File Formats

[Avid29]

Can we agree to use it in what ever formats we want but we must upload it in .dae and fbx

[MadoxLabs]

Ultimately, the data needs to be in FBX format, either ascii version 7.3, or binary version 7.4 (2013 versions). The parser I wrote only deals with this format.

[Avid29]

@MadoxLabs I know of that requirement for fbx but dae is easier to works best to exchange between most editors so that's why we should make both

[wannefab]

I don't have time today to work on it but i allready started to also export .fbx for all .c4d files, i'll finish everything tomorrow...

[Avid29]

Thanks

[MadoxLabs]

@Firestarthirty Cool. Sounds good to me

[wannefab]

dae doesn't work well with C4D... gives no working skeleton... FBX is perfect sofar ^^

btw: i exported everything i did sofar as fbx, will do the rest tomorrow

[Avid29]

@Lithagon FBX 2013 formats? I have a converter I'm just curious

[wannefab]

It's FBX 7.4 (2014)

[MadoxLabs]

I got the ASCII fbx code integrated into PoGo now and I'll start on the binary fbx code next. The files in the repo here are binary so I can start giving feedback on those after the coding. I'll also see if I can handle 2014 formats as well at that time.

I tested the ascii fbx code on the original bulbasaur model. The problems are:

• multiple UV sets and multiple materials assigned to the same mesh. I dont handle this all but I can if we have to. It will take time so best to not depend on this. Stick to one material to a vertex. Unless you want me to support multiple. Let me know.
• model hierarchy. In this file, only three Models have Geometry. There is also dozens and dozens of Models and sub-Models in a hierarchy. I don't deal with this at all right now but I will probably have to in order to do animations. Animation in general is going to be a large task. just fyi
• texture hierarchy. This is bizarre and I've never seen this before but it looks like a texture with a filename is assigned somehow to a parent texture that has no filename. (I'm not sure the nature of the connection). The parent is then assigned to the material. I dont know what this means but would prefer to not see texture hierarchies. strange.

[wannefab]

• We can split up the mesh so there's only one material per mesh, that is no problem and takes not long (i sent you a split-up version right? It would also be under "Bulbasaur_split.fbx")
• The model contains a sekleton, so that's nothing to render, you can just ignore it i guess. (3 meshes is correct)
• hmmm that sounds strange... do the textures show?

how does the animation work on your side? i can export the fbx with animations baked to each frame...

[MadoxLabs]

Ya I have the split one, havn't tried to load it yet.

I don't have my old animation code with me to look, but ya, the empty models must be the skeleton. As I recall, an animation is a sequence of timestamped matrices tied to one of these empty models. The mesh's world matrix is the combination of all the parent models' world meshes. I'll get more details when I look at my code. Its been ages.

Right now I'm just loading the fbx, not rendering it :p

[MadoxLabs]

Oh the double texture thing is a LayeredTexture, which takes a Texture and adds the alpha factor to it, so I'll add support for that.

[wannefab]

ahh ok! Nice!

Alpha Support would be good, but the bulbasaur souldn't have any alpha (and we should be able to manage without i guess, but it could be nice for the flames for example)... Ahh if you use the sourcefile of bulbasaur it has a transparency channel on one of the textures (the iris)

[MadoxLabs]

Update - the binary parser is ported to C# and it hit some corrupt data block after "FileID". Obviously its not actually corrupt, but its some data block that doesnt conform to the expected 7.4 binary format. I have to do some research to see whats going on.

edit: looks like parts of the file are z compressed and the builtin c# DeflateStream doesnt deal with it

edit2: ok I can parse the binary file. Just have to use it now!

[wannefab]

ohhh cool! =D