Brute Force (XBOX 2003) - Models

[shakotay2]

Tried to copy the process you used in the brutus picture example where you said to delete the lines in the hex data of the characters to fix them, before renaming to .OBJ but the model never changed.
Had done this on a character from another mission.

Which one?

[manbig343]

, and renaming the "objx's" to "obj" seems to either work or corrupt the file.

Well, it's a very simple fix:

example: in objects-common_7_%_.objx delete the marked lines (<<<) then rename from .objx to .obj:

# 0xe118: faceIndices= 3932, triangles= 1310 (strips?)
f 6141/6141 5/5 7/7 <<<
f 5/5 1/1 7/7 <<<
f 7/7 1/1 6145/6145 <<<
f 1/1 24225/24225 6145/6145 <<<
f 6145/6145 24225/24225 1/1 <<<
f 24225/24225 2/2 1/1 <<<
f 1/1 2/2 3/3
f 4/4 2/2 5/5
f 2/2 6/6 5/5
f 5/5 6/6 7/7
f 6/6 8/8 7/7
f 7/7 8/8 9/9
...

f 1919/1919 1807/1807 1920/1920
f 1807/1807 1921/1921 1920/1920
f 1920/1920 1921/1921 6141/6141 <<<
f 1921/1921 5/5 6141/6141 <<<
f 6141/6141 5/5 1/1 <<<
-----------------------------
result:
.
fix_objx.png

This one. I really hope we don't run into a stand still on the characters. By the way, we got every single weapon model now (they were never messed up) literally all we need to figure out is getting the characters in perfect condition, thats the next big thing. Thats why I hope you could make the special brute force hex2obj patch. Though I have no idea how hard or easy that would be to do so on your end for us. We just need brute force squads models (which tex and brutus are done, and many soldiers too that were fixed manually) and the rest of the characters from the game.

Then, perhaps if you do know, I can send you an animation file? maybe the skeletons exist within them. They have xmbs and ivds for the animation files. I wonder if they could then be extracted and converted just like the h20 files into something.

[shakotay2]

Thats why I hope you could make the special brute force hex2obj patch.

Pmed you (usually I don't create special versions, for good reasons).
(The only reason I did it now is that the advanced GetMaxFI feature could come in handy for the "official" version.)

There's still the "strips/no strips" issue. Maybe there's a simple flag which I couldn't find so far, anyways, the patched Make_H2O is my nearest hit.

Then, perhaps if you do know, I can send you an animation file? maybe the skeletons exist within them.

My spare time for Brute Force has expired now, more or less. So, sorry, no.

[manbig343]

Thats why I hope you could make the special brute force hex2obj patch.

Pmed you (usually I don't create special versions, for good reasons).
(The only reason I did it now is that the advanced GetMaxFI feature could come in handy for the "official" version.)

There's still the "strips/no strips" issue. Maybe there's a simple flag which I couldn't find so far, anyways, the patched Make_H2O is my nearest hit.

Then, perhaps if you do know, I can send you an animation file? maybe the skeletons exist within them.

My spare time for Brute Force has expired now, more or less. So, sorry, no.

Alright. I understand.

[manbig343]

it seems as though the brute force models have more vertices than thought initially?
Using this in Hex2OBJ by random guessing I nailed the full model of Flints face while having opened the .h20 file containing it

Using fake with these settings, also leads to her entire model appearing albeit covered in triangles. I think something may be happening that we didn't know before. I have followed and learned all the steps we were taught but they result to models missing mesh pieces, but looking alright beside that...but with this idk what to say or what this could really mean. The hair model also showed up but isn't made into an h20 at all.

[manbig343]

Here is the result from the hex2obj settings on the h20 containing flints head, which is objects-common.xmb_33

[manbig343]

The monster model that is loaded when using "fake" with these settings on the head h20

Note the hair appears AND her entire body.

[manbig343]

Aaaaaaaaand...

Full model of flint recovered (No hair yet, that has to be discovered specifically what area or whatever it actually is located in her h20/somewhere near its listing of data)

[shakotay2]

it seems as though the brute force models have more vertices than thought initially?

Some of them, yes.

Using this in Hex2OBJ by random guessing I nailed the full model of Flints face while having opened the .h20 file containing it

What I can say is that these are the correct parameters for the head which gives senseful uvs, "random guessing" doesn't make sense to me here:
.

flint_head.png

Crossing the "borders" to find the hair should be done different. I'd be surprised if "your way" lead to good results (for the hair).
Sadly I don't have the time for more research on Brute Force.

Also I think there's a separate H2O for the body, isn't it? (xmb_35.h2o is for arms and legs)
xmb_36.h2o is for the body, so no "random guessing" will help you here in any way.
And I don't know why you're saying: "Full model of flint recovered".

That sounds as if the patched hex2obj_BF.exe is working in a wrong way. I would beg you for showing more respect, sir.
You don't need to recover anything (except the hair, for example). Just use the corrected parameters hex2obj_BF.exe
writes as first line in the created obj file, if required.

Just in case you didn't get that. Read the instructions for hex2obj_BF.exe again, then.
--------------------------------------

BUT: there is a gap between 0xb0530 and 0xb3730 (where hands&shoes start) which is 400 vertices (FVFsize 32).
The point cloud seems to show the hair!

Well, where is the concerning face index block?
done: (click up arrow)

It's _78.H2O which will create the missing hair mesh (one face missing, afaics).

[manbig343]

it seems as though the brute force models have more vertices than thought initially?

Some of them, yes.

Using this in Hex2OBJ by random guessing I nailed the full model of Flints face while having opened the .h20 file containing it

What I can say is that these are the correct parameters for the head which gives senseful uvs, "random guessing" doesn't make sense to me here:
.
flint_head.png
Crossing the "borders" to find the hair should be done different. I'd be surprised if "your way" lead to good results (for the hair).
Sadly I don't have the time for more research on Brute Force.

Also I think there's a separate H2O for the body, isn't it? (xmb_35.h2o is for arms and legs)
xmb_36.h2o is for the body, so no "random guessing" will help you here in any way.
And I don't know why you're saying: "Full model of flint recovered".

That sounds as if the patched hex2obj_BF.exe is working in a wrong way. I would beg you for showing more respect, sir.
You don't need to recover anything (except the hair, for example). Just use the corrected parameters hex2obj_BF.exe
writes as first line in the created obj file, if required.

Just in case you didn't get that. Read the instructions for hex2obj_BF.exe again, then.
--------------------------------------

BUT: there is a gap between 0xb0530 and 0xb3730 (where hands&shoes start) which is 400 vertices (FVFsize 32).
The point cloud seems to show the hair!

Well, where is the concerning face index block? I dunno - you can't interpolate that because the blocks are assigned in an order I don't understand.
That's hard: you need to exclude all FI blocks in use, then try out the remaining ones for the hair's vertex block.

It's hard, because many of the FI start addresses from the xmb files don't match the ones from the 0000 0100 0200 search. Thus they have to be corrected. I could log the corrected ones, of course, then delete them from the afore mentioned search list... maybe next month

I didn't mean any disrespect, I meant that we have the model now using what you gave us.
We're diving deeper over time and now have Hawk completely.
You can take the time you need to rework things and whatnot. You did alot for us.

[manbig343]

The Squad is back together.

It was a long road, of tears, struggle, and desperation. Waiting.

Many steps. Learning, Trial and error. Guesswork, and searching.

All done by basically 3 people and a small amount of input from those who were here in the beginning.

Now all that is left is to bring everything into the asset pack I'll upload, and after years of a forum being lost, and then found -
The models are actually done. But there's still many more characters in the game, map pieces and hidden things.
Other file types, animations - skeletons waiting to be digged up. Hopefully we can manage to find others who can help for those.

I leave this here for now until I prepare the pack. This compressed image of a team that will return to the front lines very soon.

[shakotay2]

Parameters for the missing hair mesh (seems you've got them already):
.

Flint'shair.png

(Vertex address is in xmb file, so I need to check why I missed it.)

done, updated Make_H2O-BruteForce here (click up arrow):
.

[manbig343]

Parameters for the missing hair mesh (seems you've got them already):
.
Flint'shair.png

(Vertex address is in xmb file, so I need to check why I missed it.)

Thank you man. We had to use a hair model from another game. But i'll replace the one we had with that new one.

[shakotay2]

Ok, you may check the updated version of Make_H2O-BruteForce (see link in my post above), which creates more H2O files.

Keep in mind to treat _strips_ H2Os separately. So there's two sets of H2Os with identical params, only difference is 'strips' or 'noStrips'.

Thus each set will produce some weird obj files, which are correct for the other version and vice versa.
(Usually "strips" is the most used, afaics. So start with those "_strips_" H2Os.)

(This was the easiest way to treat this rather than searching for a strips_or_nostrips flag in the xmb files, if any.)

[manbig343]

In about a week or so I'll be sharing all the models we've got so far. Got a few more characters.