Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Pursuivant]

I've done a fair bit of research into bugs in the damage models for existing planes in the stock game, to the point of making a spreadsheet of all the DM problems for planes up to the 4.12.2 release.

Information and spreadsheet here:

http://forum.1cpublishing.eu/showthread.php?t=229446

It's now clear to me that DT, or at least certain members of DT, have no interest in fixing these problems. Yet another reason to move to mods!

While Damage Modeling is one of the most contentious issues in the game, and there are no firm metrics for how to do it, I think that we can all agree that bad damage modeling gives certain planes unfair advantages over others.

For example, the landing gear on the Stuka and D3A "Val" is modeled - you can hit the gear and break it off. But, the landing gear on the R-5 and U-2 isn't modeled and can't be hit or broken off, despite being smaller and more fragile. That means that planes like the U-2 can land safely with landing gear damage while planes like the Ju-87 can't.

Additionally, there are systemic problems with DM in IL2, which an enterprising modder, or mod-pack maker could look into.

* Central wing sections on certain biplanes are combined with fuselage sections.

* "Hooks" to break control surfaces are too big in some planes, too small in others.

* Hooks to start fires or smoke effects are misplaced on certain planes.

* Wing struts aren't modeled and can't damaged or broken - and thus can't cause wing failure - for most planes in the game. Wing struts aren't just for show, they're integral parts of the wing and breaking them can be as catastrophic as breaking a wing's main spar.

* Most planes have control surfaces which are far too easily damaged. In many cases, a single .50 caliber bullet hit will trigger damage textures in a control surface, when realistically all it would do is make a thumb or fist-sized hole which has little effect given the relative size of the control surface.

* Most planes have parts which are far too easily damaged, for the same reasons described above. In particular, the wings of a big, tough plane like the B-17 Flying Fortress or the geodesic-frame fuselage of the Wellington should be able to take a lot more damage than they currently can before even light damage textures show.

* For some planes, there is no consistency in how much damage is required to break a particular part. For example, it might take twice as much damage to remove a given plane's vertical stabilizer as to remove one of its horizontal stablizers, even if they're both about the same size and constructed in a similar fashion.

* For many planes, the amount of damage required to break the fuselage is far lower than it should be. For example, the C-47 - an airframe noted for its ruggedness - has an extremely fragile fuselage in the game.

* In some planes, due to bad DM and hook placement, hits to certain parts don't do damage, or do damage to different parts of the plane.

* In some planes, destruction of a wing makes the entire wing fall to pieces, rather than just having the wing break at the point of the damage.

* In some planes, damage textures or choices for break points in the wings aren't done correctly, so that the plane is "destroyed" when realistically it should still be able to fly. For example, tough birds like the F4U Corsair, P-47 Thunderbolt, or TBF Avenger could (barely) fly back to base with a third of the outboard wing missing, yet they crash when they suffer this much damage in the game.

* In some planes, the degree of damage required to damage or destroy an engine depends on the plane, rather than the engine (ignoring factors like oil, coolant, turbochargers, etc.). For example, the P&W R-2800 engine takes damage differently based on whether it's in a F6F Hellcat, F4U Corsair, or P-47 Thunderbolt.

* In many cases, fuel tank leaks never stop, even with self-sealing fuel tanks, or planes with multiple fuel tanks. This means that a single fuel leak will eventually drain all the plane's tanks! In all cases, a fuel leak assumes that it's at the bottom of a fuel tank, so all the remaining fuel is lost.

* Self-sealing fuel tanks aren't nearly effective enough. Realistically, the better ones could survive and fully seal after being hit by several .50 caliber bullets or a 20mm shell. Currently, a single incendiary bullet is sufficient to ignite even a self-sealing fuel tank, when realistically, liquid gasoline doesn't ignite easily (vs. gasoline vapor which is highly inflammable) and a heat source isn't going to ignite combustibles in the absence of oxygen (like being submerged in gasoline).

* While it appears that TD has corrected the problem, historically it was quite easy to destroy the control cables of any plane in the game.

* For some planes, damage textures are unrealistic and ugly. Bf-109 series, I'm looking at you.

* There's no consistency to cockpit damage textures, meaning that cockpit damage to some planes will always knock out certain instruments, but not others. In some planes, cockpit damage will wipe out your gunsight, in other planes the sight is immortal. In some planes, oil on the windscreen leaves you blind, while in other planes you still have some visibility.

Going forward, a modded version of IL2, or a flight simulator based on IL2 could look into the following damage modeling issues.

* It should be possible to model other aircraft systems - like turbochargers/superchargers, engine boost systems, oxygen tanks, radios, bomb sights, radar, and batteries/generators.

* It should also be possible to model control cable runs and accurately simulate damage to those systems - like jammed or fluttering control surfaces, or controls with limited range of motion.

* Damage should have the ability to jam canopies, escape hatches, bomb bay doors, and turrets.

* Damage to guns should damage the gun. Damage to ammo in ammunition runs should destroy some portion of ammunition and/or eventually cause a gun jam.

* Damage to gun turrets could prevent guns from elevating or turning, knock out the gun sight and/or interfere with the gunner's visibility (i.e., bullet hit which shatters armor glass).

* Damage to hydraulics or electrical systems could prevent some turrets from being raised or lowered, or from turning or elevating guns.

* Damage to electrical systems could prevent guns from being charged or fired, or ordinance from being armed or launched.

* It should be possible to model fuel transfer and shut-off systems, even without modeling dynamic center of gravity.

* Scrap "generic" damage textures and replace them with "damage decals" which appear at the exact point where a missile hits, and which show the sort of damage that missile does to the material it hits. For example, a bullet punching through doped canvas should leave a small hole, but it should leave a bigger hole in wood - with a splintered wood texture around it, and possibly an even bigger hole in metal (assuming it tumbles) with a classic bullet hole in metal texture. Shrapnel form flak should leave irregularly-shaped and sized holes. Explosive rounds should have some chance of not detonating when they go through canvas, but should leave big holes in the surrounding canvas if they hit something solid, otherwise, they should produce big jagged holes.

* Fires should have the ability to spread, and should have the progressive ability to consume fuel and oil, and weaken or even burn structural materials.

* Explosions produced when a plane crashes should be based on the amount of fuel it has aboard.

[SAS~Malone]

err... why not ask someone to 'just' make a whole new flight sim...?
that is, kind of, what you're asking here, if you think about it....

[SAS~Ghost129er]

err... why not ask someone to 'just' make a whole new flight sim...?
that is, kind of, what you're asking here, if you think about it....

I was going to say something more or less among those lines - I didn't because I thought it'd be rude but turns out I was on point after all.

[vpmedia]

Some really good ideas, but I can't tell which ones are feasible.

Potential sources of problems:
-the lack of access to the games full source code
-the game engine being capable to handle the extra calculations required for the new features.

Furthermore you would need to make java and 3d changes to a few hundred aircraft types, which could take years.

[SAS~GJE52]

I fully agree with what you say, indeed it has been said here many, many, many times before.

The problem is just the sheer size of the task you have laid out in your request....

With the huge number of aircraft in the game now, as well as the different "standards" that the original models were subjected to, it would indeed be a huge task. There are also a large number of Frankenplanes available which when hit do any number of odd conversions ....  (although I guess all mods are actually frankenplanes to some degree or other  )

I am not adopting the moral high ground here, my first efforts were simple hier conversions ....    in the early days it was considered to be the norm, but we all start somewhere. Not everyone is a 3d master from day one ...... 

There are a lot of non-modders (and some modders for that matter  ) who perhaps don't understand the complexity of the physical 3d requirements to integrate an acceptable - and fully playable - aircraft, ship, etc into the game.

I will probably upset the coding gurus here but IMVHO   , although their "behind the scenes" improvements deserve the highest respect,  the 3d work is not just "eye candy". A good 3d model and all of the LODs etc  will take many hundreds of hours longer to create than any well crafted piece of coding or accurate flight model.... and remember it is the visual appearance that draws players in. It is the shop window so to speak. Therefore damaged aircraft, etc must also look good.

What I am saying here is that attempting to do all aircraft is just too big a task for our limited number of 3d guys. Some older models occasionally  get a long overdue rework, but there are just too many to even begin to scratch the surface.

So ......... sad to say but I am afraid this request falls into the " too difficult" category ......

I am not qualified to comment on some of the recommended coding changes, however, when it comes to model creation, basing new aircraft on already game compliant models is a good way to go forward. Creating a slot for a new single engine jet new model -using say the P-80- and replacing meshes "like for like" as the model is developed helps to ensure all of the required hit boxes, hooks, LODs etc are in place.  There are enough "standard" types to be used for this approach. Intelligent use of higher poly modelling where necessary helps address the poly limits and well thought out LOD distances reduces the graphics load.  At the end of the day it is down to individual effort..... and good practice.

"He who pays the piper, calls the tune"

so.. unless you want to fund model creation ...... you could always pitch in and try to become part of the solution ...... 

G;



ps .. I wish I had a $ for every time I have seen someone say," This is what we need .. and I would love to help of course, but I don't have the time" .... 


 

[Pursuivant]

Points taken.

The list of problems in the first section of requests is actual bugs and/or crummy modeling. The list in the second section is pie-in-the-sky stuff best left to a new flight sim.

But, while this request is huge, I'd ask that the moderators not delete it.

I've identified the vast majority of the DM problems with the stock planes in my spreadsheet. That represents a unique source of information, which players and modders alike will find useful.

Also, no one modder needs to undertake all the suggestions on this list. Instead, small, incremental projects over time are the better strategy. After all, in the last 10 years or so, we've gotten literally hundreds of 3d upgrades to stock aircraft, all done by different people, which vastly improve the game. So, why not have similar effort applied to the less visible aspects of the game, like damage modeling?

Additionally, many fixes to existing Damage Models are small, potentially simple projects which could easily be attempted by novices, or done by more experienced modders who lack the time to work on a larger project.

Fixing ugly damage textures would be a worthy project for a talented skinner. And, what would make the game better at this point: good-looking damage textures for the Bf-109 series or yet another Bf-109 skin?

I'm also more than willing to do work to improve the game and can dedicate some time to it, but lack sufficient understanding of how it works to make any of the fixes I recommend, except perhaps for improved damage textures. Point me to a tutorial on making 3d damage models or assigning hooks and I could possibly do some simple work to fix things.

[Pursuivant]

Some really good ideas, but I can't tell which ones are feasible.

Can easily be solved with graphics design work:

Code: [Select]

* Ugly or unrealistic damage textures. Should be no more difficult than making a skin. Start with the Bf-109, Me-321, and Me-323.

* Inconsistent oil on windscreen effects.

* Improved cockpit damage effects, such as damage textures for various instruments.

Can easily be solved with 3d work:

Code: [Select]

* Collision model (or whatever the game uses to determine if a bullet hits a plane) differs from LoD0 (detailed visual) model. (Simple solution, use LoD1 or equivalent slightly simplified visual model for Collision model. It would play hell with FPS, though.)

* Breaking parts placed in unrealistic locations. "Weld" existing parts which aren't modeled correctly, make "cuts" for new breaking parts at more realistic locations on the aircraft. With care, revised break locations could retain existing end-cap textures from stock skins.

* Adding new breaking parts. With care, you could use existing textures from damage texture or skin for new end caps.

* Breakable gun sights and bomb sights for all aircraft = simple tweaks to existing 3d mesh for these items.

Can be solved with coding work (potentially quite simple coding work, if you might just have to change a single value):

Code: [Select]

* Misplaced hooks.

* Unrealistically vulnerable or invulnerable parts.

* Adding breakable gun sights/bomb sights.

* Randomizing gauges damaged due to cockpit damage.

* Altering hooks so that breaking one section of the wing doesn't break all sections outboard from it. For example, a wing broken from a plane at the wing root shouldn't immediately break into 3 pieces, but should stay in one piece as it falls.

* Altering fatal damage effect of small planes (like Fi-156 or U-2) so that they don't explode as violently.

Coding work which probably requires more time but which is probably possible in the game:

Code: [Select]

* Adding new damage options, like chance of control surfaces being jammed, fluttering, etc., or adding chance of supercharger/turbocharger damage.

* Reworking generalized damage model for self-sealing fuel tanks so that they're a bit tougher, have a better chance of stopping leaks, and aren't likely to catch on fire if they're just hit by one bullet at a time.

* Reworking existing fuel loss algorithm so that there's a chance that a fuel tank hit will only result in a loss of a certain amount of fuel, not all of it.

* Rework fire effects so that a fire aboard an aircraft will eventually cause damage to the section where the fire occurs, and possibly adjacent sections. (For example, a bad engine fire in a wing-mounted engine will eventually weaken or burn the wing behind it, and heat from the blaze might even burn the fabric off of horizontal stabilizers placed directly behind the engine).

* Alter animation coding for breaking parts so that parts broken from an aircraft when it's on the ground won't "bounce" or otherwise move in an unrealistic fashion.

* Alter crash-landing algorithm so that crew can be injured or killed in a hard landing, even if the plane doesn't break or explode. (For example, the Akutan Zero was captured because the pilot broke his neck when the plane flipped over, although the plane was easily salvageable.)


[vpmedia]

Nothing is easy if you have to make changes to a large number of planes. Testing also take a lot of time. How do you test for example the damage to a specific instrument? Fly around and wait until it gets damaged by a well placed shot?

[SAS~GJE52]

  Also, no one modder needs to undertake all the suggestions on this list. Instead, small, incremental projects over time are the better strategy. After all, in the last 10 years or so, we've gotten literally hundreds of 3d upgrades to stock aircraft, all done by different people, which vastly improve the game. So, why not have similar effort applied to the less visible aspects of the game, like damage modeling?   

Small incremental projects are going on all the time ...... they are small because;

1 There are only a few guys who have the skills to do it
2 Those that have the skills like to do what they want .... This is not a software/gaming house, it is a hobby.

Please do not start dictating to the modding community what they should be doing ... because that is the moment the few skilled guys that are left will also leave.


As I have said before, posting lists and directives is actually counter productive.

I'm also more than willing to do work to improve the game and can dedicate some time to it, but lack sufficient understanding of how it works to make any of the fixes I recommend, except perhaps for improved damage textures. 

Perhaps you could outline what else you will bring to the party. You have not said what your total involvement in this "mega project" will be, other than to create lists and correct some textures. If you only intend to do that and expect others to give up their free time to make things happen ... then you are going to be sadly disappointed. There required 3d and code changes require far more effort than a bit of "photoshopping" ... that is the realty of the situation.

If you really are interested then make a start yourself, tell us what you are doing and what progress you have made.  Once you have done that then this thread starts to become relevant.


If as I strongly suspect this is only another "wish list", then it says nothing more than what has been known for a long time.

But, while this request is huge, I'd ask that the moderators not delete it.   

I beg to differ, for the very reason you give, I think they should... 

G;

[vpmedia]

If I want to do new damage textures you need to do damage1o and damage2o files for light and heavy damage + default skin modifications (summer,winter,eto,pto,desert,mto,cbi) and also do a "bit" of testing, let's say two days per plane.

We got around 1000 planes in CUP, but of course some share a common damage texture and we only have to do half, thats almost three years of work. If you want to do realistic and historic damage textures, then you need to do a lot of additional research, which could easily add a few more years to this project.

[SAS~Ghost129er]

This sort of a request should be stickied in the 'DO NOT ASK' section if I may say so.

[hello]

I Like the request. I also like the listing of problems. It is not like someone has to do it, but if you happen to work on a particular plane and is has a problem, why not fix it?