RX3 Files Research by du_oro10

[du_oro10]

Hi to all guys, I'm here after a while dreaming with be able to edit hairs 3D rx3 and sadly I haven't even the minimal knowledge to achive that .

But if Rinaldo and Jorge have been able to, that means that it is possible.

So lets begin with my findings.

As many 3D files, rx3 files are composed by chunks, in wich is organized all information inside them (Read this for more info).

What I learn is that in the begin of the file there is this "header" of the file, where are listed all the chunks of the file.

I use the accessory_28_0.rx3 as example couse it is the smaller 3D rx3 inside FIFA files (referee intercomunicator).

Now, I learn too that the chunk where the Geometry (x, y and z vertex) and mapping (u and v coordinates) is found under this tag:

00 58 7A A1 begin
DF 9A EC 1E end

and that when you find the chunk, in some rx3 (I have no prooved every kind) like accessories and heads, the first byte, in its IEEE Single data you will find the vertex coordinate!!!!

I'll leave you some screens:

Rx3 header and geometry chunk tags:

x, y and z coordinates as IEEE Single value:

Confirmation with Jor1980 CFC:

Now, sadly, in hairs rx3, the IEEE Single value is not the coordinate as in the study rx3, but is a beginning.

I know that maybe some master editors already know this information and more but is good to let us all to know and colaborate to solve the hairs problem.


so now, I ask all of you guys who wanna colaborate to learn how to interpretate the u, v coordinates and the polygons definition. And then research them for hairs rx3.

 

[MrMeedolson]

Hey does this mean we could possibly edit ankle tape models etc?

 

[alex94]

du_oro10,good job,I hope that somebody can help you to make hair converter

 

[du_oro10]

Yeah I hope too, couse I have findings but don't know how to interpretate them

 

[Kurlboy]

mate and about balls do you know anything? is a dream for ball makers restore the old classic 32 panel ball (any ball can be done in that type of model)

 

[MichaelCorleone]

Went straight over my head mate . I would really have loved to assist if I had any knowledge of this stuff. It looks like some deep mathematical S*it to me.

Whats this software you are using?

Can you elaborate a little more. I could understand what you said about chunks and about different aspects of hair located at particular places .. but nothing more. May be a little more explanation of what we are trying to get to may help me understand a little more.

 

[du_oro10]

Allright, this goes slowly than I think, in two days I just can learn one more thing, the number of vertex is declared next to the chunk start, as the sumatory of the hex values od the last two bytes from a 4 group. Once again here is the screen that shows more than words:

So we have this

Characteristics shared by "common" rx3 and "hairs" rx3:
-Gometry chunk start and end tags
-Number of vertex next to the chunk start as hex

Characteristics non shared:
-IEEE Single as vertex value

I'll continue researching.

BTW what I wanna do is to decipher the rx3 common and hairs files and share that information for programmers in the forum can create a converter like Jor's CFC12.

The utility I use is XVI32 and is provided with FileMaster

 

[du_oro10]

Hi this is me again with bad and goods.
First the goods.

It looks like I learn now where the u, v coordinates are represented in the "normal" rx3 files, just besides the previous shown x, y & z vertex values (start post), there is an identificator of 8 bytes (I'm using byte to describe two digits but dunno if it's strictly correct) and then 2 bytes for u and 2 bytes for v like you see in the screen below (red).

Now, the "word" (16 bits) is the most near representation of the coordinate, at first look 0.144 (the u from CFC) doesn't look like 12461 but putting all values on a graphic, the result looks promising:

The shape of the u coordinates and the u "as integer" (word) is very nearest and the v and v "as integer" is like the transformation y=C-x (flipped over x axis).

Now this may be the u and v coordinates, I'm trying to get the conversion factor between, so thats the bad, it is taking more time but I'll keep you informed.

P.S. in order of practicity I changed to wxHexEditor

 

[darxxx]

Hopefully you'll figure sth out so all the research will not be in vain.

 

[pao4ever]

dafuq did I just read...?!?!?!

 

[( . )( . )]

i have a question, how do you make the hair fit the head? i'm using your method of making faces but the hair doesn't fit hte head..

 

[Sepak]

i have a question, how do you make the hair fit the head? i'm using your method of making faces but the hair doesn't fit hte head..

it didn't supose to do that, the tutorial was for make perfect textures

 

[alex94]

du_oro10,any news?

 

[du_oro10]

du_oro10,any news?

Hi mate, the thing is that:

I know that the 16-bits integer is related with the UV mapping, but can't figure that relation couse its not linear so I'm stucked at that point.

I hope can give you some news soon

 

[alex94]

I hope too
It's very good that you already found some info about UV
Good luck)

 

[( . )( . )]

BUMP!

up for this thread