raises one last question to me that I haven't resolved - tho in some planes as the player I find I am no longer targeted by the AI when a section is I presume rated "DESTROYED" is destoyed (with wings becoming detached or not, depends as I said on the plane and the visual not real damage), I haven't come across this in TKs planes. Maybe I just haven't come across this - it doesn't happen all the time - or possibly TK has done something differently that eludes me presently.

 

At the moment I am trialling out not having any wings detached at all, and seems fine so far. So i might go down this path, as on balance this seems preferable to me.

 

Phew..... have to say, for a moddable game, there isn't much documented about all this!

Edited September 22, 2007 by peter01

[+Tailspin 3]

Thanks for sharing your thoughts, Peter. Just yesterday I started trying to redo the lift factors for the various wing components on a few planes. I'm not changing the total lift for the wing just redistributing the values more evenly among the different sections, trying to take into consideration the different proportions for upper and lower wings, etc. Whereas if you lose a forth of the top wing you lose a fourth of the lift factor for the entire wing. I want to see if this makes any difference in how a damaged plane flys. Initial results are interesting in that, undamaged, the planes don't seem to fly any differently from the player POV. Its going to take a lot of test flying though as its hard to get the AI to just shoot one wing off.

 

ps...this experiment raises a question. Why are the lift values usually set higher for the mid/wingtip sections? Is lift actually greater on the outer sections of a wing IRL?

[+peter01 2]

Yeah, thanks Alligator Devil, I need to fix the decal problems with the Avros and Martinsyde.

 

Tailspin, sounds like your really getting into it

 

Nearly all the absolute values in the FM vary according to tables in each section, including lift. The biggest impact in moving lift around however would be to stall behavior.

 

Allocation of lift needn't be in outer sections necessarily, tho it should have some relationship to the area (massfraction mainly) and centres of gravity. My understanding is that most force/lift allocation is only really critical for stall and especially spin behavior.

 

Re how it is actually in RL, don't know. But TKs initial release versions of the FMs are I believe an attempt to do things reasonably accurately - they may be the best guide we have. Of course, like all of us, TK modified things to make them work better and he redid them with patches. We all do this because in the end you are balancing so many things, its a fine art, and a matter of getting the right result. The flight models would be a lot different if there was no AI, and even no dogfighting - these two are major seperate considerations and there are many more to consider/balance.

 

The most challenging to me currently is making the planes feel more like WW1 planes (with a jet sim engine and other limitations) and making different planes "feel" different. Involves compromises too.

 

Cheers

Edited September 23, 2007 by peter01

[Guest Charles]

I need to dig out my reference books to be precise about this, but from recollection in RL most lift is produced by the outer section of a wing. Same with a prop (which is just a wing rotating at the front) - the point of maximum thrust is at the point about 75% out from the centre.

[+p10ppy 2]

The Martinsyde and Avro problems are not decals errors. (it looks like two textures beign drawed over the same wing)

If you delete the decals you will see the errors still there. (i'm an ignorant so i can't help)

 

it "looks" like the Damage nodes are displaying all the time on the Martinsyde and Avro

atleast ive seen something like that before when tinkering with damage nodes

[+peter01 2]

Thx, your right about the damage nodes.

 

Fixed, looks lot better, new FMs in zip file for Avro504C, Avro504D and Martinsyde. No other changes.

 

Fix_For_Avro_C_and_D_and_Martinsyde.zip

 

If you just want to edit the 3 files (and 1st and 3rd duplicate entries are not mistakes), replace the DestroyedNodeName= line in existing files in 3 wing sections as follows:

 

[bottomWingRight]

.

.

DestroyedNodeName=Wing lower Middle R D

.

.

 

[bottomWingMidLeft]

.

.

DestroyedNodeName=Wing Lower Middle L D

.

.

 

[bottomWingMidRight]

.

.

DestroyedNodeName=Wing Lower Middle R D

.

.

Edited September 23, 2007 by peter01

[+Tailspin 3]

I need to dig out my reference books to be precise about this, but from recollection in RL most lift is produced by the outer section of a wing. Same with a prop (which is just a wing rotating at the front) - the point of maximum thrust is at the point about 75% out from the centre.

 

Thanks Charles.

[Firecage 1]

Peter, I have been at work for a lot of hours lately and just got all the FMs installed adn I must say im very impressed at what you have done. Flying on hard still has a few planes that get stuck in spins but I think those are planes you have not worked on thus far. Your work is a major step and equallying out alot of FM vs Fm problems I saw a while back when I was playing alot more.

 

Keep up the good work.

[+peter01 2]

Thx guys. Glad you like them. I'm human too, and like some encouragement.

 

Just should say one thing.......won't say it again, think I have tried to point this out in past posts and in the Readmes, but just to be sure.

 

Its a lot of work (have i said this before ). Think I have done about 58 different FMs. BTW, we have 73 planes in the game by my count - thank you modellers, what an effort!.

 

About half or more of the 58 FMs are my own work. The rest are substantially original work by TK (except DFWC5 and AlbDva) and Charles (about 10 of the Ateams, more or less, my own entirely) . I made changes to all, yep. Sometimes minor, sometimes very major, in performance, feel, ai etc. But without a doubt the hard work, basic feel and flyability was done by them for these......and they are very good too - otherwise I would have been tempted to redo them .

 

In a way, its a team effort. We all certainly learn from each other. Even reuse others work.

 

Charles has said to me when I was starting, the more people that do FMs, the better. He is right, and FMs can always be improved.

Edited September 26, 2007 by peter01