Topic: Discussion on Models

[Starfox1701]

Recommending Blender as the preferred DCC platform for models is OK but I think it would be good to develop export procedures for 3DS Max too. While that program is not as common because of the price tag many of the various communities best modelers use it. Doing this would encourage them to participate in this project.

[[UFP]Exeter]

As they would have good experience if we specified the requirements then they would be able to adjust 3dmas.  For example the coordinate system, orientation, scale factor etc.  I have access to an older version of 3Dmax, so can check it out later.  Worse case blender can import 3ds and export .x

[[UFP]Exeter]

I am almost at a point where I need at least 2 models to work with.

But, that means in our proscribed format.   Which is .X with dds textures.  The models can be created in blender and exported to .x   If possible I would like one done in blender to work out some details.

Also, we need to determine the details of the models in blender.  for example, does the bow point toward the +x or what?   The standard size of a model?  Naturally the model can be scaled to reflect the relative sizes.

If there is a way to read a model and now this information, I have not figured it out.

[Tus-XC]

front should be +x, it should roll around the x, pitch on the y axis and yaw about the z axis.  as to model size, 1 to 1 is always nice as it removes the need to figure out scale.  Failing that 1/20th scale is how sfc did it.

[[UFP]Exeter]

Interesting, checking some other sources indicate the rule of thumb (maybe) is for the model to point toward the -x axis.   But I am fine with your suggestion, I just need a standard from those experienced in this.

As for scale, my concern is two different models, the NCC1901A an and Klingon D9 Battlecruiser.  And due to the size of the models the battlecruiser display 10 times larger the the Constitution class.  Again what it is does not matter to me, jsut need to know to build it into the code.

[Tus-XC]

its just the way my head works, + x is the direction of travel for me, but if -x is the standard, that's fine.  My main concern is how the axis are used, roll about x, pitch about y, and yaw about z.  Scaling of the model is done by the modeler btw.  a connie and a d7 or a d9 have an actual scale to them.  for instance, a connie is about 189 m in length.  What i'm referring to is how the game engine will take that information.  will I need to change how 1 unit is registered by my software.  for example, 1/20th is my current setup (1 unit is 20 m)

[Tus-XC]

and what format of .dds should I be using?  there are a gazillion of choices in the dropdown when i try to save my file in .dds lol

[[UFP]Exeter]

that is logical given the standard coordinates for aerodynamics.  I will convert as graphics is different.  But as long as we have a standard I am good with it.

I checked out a few models an it appears the axis orientation between studio max and blender is different.  The default for milkshape seems to fit your description.  And I know that the axis can e changed.



The formats are the compression level.  Compression DD5 / AT12 (3Dc)  if using Gimp.

[[UFP]Exeter]

Found the math.

Yaw is counterlockwise around z

Pitch is countercloskwise around Y

Roll is counterclockwise around x.

And we will use +x for the front of the ship.

So pick a scale and define what it means and we will use it.

[Tus-XC]

nope, +z is up (away from earth), x is direction of travel and y is perpendicular to the plane those 2 form intersecting both those lines (at least that's how it is when dealing with satellites, which I'm fairly confident is the same for aero stuff)

edit:  opps, ya edited just as I replied

[Tus-XC]

that is logical given the standard coordinates for aerodynamics.  I will convert as graphics is different.  But as long as we have a standard I am good with it.

I checked out a few models an it appears the axis orientation between studio max and blender is different.  The default for milkshape seems to fit your description.  And I know that the axis can e changed.



The formats are the compression level.  Compression DD5 / AT12 (3Dc)  if using Gimp.

guessing that would be dxt5 in photoshop w/ an interpolated alpha... any particular way you want the textures handled? one sheet, two sheet 3 sheet...   I'm used to making one map for the diffusion and one for illuminations...

[[UFP]Exeter]

However you choose.  The system will handle up to 9 textures.  If you check out irrlicht that is what I am using for the 3D graphics.  Too lazy to learn Dx 11

[[UFP]Exeter]

You realize of course blender uses the right hand system and dx uses left.  But I think I will build the conversion into the program.

[Tus-XC]

Ok, here is a model, it simple (its just a planet) but it uses inverted faces to simulate the atmosphere (1 sided mesh, so you can see through it to the other face) which I imagine is pointless in your engine, but is something I do for sfc.  textures are 4096x2048 - its my working size, so you've been warned.   this is 1/20th scale btw (1 unit = 20 m) and the planet has a radius of 6378 km (that would be earths radius btw...)  so its large... (well not poly wise, just scale)  btw, if you dled already, that one was a radius of 6378 m (realized I used the wrong value after I exported...)

https://www.dropbox.com/s/bl3lfcs72klcbf0/Test%20Planet.7z

[[UFP]Exeter]

Thanks, however to get the model rotations working I need a defined front and top.  Maybe markers?

[Tus-XC]

copy that, i'll add markers to the texture

[Tus-XC]

new file uploaded, textures have axis labeled

[[UFP]Exeter]

thanks, that will help alot!

Did you use Blender for this?

[Tus-XC]

no, 3ds max 8

[[UFP]Exeter]

I have an older version lying around but never took the time to learn it.  Decided to hone my skills in programming.  And not really an artist. 

Would you mind publishing your specs (orientation, scale etc), it works and I can use that as the basis.

Thanks very much