This is a model format for games like One Piece Grand Adventure, One Piece Grand Battle 3 and One Piece Grand Battle! 

 

in hex editor you might find formats like .act2, .gpl2 , .tpl2 and .skn2 within the gmf1.

 

gmf1 https://drive.google.com/file/d/17PgPGv9tnwmhXHC10CN6TWnShAhdOACk/view?usp=sharing

Lots of small point clouds where a mesh is hard to build:

 

Feet:

Edited September 25, 2025Sep 25 by shak-otay

 

17 hours ago, shak-otay said:

Lots of small point clouds where a mesh is hard to build:

 

Feet:

I see, that's not bad at all... I have another simpler sample model, This one is supposed to be arlong's sword kiribachi

gmf file https://drive.google.com/file/d/15a6GbzLHHBaAOXE_WHZUYGf3idVhKLfi/view?usp=sharing

The model i provided previously was luffy's model

The vertices exist in short signed format. They are small and widely dispersed rather than clustered. Each coordinate is spaced at intervals of 2, but I'm unsure of the face index type—perhaps TStrip?

For the 30 kB gmf1 you could collect the vertex blocks using the search signature 100000 0000xx010001, with xx= 02, marking start and end

(xx= 01 is for some other blocks)

but as long the face indices are unknown it doesn't help that much.

 

 

Edited September 26, 2025Sep 26 by shak-otay

On 9/26/2025 at 5:32 PM, shak-otay said:

For the 30 kB gmf1 you could collect the vertex blocks using the search signature 100000 0000xx010001, with xx= 02, marking start and end

(xx= 01 is for some other blocks)

but as long the face indices are unknown it doesn't help that much.

 

 

I see... if that's the case, could you take a screenshot of hex2obj's entire window where you managed to get some of luffy's feet? so i can try and continue from there?

Edited September 27, 2025Sep 27 by cheeselake0

Here you go:

edit: btw, the FIs don't fit as I wrote already - thus I let hex2obj create them automatically

Edited September 27, 2025Sep 27 by shak-otay

On 9/26/2025 at 4:08 PM, bobo said:

The vertices exist in short signed format. They are small and widely dispersed rather than clustered. Each coordinate is spaced at intervals of 2, but I'm unsure of the face index type—perhaps TStrip?

Yes, that is the case since the .gmf contains armatures, textures and the static model itself.

Has anyone been able to find any leads on this? . Gmf1 is a model format that contains models and armatures, where .gpl2 is the static model. 

Well. I fear, nobody wants to do the boring job of analyzing.

Btw, giving binary files a structure (which may ease analyzing) can be done by anyone.

0x70B0 is the (or one) start of vertices here, see picture  in my previous post: