PDA

View Full Version : Want to make cokpit for FB to have your fav plane flyable? Here's what it takes:



Jippo01
10-03-2004, 11:52 AM
Ok. Since there has been a lot of discussions about all sorts of things about cockpits of the planes in FB/PF, I will explain to you all what is a cockpit in this game and how you can make one if you wish. This way everyone will get a better idea of the game mechanics, techniques used and the work involved in the process. This way we will get maybe bit more reasonable discussions about the matter of cockpits, and maybe even find some ways to improve the game together.

THIS IS NOT THE PLACE TO SAY THAT YOUR FAVOURITE AIRCRAFT HAS BEEN MODELLED INCORRECTLY OR ANY OTHER **** LIKE THAT. This is the place to understand and talk about modelling in general level, not judge in bad or good. So please respect that.

Let's start with the beginning. What do I need to have/be able to do to succesfully finish a cockpit for FB?

1. You need to have a 3d modelling program that can export files without screwing them up to 3d Studio MAX format. Of course this favours the 3d studio product line itself, but life is never fair, is it?

2. For a cockpit you need a plane to stuck it in. Either you will start building a pit for an existing plane(good), or you will make also the external model(twice the work).

3. You will need time. Lot's of it! If you are an experienced modeller and do a simple pit, you may pull it through in couple of hundred hours inc. research. If you are in the beginning of your modelling hobby or a career, looking for building a bomber - be prepared to spend up to 1000-2000 hours for the job. The maths: 1000hrs, 2hrs a day, 5 days a week = 100 weeks = 2 years of your life. Be motivated, network and form teams.

4. You also need references. Depending on your task again, if it is simple AC less will do, but more complex it gets the more you will need. The printed references I used in making of the Ju 88 cockpits weigh 13,5kg (25-30 pounds?) and I have more than 200Mb of other documents and photos on my harddrive. There are problems in finding all this material, but since you are making a lot of people happy with your work they will do their best to help you in your research. But be prepared to spend in books and trips to museums too. I visited 4 seperate European museums for several occasions while doing my research. Stuff you may need include books and old documents that show everything in the cockpit in drawings and photographs, and also describe the function of all the parts you will model:
http://www.kolumbus.fi/jan.niukkanen/cockpits/158-5866_img.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/158-5867_img.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/158-5873_img.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/158-5874_img.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/158-5877_img.jpg

I have a lot of stuff about 88, but then again I spent couple of years collecting it. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif Also you can start with less references, but make sure you know that you get more as time goes by. It isn't going to be fun to find out after a year of work you can't finish the job!


I will continue writing this and adding new messages to this thread as my time permits. Part two will come later today, and other parts giving deeper information during the following week.

Keep it clean and serious guys, trying to be helpful here! http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

CowboyTodd41
10-03-2004, 12:43 PM
All I can really say is, thanks for the Ju-88 cockpit, and hot ****! That's alot of research material!

Jippo01
10-03-2004, 12:45 PM
How do I model a cockpit according the references I have?

As I go through the process, I will take my examples out of the Ju 88 A-4 Early model made for FB. To view the finished model visit the Ju 88 fansite(link in the signature) and download the panorama player and files of the cockpit.

First think of what you will have to make. You will need to produce a 3D model where all the parts visible will be modelled. All the parts must a texture which is either handpainted, baked, or a photomanipulation. You will also need to make a damage model with varying degrees of damage. You will need to follow certain conventions to produce this model first of all everything is made of triangles (polygons), and all of these triangles will have to have their face normal facing the camera. Textures you use will have to be square sized and no larger than 256x256 pixels, this means you will have several small texture files instead of a single "skin" like in the external model. You will be limited to about 3000-6000 polygons depending the complexity of the pit (this means most objects will be very, very blocky) and 5-10Mb of textures. Tools you will be using are probably very simple modelling tools like box, cylinder, loft, and so forth. You will be also using a lot of Mesh edit and UWV Wrap moving your precious polygons around and texturing them.

When we start modelling, we will need at least a 3d-wireframe of the plane cockpit area to get the rough shape and scale right. If you have made the external model, you obviously will start with that. In other cases either contact Maddox Games or the third party who made the external model.

As the case most likely is, you will not have dimensioned factory construction drawings (drawing set can constitute from hundreds of thousands documents, about 200 000 drawings in case of Ju 88) to base your work on. Most likely you have obtained some blueprints from books or the internet, you probably also have some perspective drawings and photographs of panels. Lets go through the good and bad sides of all of these different types of references:

Blueprints pros:
- 2D and easy to read
- show also parts that are not often photographed or cannot be even seen on plane (sections)

Blueprints cons:
- they cannot be fully trusted (I have no two blueprints of 88 that fully agree with each other on details) - drawings are inherently inaccurate
- do not contain detailed information, can be very schematic
- do not contain colour information.

Perspective Drawings Pros:
- Often give the best overview of the pit
- Can have detail
- Can show important things highlighted

Perspective Drawings Cons:
- they are inaccurate and untrustworthy to degree
- Can give false information
- Colors are not to be trusted

Photos pros:
- accurate
- can contain very detailed information
- contain reliable color information (even B/W)

Photos cons:
- can be difficult to fully understand, room for judgement error
- can misguide colorwise (lighting, aging)
- are often from non-standard models and prototypes as they appear in publications much more than your bread and butter model


So where are we then? We have loads of reference but we can't really trust any of it? Huh?!

Yes, that is right. Only thing you can more or less trust is your photos, just remember how difficult they are to read and also they can misguide you.

What I found the best route is to use the blueprints along with the external wireframe to construct a rough cockpit model which contains seat, instrument panel, walls and the floor of the pit. Combine the blueprints you have and see where they disagree with each other - you can most likely verify the correct lines by using your other references about that area of the pit. But don't spend too much time, you will change everything many times in the future as your understanding of the model and the aircraft increases. But you got to start somewhere. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

http://www.kolumbus.fi/jan.niukkanen/cockpits/cockpitrough.jpg

When you get the basic shape and size together, start fitting boxes of the bigger pieces of equipment in and seeing can you fit their volumes in your framework. This is where the tweaking starts, more about that in the next part. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

heywooood
10-03-2004, 12:51 PM
Thanks Jippo - This is exactly why I visit this board anymore. Your work on the Ju88 looks outstanding and I can't wait to fly it.

I am not surprised at the volumn of reference material or the amount of time involved as to attempt a perfect render requires patience, references, and TIME. (not to mention Skill and Talent) you seem to have accumulated plenty of all that.

Thanks for putting this up.

Ruy Horta
10-03-2004, 12:52 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Jippo01:
Ok. Since there has been a lot of discussions about all sorts of things about cockpits of the planes in FB/PF, I will explain to you all what is a cockpit in this game and how you can make one if you wish. <HR></BLOCKQUOTE>

Jippo,

First you are doing a great job with the Ju 88, although we haven't seen any of this work in game yet the screen shots are stunningly detailed.

BUT, and there is always a but.

You are modeling a multi-seat bomber here and giving it as your main example. Second, you are also including the sub-variants into your work AND the extreme detail which YOU are willing to model.

Now not all a/c require this amount of work, some because they are single seaters, and others because they are just much more simple in terms of instrumentation etc.

A Ki-27 will never be the same as a B-29.

Since I am quite fanatical when it comes to this subject, I have a fair idea of what is required.

This is one extreme, there is a middle ground.

For many a/c, you'd suffice with one Aero Detail type of book and a number of general reference works.

That you went the extra mile with the Ju 88s is great and we must be thankful, if we ever get to play with her in this series that is.

xanty
10-03-2004, 12:55 PM
Well written Jippo!

I can only second you, and marbel at your tenacity and the quality of your team's work. S!

Jippo01
10-03-2004, 01:17 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Ruy Horta:

Jippo,

BUT, and there is always a but. <HR></BLOCKQUOTE>


Yeah, you're of course right. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

But I can only describe my process in my model. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif But that is not the point, I just try to get the process across. Number of hours spent will vary, many are much faster than I, people may have different attitude...

As I said from couple of hundred hours to thousand or more, that is big scale already. I don't want to freak out any wannabe 3rd party, but I also don't want to give unrealistic too optimistic view of the effort. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

But in any case, numbers are not important, I just try to describe what sort of phases there are in modelling a pit.

starfighter1
10-03-2004, 01:50 PM
hallo,
nice to see this Jippo01... http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

hope to show first a new wireframe Modelling of FW-190,..
I coordinate and work only on my free time so it takes a time.. but anyway I cooperate with a fine designer in this ..

by the way re to point 1): any chance to use 'LightWave3d' for export formats too ?

File Compatabilities: by

* LightWave 3Dâ® includes a library of license-free objects, images, textures and scenes.
* LightWave 3D File Import Formats:
o dxf
o obj
o 3ds
o lwo
o ai
o eps
* LightWave 3D File Export Formats:
o obj
o 3ds
o lwo
o dxf
o w3d
o Any other files may be imported or exported by using a 3rd party translator
* LightWave 3D Render File Formats:
o Still Image:
+ Alias (.als)
+ BMP 24 & 32-bit (.bmp)
+ Cineon FP (.cin)
+ JPEG (.jpg)
+ PCX (.pcx)
+ PICT 24 & 32-bit (.pct)
+ PNG 24 & 32-bit (.png)
+ Photoshop 24 & 32-bit (.psd)
+ RLA 24 & 32-bit, Deluxe RLA (.rla)
+ Radiance (.hdr)
+ SGI 24, 32, 48, & 64-bit (.rgb)
+ Sun 24 & 32-bit (.ras)
+ TGA 24 & 32-bit (.tga)
+ TIFF 24 & 32-bit, LogLuv (.tif)
+ VPB(.ypb)
+ YUV (.yuv)
+ Flexible Format (.flx)
+ IFF ILBM & IFF ILBM 32-bit (.iff)
o Animation:
+ 4X Storyboard
+ AVI & DirectShow (.avi)
+ Film Expand
+ QuickTime, QuickTime VR, & QuickTime Stereo (.mov)
+ Storyboard
+ Flexible Float & Flexible Integer (.flx)

Specifications subject to change without notice.

jimmie_T
10-03-2004, 11:01 PM
Please continue this series. Thank you so much and respect your work. Really hope it be included in the product!

Jippo01
10-04-2004, 06:26 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by jimmie_T:
Please continue this series. Thank you so much and respect your work. Really hope it be included in the product! <HR></BLOCKQUOTE>

It will don't worry. Oleg said it is a top priority after PF is sorted. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

I will continue the series and go through all the different phases of modelling, but now I must run!


-jippo

csThor
10-04-2004, 06:52 AM
... but now I must run ...

Why? Your wife swinging the rolling pin? http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

Jippo01
10-04-2004, 11:36 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by csThor:
_ ... but now I must run ... _

Why? Your wife swinging the rolling pin? http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif <HR></BLOCKQUOTE>

Nah. I just had to shoot something. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

DuxCorvan
10-05-2004, 02:28 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Jippo01:
Also you can start with less references, but make sure you know that you get more as time goes by. It isn't going to be fun to find out after a year of work you can't finish the job! <HR></BLOCKQUOTE>

Oh, it's so lucky they found all those references for Bf 109Z, Go 229 and BI-1 after 2 years of hard work! Now they're nailed to the real ones!

Because Oleg would have never allowed a reasonable guess to pass his incredibly high standars on accuracy level, wouldn't he? http://forums.ubi.com/images/smilies/16x16_smiley-tongue.gif

He he, I'm evil. http://forums.ubi.com/images/smilies/59.gif

PS: You've made an impressive work with that Ju 88s! Now that we know it's been so easy, we want more! http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

WereSnowleopard
10-05-2004, 03:58 PM
Hi Jippo01

I wonder if you had been contact by real Ju-88 pilots/crews to tell how much they proud of you work hard doing modeling on Ju-88 in high detail to be flyable?

Cheers
Snowleopard

Extreme_One
10-06-2004, 05:25 AM
Fascinating post.
Thanks for that Jippo and for the Ju-88 cockpit.

I don't fly bombers much but I have spent time in every cockpit in the game so far and I look forward to seeing the fruits of your labour too.

Jippo01
10-06-2004, 08:53 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by DuxCorvan:
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Jippo01:
Also you can start with less references, but make sure you know that you get more as time goes by. It isn't going to be fun to find out after a year of work you can't finish the job! <HR></BLOCKQUOTE>

Oh, it's so lucky they found all those references for Bf 109Z, Go 229 and BI-1 after 2 years of hard work! Now they're nailed to the real ones!

Because Oleg would have never allowed a reasonable guess to pass his incredibly high standars on accuracy level, wouldn't he? http://forums.ubi.com/images/smilies/16x16_smiley-tongue.gif

He he, I'm evil. http://forums.ubi.com/images/smilies/59.gif

PS: You've made an impressive work with that Ju 88s! Now that we know it's been so easy, we want more! http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif <HR></BLOCKQUOTE>


I'm sorry, but this is exactly what I asked NOT to do in the first post. And yes, also 109Z, 229, and Bi need to be made with same attention to detail. Have you spotted some errors in the cockpit? (rhetorical question, please don't answer, I want to talk about making of cockpits - not argue about different planes)

Let's just stick to the subject matter of creating cockpits, ok? If you want to ask something about that, I will try to answer.

I will write at least one more part today. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

jeroen_R90S
10-06-2004, 11:33 AM
Reading your posts, this is exactly why I respect you (and of course all other modellers/skinners etc) so much!

Looking very much forward to Ju 88, especially C-6! And the Pe-2, of course http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Jeroen

Jippo01
10-06-2004, 01:51 PM
So now that we have a framework put up, we need to think about putting the actual cockpit furniture in it. What are FB cockpit objects made of?

Well, they are usual, although not always very simple objects like faces (2 polys each), boxes ( 6 faces, 12 polys), cylinders (>8 polys), spheres (totally depends) etc. Very simple parts indeed. What we want to do for a computer game is something low-poly modelling, and this is a technique that insists on shaving off unneeded polygons EVERYWHERE. This means that all objects done with as few triangle shaped flat faces as possible. With angular objects this isn't a problem as a rectangular face only consists of two triangles and is so inherently low polygon object. Where this obsession of getting rid of polygons becomes somewhat problematic is the round objects like spheres and cylinders - if we spend only a couple of triangles trying to represent a sphere the end result isn't going to be too flattering. Well you can image. It is a constant balancing act between appearance and very tight polygon budget. And this is a compromise that cannot always succeed, something in the pit looks like ****, but what can you do if you are over your budget by 200 polys? You certainly cannot go adding more detail even if you have it in your book and it would certainly look all good and realistic in the model too. No, you want go adding this magnificent piece of detail, instead, you will go and make another round around your precisious little pit and scavenge more polys out of it loosing 200 polys worth of already existing detail.

Normally when a rendering is done for a presentation, movie, or something like that where it can be first calculated and then displayed on a separate occasion modellers do high polygon modelling. High polygon modelling is to me the normal way of modelling. It means that if we want to model a curving surface, we will make the polygons so small that it is impossible to notice the individual small angles between the triangles, if we need something we will shape it out of tiny triangles. When we do architectural modeling in our office, we will spend minute or few of rendering time for every intermediate test image we make. After we have created the model, we will calculate lighting for it with a program that gives realistic light reflections and shadows for every part. For the final renders it may take half an hour to an hour to render just the one frame needed.

Obviously we cannot use a technique that needs an hour for every frame in a computer game where people will want 50 frames per second, so we will need to compensate the lack of detail in the actual 3d-shape by using texture. Texture is painted so that it appears as a three dimensional object already in the drawing program. Texture artist will add some soft shadows, light wear and tear to protruding parts and so on. When such a texture is put on a low poly 3d-model it will give an appearance of a far more complex entity than the actual model. Here is an example of extremely simple piece of model that appears extremely complex in the game, normal air speed indicator texture:

http://www.kolumbus.fi/jan.niukkanen/cockpits/asi1.jpg

And here is what it looks in 3d-model. Just needle added on the left. Both are jus as simple as possible - two triangles forming up both rectangles. 4 polygons all together. Not much to see, huh?
http://www.kolumbus.fi/jan.niukkanen/cockpits/asi2.jpg

But when these two are combined, and alpha map (opacity layer - tells program which parts are see through) added the appearance in the cockpit is actually already quite pleasant.

http://www.kolumbus.fi/jan.niukkanen/cockpits/asi3.jpg

There are also some other objects that are almost completely made by painting a good texture and assigning them to a very simple piece of geometry. Pedals two polys each, small windows on the sides 10 polys each:

http://www.kolumbus.fi/jan.niukkanen/cockpits/pedal.jpg

Some more complex objects are needed of course, but a lot can be made by just using simple faces with good textures, or alternatively by using boxes or cylinders. One thing to remember is that all the polygons that are not visible must be deleted to save computing time and memory. This means that a box by a side wall that can be seen from only one direction may only have 3 faces, as the three on the opposite corner can not be normally seen have been deleted. This prevents the use of fully movable camera within the cockpit: although the pit is fully 3d, it will appear horrendous from any angle where objects can be seen from "behind". Small camera movement is usually ok, but say half a meter movement in the pit might already show that the pit is like a set for an old western movie. Only front sides are real, and behind them is nothing (or the desert in the movies). http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Here are couple of Ju 88 objects first seen from the side they are usually seen, and then a peek behind the set to see how they are constructed. I'm sure that if you try you will be able to trace the original geometric shapes that are behind the different parts.

Radio:

http://www.kolumbus.fi/jan.niukkanen/cockpits/radiof.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/radiob.jpg

MG-81j machine gun:

http://www.kolumbus.fi/jan.niukkanen/cockpits/gunf.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/gunb.jpg

I think next part will be about placing all these funny little objects around the pit and the problematics behind that. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

Jippo01
10-09-2004, 10:25 AM
Sorry for no update today. I have been busy with our flat which is supposed to be finished soon so we can move in.

I will try write another part tomorrow. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

Dr_shang
10-09-2004, 12:22 PM
Congratulations, Jippo, for extraordinary job and extreme dedication of Ju-88 cockpit...

This plane must be flyable in IL2-FB at the begining

IV_JG51_Razor
10-09-2004, 01:51 PM
I'm sorry, I know this is definitely NOT what you want to see in this thread, but the devil made me do it!

Jippo, your work is absolutely beautiful, as well as all the research that went into it. Nobody could possibly question that. However, I think that there are a lot of people out here who might question Oleg's total fanaticism over historical accuracy, to the extent of excluding some aircraft on the basis of lack of data to produce an accurate cockpit.

I bring this up because, as a pilot for Erickson Air Crane co, who operates a fleet of 19 S-64 helicopters, and has delivered two others to foregin goverments, for a total of 21 Cranes around the world. I can tell you for absolute certain, that if you did as much research on this aircraft, as you did on your JU-88, you'd fail miserably at coming up with a completely accurate cockpit. Out of 21 airframes out there that we operate, or have sold, there are not two in the whole group that are identical!

As an ex-Army helicopter pilot, I can assure you that, if you walked down the line of 12 Hueys in our company, you would not have found any two that were perfect matches. Some instruments might have been moved, or relocated, different model radios instaled, different color schemes (both inside and out), etc, etc, etc. For sure, you can make a pretty good example of what your research shows to be what any particular aircraft was at the time, but it will never be exact.

If you are not able to determine what the EXACT color of certain components were for a certain model plane, wouldn't it be reasonable to assume that the color of the same component of a different plane (of the same country) would do? If the dimemsions of a cockpit can't be exactly determined for a particular plane, don't you think that they could be approximated well enough for a 2D presentation in a flight sim? Most countries, or branches of their armed forces used very similar types of gages for airspeed, altitude, vertical speed, attitude, etc, etc, etc. If a modeler can't find, in the course of his research, the proper gage for airspeed, for example, wouldn't it be acceptable to use one from another plane of the same country/branch of service?

I'm only asking these questions because there is a lot of frustration out here due to the lack of several important planes. I also realize that, as third party modelers, you guys make whatever you feel like - as long as Oleg accepts them. My frustration lies in the refusal, on Oleg's part, to accept a particular model due to lack of adequate research material. If, indeed, no such material exists, then who can possibly say that it didn't look like this?!! I will say that I appreciate Oleg's high standards, but I feel that in some cases they have hindered the development of this wonderful game of his.

Y'all may now return to your regularly scheduled program. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Jippo01
10-10-2004, 11:43 AM
Razor, you bring up many good points in your posts.

Let me go through some of them in detail.

1. Variation between cockpits. Yes, most cockpits vary from each other in details. This is actually not a problem, more a relief for the modeller as this means that the cockpits don't need to be absolutely factory specs to be realistic. Some artistic freedom and personalised cockpits are allowed. In fact I have taken neat non-standard features from several planes and included them in the Ju 88 model simply because the make the pit look better. Point is to be realistic, but not trying to make a perfect copy.

2. Instrument types are in general the same between German planes. They vary by the manufacturer slightly, but instruments of any of the manufacturers could be included in any plane of any factory. Also we exchange instrument between planes. Same instrument textures are used in many planes, as are guns and other objects that are standard. For example He-111 top gunner uses gun from Ju 87 model.

3. Modelling for FB is not exact science, and there cannot be fully accurate cockpit - actually they are quite the opposite. Compromises have to be made all the time, shortcuts have to be taken, accuracy of modelling has to be sacrificed for game speed etc. The game is fully 3D and that is the biggest difficulty. Pictures of old planes have been taken from the most interesting parts like panels and guns, but nobody was interested behind gunners back. But in this game we can turn our heads and see what is happening behind us in the game - there needs to be something and that something has to be based on reality on a level (if not 100%). Often such points are the reason for "lack of references".


I hope that I could clear some points.


-jippo

Jippo01
10-10-2004, 02:38 PM
Now that we know what kind of parts the cockpit is made of, we shall take a look on how to build up the cockpits using such parts.

When we have made a rough framework of the cockpit area with the help of the external model we can start finding out how we can fit in the parts we need to. Again drawings are great if we happen to have them, but they are not to be trusted. Drawings are inaccurate. But if we have a nice drawing that shows for example the overall layout like this one for 88:

http://www.jagdgeschwader53.flugzeugwerk.net/ju88/Ju88A-4/Cockpit/Ju88A-4heck.jpg


We can use it for finding out what sort of parts we are about to stick in the pit. It shows very well the layout and how many objects are needed. This time we are concentrating on the 6 radios and the two machinegun ammo boxes on the left. We just want to get these parts in the pit and have their overall relations matching each other.

This makes modelling with less references difficult. If we do not have enough pictures from different angles we may not get the parts looking ok at all, but the real reason for the need of good references is that they save hours and hours of your time because when you are able to do things near right in the first time you don't need to come back to adjust for ten times. Remember that since we don't have any factory drawings with dimensions at hand the only things we can use to dimension the parts are the parts themselves. We can read the drawing and note that radios in the top row are rectangles that are approximately 3/4 tall compared to their width, and the radios are about 1/4 tall compared the the height fo the cockpit rear armor. This allows us to dimension parts relative to each other and allow us to place them realistically in the cockpit. Problem with this technique is that it is fairly inaccurate, and modeller will at points notice that he has made some parts wrong size and has to go back and correct everything according his new, better knowledge. So it isn't simply about building the pit once, at least I have had to do it several times.

It is better to have good pictures of the parts instead, if it is at all possible it would be great to see and maybe even touch some of the parts before starting to model them. This way you will get the sizes and shapes correct already on the first try. Often this is of course impossible, but if it is possible to do so, then do. It'll save a lot of sweat later on.

http://www.kolumbus.fi/jan.niukkanen/cockpits/su1.jpg

Here is an example photo of the same drawn radios in a museum. I guess there isn't any question which one reference (the drawing or the photo) gives us more information.


Now let's have a look at the environment modeller works in:

http://www.kolumbus.fi/jan.niukkanen/cockpits/su2.jpg

Usually there are 4 views of the model open: side, top, front and perspective. This allows modeller to judge the scale and shape of the modelled parts in three dimensions, and gives him possibility to move the camera to the same position as in e.g. reference photo to check if the appearance is as it should be. Checking the correctness of things in game is almost impossible as the player is tied to one perspective only, that being the one of the pilot, and in the modelling program it is very easy because all these views can be viewed simultaneously. Also blueprints can be used as a refernce when modelling in the non-perspective view ports as can photos from different angles used in perspective viewport. This doesn't mean that the modeller does always perfect job and there are no mistakes in FB/PF cockpits, far from it. It just means that when people find a "mistake" in the model, it is almost impossible to prove the error without having access to the model in the modelling program.

For example I used the sideview superimposed with a side view of the plane when I did the cockpit bars. But even though I had very good side view bleprints of the plane I had problems. The reason being that 3 side views I had, all of very good quality, didn't agree with each other. I still had to resort in using photographs from different angles to get tight spots appear as they should. This is awfully difficult and time consuming task with three-dimensionally curving spatial structures like canopy framing. It is doable by using combination of all possible viewing angles and combinations of side, top and frontal blueprints and the verifying the result with photographs. But it is so **** difficult thing to do, that when people ask to model refraction too I cannot but laugh - it was hard enough to get it close to right in correct 3d. I wouldn't dream to start doing all sorts of light reflection angles calculations, and then trying to change the model which matches a bunch of blueprints and photos to something that wouldn't anymore match anything else than my fancy calculations. Thanks, but no thanks!

But back to our business of making radios when we have all the details right, we will have end result which will stand the test of double checking with the references. The drawing gave us general layout, photos gave us the details we need and based on these two we are able to construct a model which is more or less accurate in dimension and appearance even if we wouldn't have real dimensions to start with:

http://www.kolumbus.fi/jan.niukkanen/cockpits/su3.jpg


-jippo

heywooood
10-10-2004, 02:52 PM
WOW Jippo...what a huge series of tasks.

Jungmann
10-10-2004, 09:08 PM
Thanks for letting us inside the shop, Jippo, and letting us see what you do with your talent.

S!

Jippo01
10-13-2004, 01:17 PM
Next part is: how do the textures work in FB/PF cockpits?

Many of us have skinned aircraft in FB, and maybe most know what it means? Aircraft external model textures are a single graphic file where all the parts of the model have an area that can be painted to color, and the aircraft in the game will display that color in that specific part.

Cockpit texturing and textures work in similar way - there are texture files that can be altered and that change will appear in the cockpit. But there are still several differences between external 'skins' and internal 'skins'. While the external texture is a single 512x512 or 1024x1024 pixel image, the internal model textures consist of several (on complex pit dozens) 16x16, 32x32, 64x64, 128x128, 256x256 files that can have textures for multiple parts in the cockpit. Alternatively a single part of the texture can be mapped (this is what placing of the textures is called, they are mapped in the right place) on several different parts in the cockpit.

Sounds complicated, doesn't it? And it just doesn't sound like complicated, and it is very easy to loose track of different textures in a complex pit. I sure know I lost track several times. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Here is what a texture may look like:

http://www.kolumbus.fi/jan.niukkanen/cockpits/tex1.jpg

The example file is quite a simple one - it is the textures needed for the MG-81 machineguns in the Ju 88 cockpit. But also it is a good example how the textures work. Maybe you can spot the different parts of the machinegun from the texture: there are the hand protection cover of the (possibly very hot) barrel, the barrel itself, the top and the sides of the box, the wooden pistol grip and the box for safety etc...
This same texture is used for all MG-81's in the Ju 88, and this is why you will see the same serial number in every MG. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif We could have just as easily made everyone with a different number, but then the need for graphics memory would have quadrupled when talking about machineguns in the pit. If you want to visit the actual gun modelled, you can see it in the machinegun vitrine in RAF museum Hendon. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

http://www.kolumbus.fi/jan.niukkanen/cockpits/tex1b.jpg


Here is another example of simple texture, but this time one texture is used with many different pieces of model, namely different sign plates around the pit:

http://www.kolumbus.fi/jan.niukkanen/cockpits/tex2.jpg

Here are different textures that were used in the pit:

http://www.kolumbus.fi/jan.niukkanen/cockpits/tex3.jpg

Now you should be able to get the picture of the textures. Let's briefly look how they are applied to different parts of the model. Applying a texture to part of the model isn't enough, we also must tell the 3d program (and the FB/PF engine too) which part of the texture is shown on the different polygons. The tool used in that is called "Unwrap UVW". With this tool it is possible to align the polygons of the model with the parts of the texture you want apply on the polygon in question.

http://www.kolumbus.fi/jan.niukkanen/cockpits/tex4.jpg

In this picture you can see the heat shield of the machine gun. The white lines on top of the texture show the edges of the polygons used in the three dimensional object which represents the barrel. The cornerpoints (highlighted with red) of the polygons can be freely moved in three dimensions in Unwrap UVW tool: object remains physically the same, but the texture can be aligned on the object no matter what shape it has.

I personally disliked this job as one needs to be very accurate, and especially mapping complex curving objects can be 'intellectually challenging', at least on my level of understanding of the program. But without the texture and mapping the model is just a collection of boxes, looking stunningly dull. It is not the modeller that makes one model stand out from the crowd, but the texture artist. Here I must mention that the texture artist for the Ju 88 is a very talented German chap called Dietger Pohl, and the textures you have been enjoying in the screenies (and will be enjoying in the game later) are made by him.

Next part of this thread will be about the viewpoints and connection/interaction between the external and internal models. This will also be the last part of the series, so if you have some questions raised by my writing, please don't hesitate to ask. Now is the time to sharpen your teeth and rip me into pieces, ask difficult questions. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif I will be happy to answer if I know the answer. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

Bearcat99
10-13-2004, 01:33 PM
Thats why I dont complain too much. I couldnt even begin to put something like this and it would probably take me years to learn the process.... so I just have faith in Oleg & 1Cs attention to detail. The man has proven himself to me time and time again as far as I am concerned.. so even if something is a tad off.... I trust his judgement moreso than I would someone elses as far as thier doing thier best to render it as close as they can to actual data. Ive seen so much conflicting data... Ive seen guys bring up charts and pictures for one plane or another to prove opposite points.... like the FW 190 cockpit thing. Ive seen pics and read reports that went both ways. When I take into account the differences in pilots,climate etc... Ill just trust 1C and enjoy the mess out of the product.

DD_NL
10-14-2004, 02:22 AM
This is a great topic. I hope all those whiners that complain about the game not having this or that airplane or flyable read this. Maybe they get an idea how much work is involved in making a flightsim. http://forums.ubi.com/images/smilies/mockface.gif

Gunner_361st
10-14-2004, 06:56 AM
Excellent thread. It really leaves you in amazement at just how much work in required to get a high-quality cockpit for an airplane done for this Sim.

Hats off to you Zippo and all the other 3d Modelers who take the time to give us non-modeling folks an idea of just how lengthy the process is.

I have two questions; How much does the 3d Max Studio software cost and just how powerful does your computer have to be to run it and display all those models and such?

Jippo01
10-14-2004, 07:54 AM
The local price quotes and a demo of 3d studio max can be found from publishers website:

http://usa.autodesk.com/adsk/servlet/index?siteID=123112&id=3422340

Computer power is not at all essential when working on a low poly model. I think 1GHz machine would run Ju 88 model just fine. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

Atomic_Marten
10-14-2004, 02:19 PM
Excellent.

Jippo01
10-20-2004, 02:25 PM
How does the cockpit fit in the game environent?

This is the final part then, in which we will look how the cockpit model fits in the game environment. We will also have a closer look on how the viewpoints for players are created and how their locations can be verified to ensure a realistic view.

There have been discussions about cockpit strut widths and how they are so wide because they have to cover the struts that are on the on the external model. Well this is not true.

External model is displayed whilst you as a player sit in the cockpit, but not the actual cockpit area. Only parts that can be seen outside the pit are shown including maybe parts like wings, tail and tailplanes, and engine nacelles if we are talking about multi-engine planes. Having external model cockpit parts in the view would just cause problems without giving anything in exchange, so they are not in.

Here is a view with the whole cockpit model attached to the fuselage parts that can be seen from different cockpit crew positions. Actually for importing the game the cockpit has to be further optimised with deleting all the parts of the cockpit that cannot be seen from the exact crew position in question. This way we end up having several models for planes that have multiple crew positions instead of just one.

http://www.kolumbus.fi/jan.niukkanen/cockpits/fin1.jpg


Say we want to make a pilot position model out of the complete pit. We take the complete cockpit and we delete all the aft facing machinegunner positions and along with them everything else from the rear cockpit that is blocked out of the view by seat and other parts like that in the front part of the cockpit. Since we couldn't even see them in game, we wouldn't even know that have been deleted. But taking them out relieves strain on the memory usage and processoring. So more FPS for every part deleted - infact for every polygon deleted. That is why we will still go through the front of the pit as well and check all the parts for back faces that cannot be seen by the pilot, and edit them out of the model as well.

So how do we set up viewpoints in the models? Viewpoints are made using object called camera. Camera is not a 3d-objects as such, but it is actually what it sounds like - a camera for viewing the 3d-model around it. Camera can be rotated and tilted in the same fashion as the view can be in the game. In the game it also moves up, down and sides a couple of centimetres as a part of the camera shake and presentation of G-forces. But modeller can neglect this movement in all other ways but making sure that he wouldn't delete faces that could be revealed by it when optimising the model.

Camera object which shows the view cone:

http://www.kolumbus.fi/jan.niukkanen/cockpits/fin2.jpg


There is one fact that determines the camera position in the model more than anything else. It has to be in the line of sight through the gunsight in all positions where weapons are present. In pilot position it has to be straight behind the reflector sight, and in the defensive mg positions it has to be behind the line of front and rear sights. This is simple physics - to aim at a target your eye, the sight and the target have to be located on the same line. With iron sights there are two sights which both have to be on the line.

Cameras can be placed on the sight line very easily (cameras, sights and weapons highlighted in white):

http://www.kolumbus.fi/jan.niukkanen/cockpits/fin3.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/fin3b.jpg


The picture above is from the Ju 88 C-6 heavy fighter cockpit and shows a special case where the gunsight and the camera both are at a five degree angle downwards. Guns on C-6 were installed in a five degree downward angle, and this means that to hit the sight line must follow accordingly.

Modeller can if he wishes to do so can also use other means to verify the geometry in his cockpit. I used human 3d models to test the geometry and ergonomics in the pit. This stage is not by all means necessary, but can give additional security to the modeller. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

http://www.kolumbus.fi/jan.niukkanen/cockpits/fin4.jpg

http://www.kolumbus.fi/jan.niukkanen/cockpits/fin5.jpg

I hope that these posts may have given a bit more clearer picture of what goes on behind the scenes in this series of games. Maybe even som of you are tempted to try out your skills as a modeller. If you should have such ideas, I can not but to encourage you to try it out. It has been a bit hard at times, but I have enjoyed it immensily and gotten a lot out of it too. And I have to say that it is quite rewarding to have your own plane flying in the best ww2 flight sim of the world, I do get a kick out of still. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif and I know that when the 88 comes out.... http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

http://www.jagdgeschwader53.flugzeugwerk.net/ju88/Project/finalskins/A-4_summer_01.jpg


Cheers for reading! http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

FF_Trozaka
10-20-2004, 03:38 PM
Awesome stuff , Jippo, thank you for taking the time to explain all this, Great job on the bomber!
Salute!

Jippo01
10-21-2004, 07:07 AM
WoW! This board is hot these days! http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif This thread dropped almost 2 pages in 17 hours!

Bumbage for Euros. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif


-jippo

actionhank1786
10-21-2004, 10:32 AM
Hey Jippo, did they have versions of the Ju-88 that had solid noses packed with guns like the US and their B-25s and A-20s (i think i read somewhere they did) and if so, are you making one? That would be awesome

crazyivan1970
10-21-2004, 11:26 AM
Jippo, there are no words to describe appriciation for this, i made this thread sticky, i want everyone to see what it takes to make something flyable.
Hats off to you man, and to all other 3rd party modellers who make such a great contribution to this sim.

Will copy it to GD as well

trumper
10-21-2004, 12:34 PM
http://forums.ubi.com/groupee_common/emoticons/icon_smile.gifCongratulations and thanks,this ought to be made a sticky then when people start to whinge they can look at all the work that needs doing http://forums.ubi.com/images/smilies/11.gif