[FltLt HardBall]

I have collected a lot of modelling info over the years and saved it for posterity. I figured I'd better post the stuff here so that it may prove to be of benefit to those interested in 3D modelling for IL-2, and perhaps a greater understanding of the process will help to unlock some of the secrets.

All of this is the work of other people; I'm going to reproduce it here without their permission, but if you are the creator of any of this work and you want me to remove it, just let me know (although it was originally released in the spirit of educating the community, so I'm sure nobody will mind).

Also, I have an original .3ds (3DMax) file of an IL-2, and a truck model. PM me if you'd like me to send it to you. Obviously you will need the 3d Studio Max program (version 3, 4, or 5) installed on your computer to be able to view it.

Mods: There's a lot of stuff here, perhaps we could make a dedicated "Tutorials" section, otherwise this will be a very long post...


---------------------------

Here goes.

First off, a post from Freycinet at SimHQ back in September 2005 on LODs (Level Of Detail):

"Basicly we can't "draw" the plane and then have it magically transformed to 3D model. It's done 3D all the time. We can have 2D drawings as guides. When model is viewed in wireframe mode without prespective from top, left, right, etc. it should match the 2D blueprints. Basicly modeling is just placing vertices and polygons in 3D space and using different primitives and tools to make it easier.

As you may know, planes in this game are not just one model. This is the cold reality that hit me to the face back in 2002 when I was doing the H75. Oh fok... I gotta do about 20 different planes.


So please don't whine to Oleg that why plane x wasn't included?


LODs (level of detail) explained:



Some LODs in wireframe


And of course all models must be cut to pieces correctly so that the plane breaks down nicely.




So in Fokker's case, this means 505 different parts, 37635 polygons and 25852 vertices. I must say that making cockpit is much easier, but I know that Gibbage doesn't agree with me


---------------------------


Luthier's modelling overview from IL-2 Center going back to 2003:

The Model
Models are accepted in a form of 3D Max files (*.max), from both 2.0 and 3.0. If you're using non-standard plug-ins that we might not posess, make sure to collapse your objects to a editable mesh. (Edit: only 3DMax 3, 4, and 5 is acceptable now, AFAIK)

A flyable plane model is divided into multiple sections/objects, namely:
* fuselage (always being the central chunk of a model)
* tail fuselage(s)
* vertical stabilizer(s)
* rudder(s)
* left & right horizontal stabilizers
* left & right elevators
* wing, divided into: left inner, right inner, left middle, right middle, left outer and, right outer parts
* ailerons
* flap(s) (if any)
* engine(s) (if any)
* propeller(s) (if any)
* movable gear parts, including wheels and bay doors (if any)
* bomb bay doors (if any)
* movable canopy part(s) (if any)
* nose part (for multi-engine bombers)
* leading-edge slots (if any)
* diving brakes (if any)



Orientation and Pivots
The orientation of the whole model is clearly shown above (flying towards -Y axis). The model should be built around its center of gravity.

Every movable part's pivot should be oriented so it normally rotates along its Y-axis (green).



The control surfaces' pivots should be oriented as shown below.



Plus, no left-handed pivots allowed. So you must "reset transform" manually for every object that was mirrored during the course of model creation.



Materials and Textures
A plane's "skin" is a 512x512x32bpp .TGA file and it should cover all the surface with the exception for u/c wheels, propeller, pilots and glass parts that use special texture reference within the game.



There are also 2 special textures to mark the a/c damage (in addition to possible geometry change) that contain 2 grades of combat damage (namely lower and higher).




We would appriciate these as separate layers in a supplied .PSD file. The resulting damage .TGA file should contain damage only, masked with alpha channel. This approach allows to paint damage texture over any skin the a/c is currently wearing, either supplied or user-made.




The texture utilizes the alpha channel. Thus, you can make a cockpit outline using either geometry or cutting out the holes in the alpha.





If you feel you need to show some holes (in the wings, e.t.c.) due to battle damage, tend to also make them alpha-cut to save triangle counts.


There should be two places on the texture that the propeller spinner refers to - one for the standing airscrew, and the other for the rotating one.




Geometry and Smoothing
The model's surfaces can be smoothed and faceted in the same way the 3D MAX does. The smoothing groups are imported into the game "as is" and the result is next to the same you can see in the MAX.

The scale can be any, but a 20/1 scale model will be appreciated most.

Produce every surface in a "fan" style. Like the MAX itself, the game engine smoothes the surface normals without their weight in mind; and if you want to make the surface smooth over an edge, the amount of faces connected to a vertex on that edge is to be the same to the "left" and to the "right" side of it.








LODs and Perspective Correction
The engine shows different models at different (freely definable) distances. It is recommended that you reduce the number of polygons by 2 times (appx.) each next LOD. Smaller parts should be omitted from a certain distance (i.e. melting the aeleron and the outer part of the wing into one simple prism since you can't tell one from another at 150 m away). (Distances are measured in a 640x480 window from a 70 degree FOV camera as the base, and then rescaled if you change the resolution and FOV).



Even a 3-part console collapses into one at a distance.



While it's necessary to omit small part from displaying at greater distances, the accelerators' tendency to collapse small triangles drives you to make small details larger than they are (i.e. a 7.62mm barrel is noticeable only if you make it a 30mm in diameter). Plus, you must make every detail thicker as the distance to it increases (Mind that the minimal visible element at a distance of 100 meters is as big as 0.2 m; and you are to make a plane's wings as thick as 2 meters to make them seen at 1000 m away).


-------------------


A compilation of IL-2 modelling info from Terrence "Majesty5" Daniels (maker of the AR-196 float plane) back in Nov 2003:



This page is the index page for the Il-2 Modeling Bible. It should be part of a large .zip format file containing a number of .html files, and the following folders:

1C-Maddox Models, with sub-folders Bf 109F-2 Cockpit and Il-2 External. These folders contain examples of meshes used in the Il-2 series.

Guides contains copies of the old 1C: Maddox modeling guide from late 2000 in .mht format, and a "Cockpit FAQ" compiled in late 2001. Edit: this is Luthier's guide (posted above.)

Plug-Ins, with sub-folders MAX 3, MAX 4, and MAX 5, contains plug-ins used by 3DS MAX (see Software page for descriptions).

This "Modeling Bible" is intended to collect all available information on 3rd-party 3D modeling for 1C:Maddox Game's Il-2 Sturmovik: Forgotten Battles flight simulator, published by Ubisoft Entertainment. It was compiled in late November 2003 in response to requests for more information from new modelers following the demise of Il-2 Center and the move to Netwings.

Although Forgotten Battles is nearing the end of its lifetime as a product, and it is anticipated that new models will not be accepted for integration into FB after mid-2004, this information will hopefully still serve as a useful basis for modeling add-ons to future 1C:Maddox Games projects.

The model system used in FB is anticipated to be carried over in some form or another into the new products such as Battle of Britian and its successors.

-----------------------

Externals

General Overview of the Il-2:FB Aircraft Model
This page gives a general overview of the Il-2: FB 3D model. For specific details about the model, such as parts division and naming, textures, and other structural details, please see the subject-specific pages linked from the main menu. The chart below shows the Il-2 modeling system; what is represented below is what makes up an entire "model" for submission. The end result of all the work below is contained in a single file which is submitted to 1C:Maddox Games for inclusion in the game. Each individual item on the graphic is explained in the accompanying text.



Compete D0_00 Model - This is what is generally referred to as the "final model", though in reality a competely finished model can be thought of as eighteen seperate models, plus associated small parts. This is the source from which all other models will spring. It is the full-poly-count, undamaged, close-range (0-25m) "perfect" model of an aircraft. Its completion is really only about 25% of the overall modeling process. The model is constructed in Metric measurement at 1/2 scale, 50% of the original aircraft's actual size. The most efficient means of construction is simply building the D0_00 model at 1:1 scale and then scaling it down 50% when construction is finished. Polygon limit is ~3000 for a single aircraft, and ~3500 for a twin-engined aircraft or bomber.

D0, D1, and D2 - These are Damage levels. D0 is "perfect", undamaged (simple enough). D1 is "light damage", theoretically a uniform smattering of bullet and shrapnel holes over the airframe. Ideally this is represented with a change in texture only, although modelers may build some minor damage in 3D if they wish. D2 is "heavy damage". Holes have been blown through the airframe, represented either by literally cutting 3D holes and leaving tattered 3D debris hanging in the virtual air, or by cutting alpha holes in the texture. Both methods can also be used at the same time.

Caps - These are related to damage level 2. "Caps" are exactly what they sound like: they cover the empty hole on a major part when its neighbor has been blown off. For example, if Wing_Mid is blown off, Wing_Inner will now be capped by Wing_Mid_Cap instead of Wing_Mid. Wing_Mid_Cap will be a torn-up stub of wing skin with bits of rib or spar hanging out. Caps are built for LOD 0-2.

Internals - a.k.a. "D4" parts. These are independent internal structural parts, related to each major part, which only appear in the D2 state. They are not strictly necessary but they add a lot to the model when D2-level damage has been inflicted. They carry the same name as their parent in the structure. For example, ribs and internal structure that would be visible inside of Wing_Mid_D2 would be called Wing_Mid_D4. Please note that the term "parent" here is only in a relational sense, and has nothing to do with the Linking function in MAX. Internals are constructed only for LOD 0.

Levels of Detail (LOD) 00-05 - As an Il-2 model recedes from the viewer in virtual space, its polycount is reduced step-fashion using the Levels of Detail system. At each step, the receding model is actually replaced by a new, lower-poly-count version. Each LOD is used across five different virtual ranges. Additionally, levels 3, 4, and 5 are actually increased in size over the closer models to maintain a visual presence at long range. LOD 3 is 110% the size of LOD 0; LOD 5 is 125% larger.

NULL Parts - These are ghost parts extending from each individual part to ensure good surface smoothing between parts. This concept has given modelers a lot of trouble, and so it is explained on its own separate page with screenshots and full description.


Suggested Building Order
Many of the sub-components of the model system have their own complexities and idiosyncracies. They are described on their own pages, linked from the main menu. Building a model for Il-2:FB is a somewhat complicated, multi-step process. The most efficient order of construction is outlined as follows:

1. D0_00 Model with texture.
2. D1_00 and D2_00.
3. D0_01 - D0_05.
4. D1_01 - D1_05.
5. D2_01 - D2_05.
6. Caps_00 - Caps_02.
7. Internals (D4 parts).
8. NULL parts added to all iterations of D0, D1, and D2, as well as caps.


Submission and Integration Process
Once the 3D model is complete, it is sent to Oleg & 1C:Maddox for integration into the game. This step is as simple as posting a .zip file which contains the .max file, textures, and any other necessary information in a modeler's personal webspace (or perhaps web space loaned from another person), and letting Oleg know via e-mail that the model is complete and ready for downloading. At this point, it is very, very important to know that building the 3D model is only a small part of getting a model into the game. The model must be imported into the engine, the actions of the various parts must be programmed, AI routines have to be written, and then the new model must be tested. This process can take anywhere from three to six weeks or more. All this work is done by 1C: Maddox staff in their spare time, in between actual paid work and product deadlines. It is unfair for anyone to demand updates on the progress of model integration, or take 1C: Maddox staff to task for percieved "slowness" in work to integrate a new model into the game. It is simply a matter of professional courtesy, and of "not looking a gift horse in the mouth."


When and how are new models released?
After successful integration and testing, the new aircraft have historically been released in batches as part of a patch or, in the case of Forgotten Battles, a new game entirely. Often the first view of a new model in-game is a screenshot in one of Oleg's regular Friday development updates.

-----------------------

Parts and Naming Conventions
This page describes the Parts and Naming Conventions of the basic Il-2 Aircraft 3D object(s).

Parts
An aircraft model in Il-2 is not a single 3D object, but rather a collection of parts flying in formation. The airframe is divided into these parts for purposes of locational damage modeling. Generally the fuselage and wings are each divided into thirds (fore-middle-aft fuselage, outer-middle-inner wings), with horizontal stabilizers, vertical stabilizer(s), and individual control surfaces as separate individual parts. Other individual parts include propellers, moving and/or ejectable canopy parts, turrets, antennas, or generally anything that needs to be separate for purposes of realistic movement. Divisions in the 3D airframe should ideally be made at structural break points of the real airframe. These should be decided on with the help of photos of real damaged aircraft. For example, the Ar 196's fueselage is divided at the front bulkhead where the engine mounts attach, based on two different photographs demonstrating that the real Ar 196's front end broke at this point in crashes. A similar rationale was used for the division in the wing of the J2M3 Raiden (currently under construction).

Naming Convention
Part names are in the format [Name]_[Damage]_[LOD]. Name describes and gives the location of the part; Damage is in the format Dx, where X is the level (0, 1, 2, 3, 4), and LOD is in two-digit format, 00 through 05. This convention describes the part completely and creates an alphabetical list when the parts list is viewed in MAX. No spaces are allowed in the name, and each individual word should be capitalized.

Below are diagrams of an example aircraft, the Arado Ar 196A-3, with individual parts labelled and color-coded.






Proprietary Names and Special Cases
What happens when an aircraft has a unique part (or parts)? How are they named? The best guideline in this case is to simply follow the naming convention as best as possible, and stay logical and consistent. Please notice that there are some proprietary names for parts in the Ar 196A-3 example, although in (some cases, the name was broken separately from the location indicator to ease readability of names during construction; the naming convention dictates that WireX_L_D0_00 should, perhaps, have been named WireXL_D0_00, for example):

* The sliding part of the canopy was called "Haube" because the modeler needed a word to differentiate it from the windscreen and aft enclosure before they were attached to the fueselage. It became Can_Haube_D0_00 (the other parts were Can_Wscreen and Can_Aft during building).

* Wires were labelled as such generically. WireAnt is the radio antenna; WireX_L is the x-shaped set of bracing wires on the Left side; there is a similar set, WireX_C (for Center) between the floats, not shown.

* Struts were described by form and location. Three types of struts exist, X-shaped support struts under the fueselage, V-shaped folding struts between float and wing, and main float struts. These became StrutX, StrutVee, and StrutMain. Strutmain's location modifiers were A(ft) and F(ore) as well as right and left.

* FloatSteer is the name for a small guiding piece between floats and fueselage. On the real aircraft the float rudders pivoted vertically on the guide pieces for extended/retracted, and the guide pieces rotated left and right to steer the rudders. In case 1C:Maddox wanted to program realistic motion for these parts, the modeler built them to function realistically with appropriate pivot points.

* Absent any direct guidance from 1C:Maddox or fellow modelers, the modeler simply named the MG 15 machine gun as MG15_D0_00

In the case of more complex aircraft, such as aircraft with multiple wings, tails, or engines, the parts should simply be named descriptively and consistently.

-------------------

Il-2 Aircraft Materials setup in 3D Studio MAX
This page describes the setup for "Materials" for each individual aircraft in 3D Studio MAX. It is based on information extrapolated from the original Il-2 mesh example .MAX file, circa late 2000. The author is not aware of any change in this system in Forgotten Battles.

Il-2 aircraft use the "Multi/Sub-Object" Material type in MAX. The number of sub-materials is set at 32. All parts of the model are assigned this Material, and their selection of the individual Sub-Materials is assigned by changing the material ID at the polygon level. Unfortunately the contents of all 32 Sub-Materials are not known; What is known about them is described below. The information is based on a viewing of the Il-2 example .max file in 3D Studio MAX.



Material Name is in the format [aircraftname]_o. For example, "ar196_o". Please note that this ends in the letter "o" and not a zero. Letters "p" and "q" are for LOD textures, installed by 1C:Maddox.
Material Type is set to "Multi/Sub-Object". Number of Sub-Materials is 32, set up and numbered as follows:

01. NULL
Diffuse 192 red (192/0/0)
Self-Illumination 100
This is the infamous NULL Material, mysterious but absolutely necessary for correct smoothing of surfaces when the aircraft is imported into the game. It must be named "NULL". Building NULL-Material surfaces adjacent to each individual part, for all damage levels, out to the first three LOD's (L's of D?) is the last thing you do before sending off the model to Oleg & Team. Basically for each individual part, you've got almost the whole rest of the airplane copied and attached with Material ID set to the Big Red 1. Please see NULL Material for a full explanation of the hows and whys of building NULL Material parts.

The following twelve Sub-Materials (#02-13) are the most important ones. They are the Materials you see as the surface of the airplane in MAX and in the game.

02. Gloss1_D0o
Specular Level: To taste, but should be set higher than Matte materials (08-13). Ar 196A-3 is set to 37.
(Map) Diffuse Color - Bitmap: (THIS IS YOUR SKIN) - "skin1o.tga", 1024x1024x24 bit Targa.
(Map) Opacity - Bitmap: (THIS IS THE SKIN'S ALPHA CHANNEL) A grayscale targa.
I build and use the opacity map as a separate file from the skin. This is not technically the correct way to build the alpha information, but it is actually more convenient during construction (see below).

03. Gloss1_D1o
Specular Level: Same as above.
(Map) Diffuse Color - Composite: Map 1 is skin1o.tga, Map 2 is damage1o.tga
The file damage1o.tga is intended to be a 32-bit targa with your damage effects like holes, burn marks, etc. (Damage level 1 in this case) as well as your alpha holes built in as a single image. I set up the slot for this file, but I leave "Map 2" unchecked and build without it, because for me it's easier to render & check transparencies during construction by using the skin's alpha as a separate file in the "Opacity" slot. The downside is that late in construction, MAX is looking for damage1o.tga and damage2o.tga, and it takes a bit of fiddling and some creative use of Photoshop Actions to build & check textures on the fly. However if you use the Composite method, the aircraft renders with a bright metallic-looking whitish surface, which is difficult to interpret when testing damage textures & effects. So I do the alpha separately.

How does this work out when you are done with the plane? Well, for final submission of the Ar 196A-3, I sent the following to 1C:Maddox:

* Ar196A3_HighRes.max (3D file)
* skin1o.tga (the main skin)
* A_skin1o.tga (the alpha channel for D0, needed for first loading of the file when they get it - remember, I have the Alpha map separate in the Opacity channel) alpha&damage.psd (a five-layer Photoshop file, containing black & white alphas for D0, D1, D2, plus two regular paint layers for the D1 and D2 effects).

...These files theoretically gaive them everything they needed to set it up for import, and also, I wasn't told that I did anything wrong. So, this method worked for me!

04. Gloss1_D2o
Same like above, but with Diffuse Color - Composite - Map 2 is damage2o.tga, which would contain second-level damage effects & holes in the alpha.

05. Gloss2_D0o
06. Gloss2_D1o
07. Gloss2_D2o
All as 02-04 above, but with 2-sided ON. This is the only difference.

The two-sided materials can be used for certain undamaged parts, but they are most useful for damage effects. Let's take an example, say, WingRMid_D2_00, which is the center section of the right wing in the D2 state. Imagine your airplane is attacked in the game, and it is struck in the wing by heavy cannon fire. The D0 piece is instantly replaced with the D2 piece, which has large holes blown through the wing via the alpha channel... But if you didn't build it with the two-sided material, then you will be able to see completely through the wing section, no matter what the angle. At a low enough angle you will see through things that just shouldn't be transparent... As if the whole underside of the wing didn't exist. This is obviously visually "wrong". To get around this you use the 2-Sided material for the D2 part. Now when you have a large alpha hole in the wing, you can see the opposing surface of the wing interior, basically just showing the texture on both sides of the surface.

There are two problems with this though. First, there is a performance hit for this (when rendering in 3DS MAX, and in the game as well), so use it only when appropriate. The other catch is, if you have a light undersurface color (as most camouflaged airplanes do), when the Material is two-sided, if a hole in the upper surface is larger than a corresponding adjacent hole in the lower surface, the lower surface may show bright camouflage color.

For example, let's say that you've modeled the effects of a cannon shell entering the wing from below, making a small hole, and then exploding against the inside of the upper wing surface, making a LARGE hole in the upper wing. Then there will be a light-colored undersurface, double-sided and plainly visible from above through the large hole. The I-16's wing, when heavily damaged, is a good in-game example of this phenomenon, which is especially apparent on VVS aircraft with their bright sky-blue bellies. Theoretically you could build an interior surface for the wing and map it separately to a dark color, but then modeling ragged-edged holes would require a large number of polygons. Using a material with the 2-sided option is the most effective compromise. A suggestion for getting around the undersurface color show-through is to try and model the holes in the upper surfaces to be smaller than the holes in the lower surfaces, or offset them somewhat (assume that a cannon shell has come in at a shallow angle).

08. Matte1_D0o
09. Matte1_D1o
10. Matte1_D2o
These are the same like 02-04 above, but with a lowered Specular Level. Set it "to taste", but lower than the Gloss materials. This is how the different surface finishes (metal & wood) are simulated on VVS aircraft, and it comes in handy for simulating metal & fabric on other planes. Adjust this as necessary. For example, with the Ar 196A-3, judging by various photographs in all sorts of lighting situations, the fabric surfaces were only slightly less shiny than the metal painted surfaces... So there isn't much of a difference in specular values between Gloss & Matte on this aircraft. But every aircraft is different... References, References, References!

11. Matte2_D0o
12. Matte2_D1o
13. Matte2_D2o
All as 08-10 above, but with 2-sided ON. Used exactly the same way as Gloss 2-sided materials. But if you are building damage effects for fabric-covered control surfaces, be especially careful of how the lower surfaces show through... Items like ailerons are much thinner than things like wings and fueselages, and therefore light-colored undersurface show-through will be much more apparent.

15. Propeller
Opacity 10.
Specular Level 75.
Glossiness 16.
Soften 0.1.
There is a Diffuse map for this Material available, installed by 1C:Maddox. When you've built your propeller, copy it, and replace each blade with a square that's as big across as the propeller arc. Rotate the square so that the chord of the square is the same as that of the propeller blade (it will look like a "/" from directly aside). Assign this material to the squares. Make sure you map the spinner to the second "rotating prop" on your texture. Name the whole thing "prop_rot" or something similar, to distinguish it from the non-spinning prop.

30. Glass2
2-Sided ON.
Ambient 26 (nearly black), Diffuse & Specular are both 128 gray.
Opacity 03.
This is the material for the canopy glass.

31. Shadow1
Everything is at the default settings, but named "Shadow1".

32. Shadow2
Same like above, but with 2-Sided ON.


----------------------------


Textures and Damage
This page describes and explains the texturing requirements and damage system.

Textures
The textures in Il-2 are linked to the damage system. Standard texture format is the targa (.tga) with alpha channel (32 bits). Textures for the aircraft should ideally be 1024x1024 in size, though 512x512 was the size for the original Il-2 game. Although 1024x1024 is not the official standard, since Forgotten Battles improved on the original game, it is generally accepted to build at the larger size for FB. If a smaller texture is required the 1024 can be scaled down; of course, the reverse is not true for 512.

The alpha channel of the targa holds anything on the airframe that requires transparency (such as canopy glass) and additionally, in levels of damage, holes in the airframe.

Damage
As explained in the general overview & parts & naming, an Il-2 aircraft is really a collection of parts flying in formation. There are three basic sets of parts, each corresponding to a "level" of damage. Levels are explained as follows:

D0 - No damage: This is the aircraft completely undamaged in normal running condition.
D1 - Light damage: Small holes from small or medium caliber machine guns and shrapnel.
D2 - Heavy damage is Light damage, plus large, threatening holes blown in the individual parts from cannon shells. Internal structure may well be visible, represented by D4 parts (see Caps & Internals page).

------ Image damage_d1.jpg missing -----



The three damage levels are represented by the three different sets of 3D parts, and the corresponding textures for those 3D parts. All textures for damage levels are built up from the pristine D0 texture in the same way that the 3D construction of damage levels are built up from the D0 3D parts. The following table outlines a method of working with the aircraft texture in a .PSD file to build the damage textures.



The final result of this will be three different combinations of undamaged texture, damage markings, and damage holes.

Radical changes to the airframe - what happens?
In the D2 state, it is likely that the internal structure of the aircraft - ribs, spars, etc. - would be visible through the large holes in the aircraft. This is simulated simply by building actual 3D internal parts that correspond to each D2 part. The design of these parts, of course, should be based on the internal structure of the actual aircraft being simulated. And when things get really bad for the virtual pilot, and pieces of his aircraft are simply blown clean off, then Caps come into play. Caps are exactly what they sound like; they cap the hole left by a missing part on its adjacent surface. Please see the Caps & Internals page for a full explanation of these additional parts and methods for their construction.

Special Cases

* The propeller (or propellers) does not have a D1 state; it is either in perfect condition (D0), destroyed (D2, with bent and/or broken blades), and "running".
* Flexible machine guns do not have damage models either. They exist only in the D0 state.
* Antenna wires (and other wires) also exist in the D0 state only. If they are damaged or broken, then they simply disappear.

--------------


Caps and Internals
This page describes and explains the 3D objects that show internal damage, and those that "cap" open surfaces when the adjacent surface is shot away in the game. As explained on the Damage & Textures page, these parts come into play when severe damage has been inflicted, and when the situation gets really bad and whole chunks are blown off of the virtual airframe.

Caps
Caps function exactly as their name implies; they cap the hole in the adjacent surface that remains when a part is blown away in the game. They take their part names not from the surface that they cap, but from the piece they replace; i.e. a cap that sits on the end of WingLIn_D2_0x is named WingLMid_Cap_0x.






Cap Construction
The quickest way to construct a cap is directly from the adjacent part of the surface you wish to build a cap for. In our example, after mapping and texturing is complete, to build a cap for WingLIn_Dx, make a clone of the complete WingLOut_D2_0x and use that as a basis. From here it is a matter of 3D surgery; use the mesh editing tools to cut away about 3/4ths of the structure. Imagine that you wield a virtual cutting torch, and think about what a large-sized explosion inside WingLOut would do. Use photographs of actual damaged aircraft for reference. The 3D skin of the aircraft would be peeled and torn in various ways; this can be built directly in 3D. Remember also that the cap's construction is based on the D2 level of damage, so use the pre-existing holes and D2 texture features to guide your construction of the cap. If you cut the 3D surface through the middle of a large alpha-channel hole in the D2 part, you will end up with a lovely and realistic ragged edge showing in-game. It also looks good to have internal parts of the aircraft's structure visible with the cap. These can be taken from the Internal parts, outlined below.

Internals
In the D2 level of damage, the large holes inflicted upon the airframe will make areas of the aicraft's internal structure visible. These are simulated with independent internal parts which are called into play when necessary by the game engine. They are given the designation "D4". This name is not official, really, but the damage parts in the Il-2 example .max file are labelled "D4" so that convention seems to be acceptable. It should be noted that it has not been made clear by official statement or example what exactly the "D3" parts may be. If a modeler wishes to name their internal parts "D3" it will probably not cause anyone any stress.





Internals Construction
The construction of Internal parts should be based on historical reality. It may not be entirely possible to obtain full internal schematics of the aircraft being modeled, but an effort should be made to locate references for what the aircraft's structure looked like. Most modeling books on a given aircraft type contain one or two photos of the aircraft on the construction line or in a damaged condition in the field; these can be used to guide D4 part construction efforts. For practical examples, spacing of ribs will usually be evident in panel line patterns on an aircraft's skin; often it is enough to build a set of planes perpendicular to the wingspan at the appropriate intervals and build a Boolean object. Subtracting a copy of the wing part and cleaning up the result will provide ribs. Spars can be built with simple boxes. These parts do not all have to be specially textured; mapping the spar's entire surface to a small square of an appropriate color on the texture, and mapping all the ribs to a generally correct image of one large rib (with mapping adjustments to avoid distortion and placement issues) is enough. If the modeler maps the ribs large enough, they may even be able to build alpha holes in the ribs. Remember that all mapping for damage parts starts from the D0 texture, so the texturing for the internal parts should be included somewhere on the standard texture. This can be seen in the skins for individual aircraft provided with the game.

On some aircraft which have fabric surfaces stretched over a metal frame, the metal frame can be represented directly with the D2 surface texture itself. Simply make a positive alpha of the structure on the for the part, and color the "exposed framework" differently from the rest of the skin on the D2 texture's paint layer.


-----------------------


LOD System
This page describes and explains the Levels Of Detail system used by the game to reduce poly count for distant objects. The "Level of Detail" system is a common technique in 3D games to reduce the polygon count of visible objects as they recede in the virtual distance during gameplay.

The LOD system for Il-2:FB uses a total of six different Levels of Detail. They are numbered 00-05, and the corresponding number for each LOD is included in the individual aircraft parts. See Parts & Naming for details. The general rule for each LOD is to reduce the polygon count from the previous LOD by half. If LOD 00 is the full-detail "perfect" model at 3000 polygons, then the modeler should aim to make the next LOD 1500 polygons. This is not an absolute limit, but a suggested guideline. The LOD's also trigger the switching of models at certain virtual distances within the game, and as the aircraft recedes in the distance, the far-distance LODs may actually be scaled up slightly to maintain a visual presence on the screen. There are also tricks and techniques for making good-looking, good-functioning LOD's in the game. All of this information is outlined in the following breakdown.



Building LODs is one place in Il-2 modeling where plug-ins can come in handy... Specifically, Polychop (see the Software page for further details). This plugin is extremely handy for helping with polygon reduction. It does not do all the work for the modeler, but it is a great time saver. The best method is probably to do a first pass with Polychop, adjusting settings until the result looks acceptable, and then manually editing the mesh further to achieve the desired results.


------------------


NULL Material
This page describes and explains the NULL Material. The quickest explanation of how the NULL Material works was posted by Fennec at Netwings. What follows is his description:




Applied to the Il-2 aircraft, this concept results in in each part having "ghost" versions of adjacent parts in the NULL Material. This is easier to show with screenshots than to describe verbally. Below is a gallery of screenshots of the Ar 196A-3 that illustrate how null parts are extended from each individual main part of the airframe. All NULL parts are constructed from "perfect" (Part_D0_0x) damage state. In damage levels D0 and D1, the fueselage sections, wing sections, and tail surfaces all have NULL parts. In the heavy damage level (D2), only the fueselage, vertical tail(s), and horizontal stabilizers need to have NULL parts.





NULL Part Construction
NULL parts should be the last step in the models' construction. They are surprisingly time-consuming to build, they turn the aircraft bright red in the MAX viewport during construction, and they increase file size & file load times in MAX. However, NULL parts actually do-not increase polygon count in game, because they are used only during the model import process. They are not in the game engine during play. The best construction method is to simply clone the whole D0 model, change all Material ID's to 1, and then clone, edit, and attach the new NULL parts to each individual part of the D0, D1, and D2 models as required. Repeat this procedure for LOD 00-02.

The exact Material parameters for NULL parts can be found on the Materials setup page. They can also be viewed directly in the Il-2 .max file from 1C:Maddox Games.


---------------------


3D Construction Techniques, or "How do I...?"
This page describes various techniques, tricks, and caveats in building a model for Il-2 that have not necessarily been officially documented. They are "modeler's lore", information most often given as the response to a question posted on the modeler's forums. Some of them are generally known 3D techniques, others are more specific to Il-2 modeling.

Smoothing Groups
Smoothing groups tell the MAX rendering engine that the shading of all surfaces in the same group should be averaged together. What this means in non-technical language is that generally, all surfaces which face the same direction should be in the same smoothing group. The numbering of groups is totally up to the individual user, provided that the numbering system is consistently used throughout the model. Here is the system used on the Ar 196:



9-14 alternates for above, used to differentiate matte material from shiny material. 25-30, same as 1-6, but for clear parts. All other numbers can be used as necessary to differentiate individual features on the airframe.

Double-Sided Surfaces
One of the handiest construction techniques in Il-2 modeling is the double-sided or double-faced surface. This is constructed simply by building a surface, cloning it, flipping the clone's normals, and welding the clone to the original surface in place. This technique can be used to make many thin surfaces like control sticks, seat buckets, canopy glass, antennas, internal ribs, doors, trim tabs, speed brakes, and even some major control surfaces.

Canopy Frames and Glass
Canopy surfaces can be constructed in two different ways: As alpha holes with one of the double-sided Material variants applied to the appropriate geometry, or as distinct 3D panels and frames with double-sided geometric surfaces.

Intersection vs. Integration of Parts
On small surface items such as gun fairings, bulges, air intakes, antennae, etc. often it is better to built the item as a separate surface and simply attach it in place to its parent object than to carefully integrate it vertice-by-vertice into the surface it is attached to. This means that the surface of the object intersects its parent surface, rather than integrates with it. This is acceptable to the game engine, and it helps lower the polygon count. It also means that there will be a visible difference in surfacing between the intersecting objects, even when they have the same smoothing group. Usually, though, this is not much of a problem, and in some cases it can actually be desirable. The modeler may even wish to go so far as deliberately assigning the object to a different smoothing group to make sure it stands out from the parent surface.

Antennas and Other Small Parts
Antennas, wires, etc

Cowlings and Engine Faces
Engines have been a sticky spot for several modelers. When constructing cowlings and engine faces, the general technique is to build everything as one single object. It is also important to remember to assign different smoothing groups to the engine face, the inside of the cowling that surrounds the engine, and the outside of the cowling. If these surfaces are not given different smoothing groups, the result will be dark artifacts in the corners where the angle of the surface changes. And although the basic face of the engine should be built to connect with the cowling interior, parts that protrude from the engine face (such as crankshaft casings) can intersect with the surface of the engine face, rather than be directly geometrically integrated with it, vertice-by-vertice (see section above on Intersection vs. Integration of parts).

"Standard" Parts, such as flexible machine guns
For parts such as the flexible MG 15, there are standard meshes and textures used by 1C:Maddox modelers that have been distributed, but not as a single official package. The exchange of such parts has been on an informal, as-required basis, which effectively means that there are individual components floating around "out there somewhere". It is best in these cases to inquire on the modeler's boards, or failing that, ask the staff at 1C:Maddox directly if they have a required part (such as an MG 15 or flexible ShKAS) handy.


----------------


----------- Cockpits page empty -------------


----------------


Ordnance
Ordnance... Hi: The new LW bombs are 256x256...and that is for the big ones. the Sc250 and smnaller, together with cannons, the Ju87G pods and some others come from a "SINGLE for all" 256x256 map. LOD0 for those bombs are 200 more or less...but could go to 300 if it is complex in it make (like the rockets for assisted take-off) I am also starting to do some PC1400 and SC1600m and maybe some others. I specially want to get the equipment containers for fdropping supplies...(Stalingrad?) Anyone has some PICS!!!???


-----------------


Ground Objects
Ground Objects... Few observations based on what Oleg had sent me: (please note that most of the tips posted above also apply (like file format, naming convention etc.) and here are the major differences) - There are 3 LODs only, usually with shadow and DM. - Textures are 256x256 for LOD0, 128x128 for LOD1 and 64x64 or 16x16 for LOD2, 24bit TGA files. 32bit can be used for making windows via alpha cut (so do not model the frame in the mesh). - Max polys for LOD0 (only) seems to be no more than 500, usually much less, but all depends on the objects modeled... - If there's open driver(gunner) position (like in Jeep) there can be a simple figurine. - LOD2 and it's shadow are basically a (squashed) box. LOD0 shadow is often the same mesh from LOD1 external.

Here are the sample models I had a chance to examine:
------------------- Valentine MkII tank ------------------- LOD0 - 392 LOD1 - 254 LOD2 - 32 Shadow LOD0 - 265 Shadow LOD1 - 145 Shadow LOD2 - 20 textures: 256x256, 24bit TGA 128x128, 24bit TGA 16x16, 24bit TGA ------------------

M4A2 Sherman tank ------------------ LOD0 - 480 LOD1 - 170 LOD2 - 30 Shadow LOD0 - 170 Shadow LOD1 - 30 Shadow LOD2 - 20 textures: 256x256, 24bit TGA 128x128, 24bit TGA 16x16, 24bit TGA ------------------

Willys Jeep ------------------ LOD0 - 502 LOD1 - 174 LOD2 - 64 Shadow LOD? - 88 DM LOD0 - 302 DM LOD1 - 150 DM LOD1 - 64 DM Shadow LOD? - 56 textures: 256x256, 32bit TGA

[Vpmedia]

Thanks a lot, this is very usefull!

[RAF_Magpie]

hehehe I'm just waiting for him to say:

And here's one we prepared earlier

Wow! This will be very handy

[Guest]

Thankyou so much for this - i have passed this thread over to our guy who is working on the 3d converter for us

[FltLt HardBall]

No worries, guys - just my little bit to help out. I was hoping all this stuff that I squirreled away over the years would come in handy one day!

By the way, I just discovered that also have a max file of a BF-109 cockpit with all the original textures. If anyone wants it please PM me.

[Nightshifter]

By the way, I just discovered that also have a max file of a BF-109 cockpit with all the original textures. If anyone wants it please PM me.

Got a feeling that some folks here will be biting your hand off to get a hold of that

Thanks for the time you put into all this.

[Trooper117]

Outstanding... a good comprehensive look on a pretty complex business..

[Bee]

Really impressive. It would take ages to work all that stuff out by trial and error. Hopefully an insider mapmaker will come forward with similar treasure too.

S!

[Hellzone]

Bravo FltLt HardBall

Great Find. :)

[asheshouse]

Wow!

I knew the aircraft models were complex but I didn't realise how complex.
--- and the BoB aircraft will be on a new level!!

[FltLt HardBall]

Majesty5's tutorial continued. (The first post won't accept any more content!)


--------------------


Reference Issues and Techniques

Here are some randomly collected tips and techniques that might help a modeler with references.

The greatest weapon against reference confusion is comparing data. This is time-consuming, and it can be exasperating, but when, for example, three drawings all show the same part with a different shape, the only recourse is careful study of the drawings against period photographs. When in doubt, go back to the real thing. Photos, drawings, and good historical data are the trinity of references needed to build a good model.

A common problem with collections of basic aircraft measurements is rounding of the true numbers. It can be as small as rounding 9,696.5mm to 9.7m, or as large as 12.44m to 12.45, but when a modeler is trying to find out the exact dimensions of something, sloppy data can throw off many other related measurments.

A related problem is matching drawings correctly to measurements. Sometimes two different technical artists will draw the same object out to the same measurement, but in different ways. For example, is a given wingspan measurement taken from the centerline of the aircraft, or from the root of the wing at the fueselage? Or maybe even from the structural root of the wing inside the belly? This sounds like nitpicking, but it is an actual example encountered in the building of an Il-2 model. Now, if two different artists draw out to the same wingspan measurement while starting at different points, the entire shape of the wing may be rendered indeterminate.

A knowledge of basic aerodynamics is a good thing to have. This can be taken further; the more a modeler knows about airplanes, the better the final product. For example, a single-engine propeller aircraft's vertical tail is built slightly turned to one side in the vertical axis, usually about one or two degrees, to counter-act engine torque. Another example is wing shape. Most aircraft wings have two different airfoil shapes, one at the tip and one at the root. Furthermore, the wings often have what is called a "washout"; the tip airfoil will be canted downwards with respect to the root airfoil, to make sure that the wingtips don't stall before the roots and potentially cause a spin. This is a small detail on the model that is often overlooked by new modelers, resulting in a clunky, board-like wing shape that is inaccurate, and worse yet, looks wrong. If a modeler does not know what an aircraft should look like, their model will not look like an aircraft.

A quick formula for determining the angles of aircraft parts using drawings is Rise over Run, ASIN. This works best for digital copies of drawings, using the Marquee tool and the Information palette in Photoshop. If a modeler wishes to know the angle of a wing, for example, run the Marquee from the wing root and place its opposite corner at the end of the straight part of the leading edge. Divide the box's height by the length and hit the ASIN function in a calculator. This will give you an angle. Take multiple measurements if necessary.

Running a photo through Photoshop and playing with the levels may bring out key details that would not be visible otherwise.

When modeling or looking at references, it helps to keep a notebook handy to write down details about an aircraft that may be forgotten later. The difference is in the details, and there are lots of details on even the simplest aircraft that can be overlooked. On the structure, remember to look for pitot tubes, static ports, antenna fairings, wires, steps for the crew, wingtip and fueselage lights for navigation and landing, gear up/down indicator pins on the upper surface of the wing near the cockpit, tie-down or other attach points or pins, and flush mountings for cables or equipment. Texture features to look for include gas filling ports, signs for rescue equipment, gasoline octane, lift points, and warnings around areas such as propellers and control surfaces.

There is no such thing as a perfect reference book. Even the best of them have a mistake somewhere. This is another reason to always, always compare references.


----------------


An Informal History of the Third-Party Modeling Efforts

2000

2000.02.xx While a growing fan base gawks at the incredible development screenshots showing up in the Friday Updates at the Blue Byte forums, potential 3rd party modelers begin to discuss the feasability of a 3rd-party expansion effort. After some discussion outlining the policy of "developer quality control", 1C:Maddox releases the first modeling guide.

2000.08.xx Evidence of a modeling effort begins to surface in the form of model update posts at Il-2 forums and even an occasional screenshot. The beginnings of a community starts to evolve.

2000.12.09 An early version of the Majesty5's Arado Ar 196A-3 becomes the first 3rd-party model tested in the Il-2 engine.


2001

2001.05.09 Il-2 Center, paid for, built, and run by Luthier, is founded. 3rd party modelers begin to gather and coordinate modeling efforts.

2001.06.20 Il-2 Sturmovik beta testing begins with a pool of 100 testers chosen in secret. Beta CD's are delivered by special international service in a small package. For several days until everyone gets their CD, doorbells cause excitement and anxiety in 100 people across the world. Receipt of the Beta CD prompts excited posts on the secure Beta forums.

2002.09.11 Nihilistic loser assholes hijack planes and crash them into buildings in New York and Washington. Having witnessed the attacks personally from the front of the Staten Island ferry, Luthier takes leave from Il-2 Center for a few weeks. General sim-flying activity declines as people feel the need to refrain from simulating using aircraft to kill other people. (someone) takes up temporary administration of Il-2 Center in Luthier's absence.

2001.10.30 Il-2 Sturmovik beta testing is complete after four months of testing and three package updates to the core beta, which fix various bugs ranging from flight model issues to purple clouds, while adding new features such as trees.

2001.11.20 Il-2 Sturmovik is released. Flight simmers dive in and expend precious hours of their lives flying strange new aircraft of the East against well-known aircraft of the West. The absence of the Bf 110 is noted. The differences between Il-2's simulation of aircraft vs. that of other games begins a renaissance in simmer's conceptions of warbird performance. Old myths crumble, while new questions are raised. Proving that the Cold War isn't over, xenophobia, ignorance, and even genuine curiosity lead some to question Oleg's methods and partiality. "Is Oleg a secret Communist? Everybody knows that any flight data from Stalin's time would be biased!" Despite the fact that lying to Stalin was a wee bit on the risky side, and Oleg's claims to an aeronautical engineering degree and his own avowed preference for the Bf 109, the phenomenon of Luftwhining is born. Legitimate complaints and questions about German aircraft modeling are lost in a sea of irritating troll posts. In the meantime, most people have their cake and eat it, too.


2002

2002.03.xx Addressing some of Il-2's weaknesses, Vadim "Starshoy" Kolosov releases his Dynamic Campaign Generator for Il-2, providing players with a great new way to expand gameplay. Many people begin putting together historical missions and campaigns for the game.

2002.06.14 The 1.1a patch for Il-2 includes Luthier's BI-1 and Zurawski's PZL P.11c, both entirely new flyables. Lending credence evidence to the theory that "Those who can't, complain", a small minority of people wonder just what the hell a Soviet rocket plane is good for, saying that a 3rd party modeler's small amount of unpaid free time should be spent on a different sort of completely free bonus that would bring their idea of "more" to the game. Some modelers, in response to this new trend of having their intelligence questioned, begin asking the community what sorts of aircraft would be most appreciated. Some members of the community respond with spirited, positive movements lobbying for the inclusion of smaller air forces such as the Italians, Romanians, Hungarians, and Finns.

2002.07.xx As discussions of what should (and should not!) be added to Il-2 continue, some fans come up with the idea of a World War 1 package or add-on. Oleg reveals that the current Il-2 engine could concievably handle everthing up to 1950's jets, with the exception of helicopters. He also agrees that if he recieves a package of flawless 3D models that don't need any corrective work from the 1C:Maddox team, he will try and take the time to put together a WW1 patch. Over the space of a few weeks, a World War 1 project is born at Il-2 Center, and soon gets its own forums. Although the level of activity is low, a couple of the more prominent modelers begin building WW1 aircraft behind the scenes. These eventually result in a handful of really wonderful screenshots, but as time progresses and Il-2 Center activity fades, the project seems to lose momentum and disappear. Its status is currently unknown.

2002.08.xx Il-2: Forgotten Battles is announced. Originally concieved as a large add-on, over time FB morphs into a whole new sim building on the original Il-2 core, adding many game engine enhancments, new theaters, new campaigns, and dozens of new aircraft. Answering the hopes of those who wish to see more "minor" nations represented, it is revealed that the new campaigns will be set in Finland and Hungary.

2002.09.01 The highly anticipated 1.2 patch for Il-2 makes the Ju 87B-2 flyable, courtesy of Jippo's cockpit. Legions of Stukaholics rejoice, and while people enjoy the experience of piloting the infamous dive bomber in the best WW2 flight sim out there, they look to the future and begin to dream of piloting the Ju 87D in a tankbuster role.

2002.10.xx The Killjoy Saga unfolds across the Il-2 modeling community. Someone calling themselves "Killjoy", later determined to be (at best guess) a 15-year-old from somewhere in the Low Countries, descends upon the Il-2 Center boards. It stars with a harmless question about using Milkshape 3D for Il-2 modeling. "No" is not a good enough answer, evidently, and soon the question is repeated, again and again. Frustrated at making no progress in changing things by making five-word posts every ten minutes, Killjoy requests that he be allowed to model the Mosquito. Luthier, who by now is not around much due to his job, gives the newbie the benefit of the doubt and approves. Three weeks of poorly spelled spam posts follow, interspersed with screenshots of a poor-quality model thrown together in Milkshape which utterly disappoint real fans in the community who were overjoyed at the prospect of a Mosquito in Il-2. Gibbage, no master speller himself, gets into arguments with Killjoy in a series of epic battles that are unreadable for most English-speaking people. Killjoy then schizophrenically abandons the Mosquito and moves over to the Gloster Meteor, which soon recieves the same treatment as the Mosquito. He then posts screenshots of other people's professional 3D work as his own in Oleg's Ready Room, and admits that the shots are not his work after thorough questioning by skeptical posters. Suspecting that Killjoy is behind recent e-mail threats and attempts to hack the site, Luthier bans Killjoy and his recently discovered aliases from Il-2 Center. The death throes of spam posts last another week or two before finally dying away. The whole affair leaves a trail of astonished people in its wake. The real victim, however, is Gibbage, who is left with permanent psychological scars and a temper that automatically explodes on the mention of the word "Killjoy".

2002.10.xx In a surprising twist, Luthier reveals on the Modelers-Only Il-2 Center forum that Forgotten Battles will have the capability to show individual faces on the virtual pilot's head. This means that 1C:Maddox needs photographs of individuals for pilot's faces. It eventually leads to modelers posting their pictures and a revealing their real names to each other. Now the modelers know each other by name, and correspondence becomes more personal and more frequent. An observer might even say that an interesting, and somewhat overdue, bonding has taken place.

2002.11.xx A similar phenomenon to the "WW1 Project" now develops around the subject of late- and even post-WW2 and experimental aircraft. Advocates of post-WW2 aircraft debate with those who want to see wild experimental prototypes in the game. Historical purists have their say as well, with one person even suggesting that aircraft with production runs of less than 100 examples shouldn't be included in the game at all. Others point out that obscure, low-production Russian types (the La-5FN) are included, so why not aircraft like the Ta 152 and Do 335? The argument now suddenly breaks down into a Luftwhiner vs. Evil Communist Biased Game Developer fight. Gibbage, ever the diplomat, deliberately throws gasoline onto the fire by announcing that he is developing the Lockheed P-80 Shooting Star (a borderline late-war aircraft) in a brutal shot at the Luftwaffe fans, who are already upset that the Me 262 will not be available until Forgotten Battles adds multi-engine management. This ignites three-week-long flame threads in the Ready Room. In the meantime, less inflammatory discussion goes on in the background at Il-2 Center, with most parties deciding that more aircraft of any sort can only be healthy for gameplay. Eventually plans for a "1946" package evolve, which is intended to contain at least the Go 229, Do 335, and P-80. Majesty5 even formulates secret plans to build the F7F-3 Tigercat for this package, but is unfortunately thwarted by lack of references. Xanty's Do 335 also fizzles out due to a diabolically timed massive hard drive crash. Gibbage succesfully completes the P-80 and Go 229, however, both of which are expected as flyable add-ons sometime in the future, but the larger "1946 Project" will disappear as a collective effort when Il-2 Center eventually goes down.


2003

2003.01.xx As Il-2's lifetime wears on and Forgotten Battles takes shape, the steady, systematic decline of Oleg's Ready Room, a place where Oleg can theoretically interface with the community, begins. Some very real flaws and drawbacks to the game attract a fair share of criticism, which unfortunately begins to snowball. Perspective is lost as people realize they can take personal shots at the developer without fear of retaliation from lax moderators. Smelling blood, professional whiners, loonies, agitators, and kiddie trolls (later referred to by Oleg as "some not nice people") bombard the forums with noise, and Oleg's online presence becomes rare. Never a place for the sane, soon the board is almost fully populated by people with psychological dysfnuctions that are common to the Internet, such as trolling, martyrdom, and persecution complexes. Twenty-page argument posts between socially stunted, overtly nationalistic airplane nerds make it difficult to tell what is real and what is relevant. People who actually know what they're talking about begin to correspond in private rather than public. Many other regulars of Il-2 forums who actually have lives and functioning brains decide that it's time to unplug and get on with real life for a while. Some of them quit the Internet entirely. The 3rd party modelers begin to disappear one by one, into various places like college, the military, or other personal work. In the meantime, the Ready Room becomes one of the worst places on the Internet, trailing only behind fetish-picture-trading IRC channels, usenet conspiracy groups, and boards featured on the "Weekend Web" at Something Awful.
2003.03.04 Il-2 Sturmovik: Forgotten Battles is released. It includes the new flyable TB-3 and I-16 cockpit by Luthier, the Me 262 cockpit by Vibert & Derrien, Gibbage's PBY-6, Walter Pedrao's P-47, Xanty's He 111 external re-work and cockpits, MW's B-239, Radek's P-40, SB's Su-2, and Majesty5's Arado Ar 196A-3.

2003.07.xx Il-2 Center's front page is defaced by a collection of recessive genes masquerading as an illiterate Portuguese-speaking teenager. As the weeks wear on without the obvious problem being fixed, the small number of modelers remaining begin to wonder where Luthier is. Discouraged by the decline of the community and the drying up of the modeling effort, and busy with other things, activity has declined to the point of an occasional "check-in" on the forums.

2003.08.xx Il-2 Center's remaining online content (really only the front page and the forums) is suddenly, mysteriously deleted. Bewildered modelers figure that if that doesn't bring Luthier back, nothing will.

2003.08.xx After months of no activity, a small group of modelers decide to move to a new "home" online. Under the coordination of Gibbage and with the help of Niel Washbrook, a move to Netwings takes place.

2003.09.xx Oleg announces his future plans for Forgotten Battles. He reveals that there will be one more patch, followed by a 1C:Maddox-developed paid add-on package. Additionally, he reveals that a European campaign with a new Normandy map has been in the works by an independent developer. As a warm-up for this European add-on, the patch will include the P-51D (flyable) and B-17 (AI). Community reaction alternates between praising Oleg as a God and decrying the imminent Communist desecration of the sacred Mustang.

2003.10.06 Piling on the good news, Ubisoft announces that the next 1C:Maddox sim will be a simulation of the Battle of Britian, expected in late 2004. The dreams of those who hoped for the Mediterranean theater are crushed, until Oleg floats the possibility that the Mediterranean theater will be tackled after BoB, using it's engine as a basis for new sims. On the whole, reaction is positive to the new announcement. Oleg also points out that the new game will very specifically model Italian participation in the Battle of Britian.

2003.11.13 Majesty5, tired of having to scroll down past half a dozen sticky posts that re-hash old information at Netwings, decides that it's time to put all known modeling information into a single, easily-distributed package. Having already compiled a Cockpit FAQ, he figures (wrongly) that it wouldn't be too much work to collect everything in one place and build an HTML index for it. And so, this Modeling Bible was born.

2003.11.19 Capping a shocking two weeks that has already seen a lunar eclipse and the death of Dr. Romano on "ER", Luthier drops a bombshell by simultaneously returning and announcing a new Il-2 add-on product under his aegis in a single post at Netwings.


-------------------

Disclaimer
1. The authors and contributors to this document have no business connection to 1C: Maddox Games, Ubisoft entertainment, or any other companies, firms, multinational megacorporations, intelligence agencies, shadow governments, cults, organized religions, alien civilizations, political action groups, fetish subcultures, terrorist organizations, or crime syndicates, that have been named, shown, or indirectly represented in this document. All information contained herein is third-party material or has been released for free for the benefit of the modeling community.

2. 1C: Maddox Games, Ubisoft, and all other individuals, companies, firms or organizations mentioned herein are in no way legally liable for anything that might happen to your body, your property, your mental health, your sex life, or that of your dependents, pets, friends, enemies, or acquaintances, through the use of the information in this document.

3. Names of various software packages are trademarks of their respective developers and/or publishers, and are used here for information purposes only, in a non-promotional sense.

4. Please note that ALL third-party work for the Il-2 series of games is only fully accepted for integration into the game after the author of the model has signed and submitted an internationally binding legal agreement with 1C: Maddox Games and Ubisoft establishing clear rules of compensation and liability. All modelers who submit their work must go through this procedure.

5. The modeling community does not endorse the copying, trading, or alteration of unlicensed software of any sort, and any requests for such within the community will be frowned upon and/or deleted. If you're looking for "warez" don't bother asking us. We are all students and working people who are not interested in spending time in Federal prison for stupid reasons.

6. Finally, all tongue-in-cheek humor found in this document is purely accidental. No animals were seriously harmed in the making of this document, but several bottles of cheap liquor suffered horrible deaths. If this document was anything other than perfect, any (purely hypothetical) mistakes in it would be the fault of the author, who can be reached here. Edit: e-mail address removed.

[FltLt HardBall]

Gibbage's 3D Max Tutorials

Edit: I have taken the liberty of correcting the spelling :wink:

Setting up your Workspace

The key to a good model is good references and an efficient work space. This is not a tutorial on how to organize your desk space, or even your desktop. That mess is up to you, and I am the last person you want advice on about THAT. This is about setting up 3 view drawings into 3D Studio's Max to help you re-create models from drawings. Here I will demonstrate how to do this with the P-63 Kingcobra WWII aircraft.

First think you will need is a good set of high res 3 view drawings. The more detail, the better. There are a few great internet sites were you can download these including a few Russian sites, but the best place is from a book for many reasons. #1 is you control the quality. Below is what a typical 3 view drawing looks like.



Note a few things with this image. First, its MUCH too low a resolution to use as a good 3 view. Get your own. Two, it uses what I call "half" views. The top and bottom views are combined into 1. This is done if the aircraft is an exact mirror on both sides, and to save page space. Also, the front view its cropped. No need to draw both wings if they are the same. Also note the sectionals. THESE ARE CRITICAL! The sectionals are the line drawings marked with numbers or letters. In this case, they are marked with A, B, C, D and so on. The sectionals shows you what the aircraft section would look like. You can see at the top, the side view is divided into section's A-K. This corresponds with the sectionals A-K. The wing is also divided to show you a cross section of the wing and its aerofoil. These sectionals help out a LOT in getting the shape of the body perfect.

When I set my work space up, I divide the 3 view into 5 sections. Top, Bottom, Side, Front, and Sectionals. I save each section alone in its own file's. The top file has only the top drawing in it and so on. You could have just 1 file like above, but I find it less complicated to do 5 files in the long run. I do this with Photoshop and cut/paste and erase the stuff I don't want. When in Photoshop, do what you can to clean things up.

Now that your 3 view is chopped up, lets work in getting it into Max. Most people set each file into the Viewport Background. I have found MANY problems with this. Scale for one is difficult to keep, and detail is another problem. In order to get it high res, you really should have them on separate planes. For this, you will need a good 3D video card. Lets start out by making sure Max will display the full res file.



Its as easy as 1, 2, 3. Click Customize, Preferences, Viewports, Configure Driver, and then put a check on "Match Bitmap as close as Possible. This will make is so textures are applied in full res in your viewport. This is critical.

Next, lets set up our materials.



Set the Self-Illumination to 100, and click the box next to the Diffuse channel. This will bring up the Material/Map Browser. Double click on "Bitmap" on the top right corner. This will bring up an explorer window. find your TOP file and click open.



First, click on "Go to Parent" button to return to the material property. Now rename your new material TOP. Now, make sure you click the blue and white checker box. This is VERY important. Without that, you WONT see the texture till you render it. That will be useless for us.

Now repeat this for the Bottom, Side, Front, and Sectional files. You should now have 5 separate materials with the 5 views. Each named by its own name. Not lets pace the materials.



Click on Plane and drag a plane out in the TOP view. This will be the object we put the TOP texture on it. Now that you have the plane set, look at the graphic file for the TOP texture. My top texture is 1274 pixels high, by 1470 wide. You can see that I made the Length and Width of the plane to fit 1274 by 1470. Click in the Length and Width box and type it in manually. Now, make sure that Generate Mapping Coords is CHECKED. Hit the M key to bring up the Material Browser and now apply it to the plane



Here is the end result on the top. You can see the drawing on the plane in the viewport. Also, its very high res. Also note that I had to reverse the Length and Width around since it was stretched. If yours looks squashed, do the same. Now do the same for the other 4 materials.



This is what it should look like once you're done. I rotated them into their proper positions. Now its time to size them up and make sure they scale well. Using the size of the drawing is not going to work. The side view may be longer then the top, and the front higher then the side. Make sure all the drawings line up wing tip to tip, and nose to tail.



Here you can see I made sure everything lined up well. Wing tips and nose tail. You can also see that I need to lower the front view plane to match the canopy and tail. Once I make sure everything lines up, I move them back. I also lowered the opacity of the side and front view to help me line things up, but once I am done I return it back to 100% opacity. Now, its time to get to work and show you how this works!!!!



You can see here how I use it. I made a cylinder with 16 sides with as many segments as sectionals. I made a material with a 50% opacity so I can see through it to the drawing behind it. That's the key. I converted the cylinder to an Editable Mesh and then aligned the segments with the sectional markers on the side view. Time to make it fit in both the side view and top view.



You can see here how I use the non-uniform scale to shrink and grow the cylinder to meet with the front and side views. For the nose, use the uniform scale to keep the perfectly round shape. We will use this for the spinner later. The body is coming together quickly. Note that in the front view the body does not fit 100%. That's were the sectionals come in.



Here you see how I am using the sectionals to align every vertex with the corresponding section. I move the sectional plane forward and align the middle and top/bottom that should be accurate. Do this for every section, but only do this to HALF of the body. That will save you a LOT of time, and you just mirror it later for a perfect body. For the the wings, use my Wing tutorial.



Here is the final result after aligning all the sections, and mirroring the body. It don't look like much yet, but remember that this is just a start.


-----------------


Low Poly Wings

The wing of an aircraft can be very complicated to model. This method is the best I have found so far to build a quick and highly detailed aircraft wing in less than 5 minutes.



First start off by creating a line shape that fits the airfoil of the aircraft. Most 3-view drawings have a sectional of the airfoil. Use that as a template. I draw the shape using the Line tool. I moved the shape away from the drawing so you could see the result better. Remember not to make it too complicated for low polygon modeling. 6-8 steps are typical.



Next, use the Extrude tool. Extrude the shape you just made to fit within the width of the drawing. Add only the segments that you need. Add 1 segment per side of the flap, 1 per side of the aileron, and two for rounding the tip. I also added 2 for the engine boom. These segments will later on help you cut the control surfaces out of the wing.



Now change the extruded line to an Edit Mesh so we can make the wing fit better into the drawing. Select the vertex edit mode, and select 1 line of vertices. These lines of vertices I call the spars because they remind me of the spars wings are made of. But a real wing is made of MANY MANY spars. Align these vertex spars with the critical points on the drawing like the flaps, ailerons, and wing tips. Again, aligning these spars will help you cut out the flaps and ailerons.



Here, select Uniform Scale, and scale along the Y Axis to fit into the drawing's constraints. You may also need to move the spars on the Y axis.



Do the same in the forward view. The wing tip can get complicated. Its best to sometimes delete the polygons on the end, and collapse the vertices at the tip.



Here is the complete wing. Next, read the Smoothing Groups tutorial to complete the wing and make it look smooth.


-------------------


Smoothing Groups

To me, smoothing groups are at the top of the most important skills to have in a professional manner. Most beginners can get away with using the "SMOOTH" tool, but the pros manipulate the groups by hand. I have not used Smooth in about 1 1/2 years. Doing the smoothing groups by hand takes a bit more time, but the results are well worth it. Here I will demonstrate smoothing groups on a simple shear. The Smoothing Group section in max may look complicated, but once you understand the basic principals about the groups, its extremely easy.


Here you see the default sphere. It automatically sets all polygons to smoothing group1.


Here is what happens when you remove all smoothing groups by unselecting the highlighted 1 in the smoothing groups section. You can now see each polygon the sphere is made of.


Here the top half polygons are set to smoothing group 1, and the bottom half to group 2. This creates a smoothing "break" between 1 and 2. Note that I have not changed the model at all. This helps reproduce the effect of a plastic ball that was made from two pieces, like a Pokeymon Ball, or a plastic Easter egg.


Here, the two halves again are in separate smoothing groups. But this time the center polygons have BOTH group 1 and 2 selected. I color coded the the different sections to help visualise the smoothing groups. Remember. Yellow and Blue make Green. The green polygons with both smoothing groups selected blends the two together.


Here you see the use of smoothing groups on the P-38 model. The fillet uses a separate smoothing group to give it a detached look. The wings have group 1 on top, group 2 on bottom, and group 1 and 2 on the leading edges. This keeps from getting the typical dark edges on the trailing edges of the wing you see a lot when using bad smoothing groups. The boom also uses smoothing groups to help define the depression for the GE Turbo Supercharger. The Smooth command WOULD NOT be able to duplicate this effect. Also, the turbo itself uses a bunch of smoothing groups. The tires also use two smoothing groups. 1 for the rubber, and 2 for the hub.


Here are the results from BAD smoothing. At first, it does not look that bad. But look at the details. A lot of them are missing. Also the radiator intakes got REALLY messed up.

[Guest]

I have many doubts about the 3d modeling.
Already tried in various ways understand and design but I am Brazilian and I have difficulties in English language
Someone could help me in the modeling of this plane?




The links below are all topics that posted asking someone to make the plane.
And nobody helped me, I am trying to do it myself
http://www.simviation.com/cgi-bin/yabb2/...1204136699
http://www.sim-outhouse.com/sohforums/sh...hp?t=21406
http://www.sim-outhouse.com/sohforums/sh...hp?t=60407
http://www.simnetwork.com/SIMNETWORK/for...ead_id=536
http://www.alphasim3.com/forums/index.php?topic=7624.0
http://forums.surclaro.com/ftopic14729.html
http://www.thewarbrothers.com.br/forum/v....php?t=776
http://allaircraftsimulations.com/forum/...php?t=2936

As you can see I am not lack strength of will to build the plane!
But lack knowledge in other forums was ignored, but I hope not to be ignored by you!

[asheshouse]

PabloSniper
This is not really the right place to try and learn 3D modelling.
I have tinkered with aircraft modelling but am not an expert.
The most useful help I got was from the Military Meshes site
http://www.military-meshes.com./
They have links to good tutorials on the site
One of the best I found was for a Fiat500 car. The techniques are similar to do an aeroplane.
You can get it here
http://www.3dtotal.com/team/Tutorials/f ... iat500.asp
Best not to just sign up on Military Meshes and ask for help.
Get a tutorial. Have a go and post some images of what you've done on MM.
You are certain to get someone taking interest.
Its what the site is all about.

[Guest]

Hardball,

Many thanks for your postings, some of which have disappeared for some reason? Gibbage's tutorial is great on exteriors but there's nothing about cockpit interiors.

I am thinking of having a go at modelling a plane from scratch, but have yet to come across any info regarding cockpits. Do you have anthing you could post here? Ideal would be full blown tutorial but failing that perhaps a wireframe mesh pic, another of it smoothed and then another of it textured ? At least then I could get some idea of what's invovled.

I have been trying to collect info about the modelling techniques involved and slightly off topic here, came across an article on line with an industry insider who suggested that with today's computers low poly counts were a misnomer and no longer the holy grail they once were, and that for new game designs the fps bottleneck is now texture rendering. This stuck in my mind and I wondered what you thought about the possibility of trying to up the poly count to get rid of the nut like wheels and other slightly cartoony looking bits. Say up to 5000 poly for a single engined fighter exterior? The other lods could slighty bigger than they are currently? Perhaps Hellzone or Jolly would have some idea about whether the game engine could cope. Mapmakers and the cockpit texture artists and skinners have all increased file sizes without seemingly any undue effects.

If you have any more info regarding modelling for the game please post it, I know there has been little response to what you have posted but it's a steep learning curve and it looks like a lot of work involved and most of that research! Seems impossible to find any really good free blueprints to start something worthwhile.

Anyway thanks once more for taking the time to share your information, I'm certain lots of people have an eye on this thread and keenly await further developments. I'm hoping Fatduck and Hellzone and Mr J say definitively that putting scratch models into the game can now be done, then we can start work.

Jameson