Elite Dangerous

[prefim]

Can I just check are you unpacking the OVL / OVS files before looking at them in hexedit as I'm not see any vertex patten (alphabet) in there. Could you put up a few screen shots showing the marker points in hexedit please for a small object I can follow along and try for myself. Once I get one, I should be fine.

[shakotay2]

Vulture-I0-00000000.dat, first submesh:

vulture_dat.JPG

[prefim]

Thanks for that. I got it to work by following the 8b00 numbers you gave and got the mesh out as an OBJ. I also got it to work on the previous example starting at E5D8. There wasn't any UVs in those exports from what I saw. There seems to be multiple face indices in the files as well as multiple vertex lists too.

Still trying to wrap my head around this. Hopefully others, smarter than me, can pick up on this and see where it goes. maybe make a few elite specific tutorials.

[prefim]

Seems the 2nd block of data follows immediately after the first block (at least I thought). both your examples seem to yet I tried one from another mesh and got only partial success.

[prefim]

More digging this morning. It looks like there are multiple face indices, then a block of vertices. The face indices blocks reference this one large block of vertices.

Sorry if this is all noob stuff! trying to get up to speed so I don't have to keep pestering. Hex2Obj seems to do the job just fine once it gets the right set of numbers. Still not seen any UVs transferred though.

[shakotay2]

More digging this morning. It looks like there are multiple face indices, then a block of vertices.

searching for 0000 0100 0200 gives many finds, then to decide which are the correct face indices blocks' startaddresses. (That's what you do I hope.)

Hex2Obj seems to do the job just fine once it gets the right set of numbers.

That's correct, sire.

Still not seen any UVs transferred though.

Trying out the columns marked by small blue outlines as signed shorts gives a sphere,
so guess these values might be part of the normals (didn't bother).

You might try out other columns:

vulture_uvs_.JPG

edit: got the next submesh:


(First I thought I couldn't get the additional offsets other than by guessing. Then I saw they were contained in some magic table.
So once I've more spare time I will include it into the Make_H2O project.)

[prefim]

Magic table! that sounds promising!

yeah, the 0000 0100 0200 search was a great way to find the starting point. end points seemed less obvious as sometimes it just ends, other times it the number order seems to get jumbled a bit. Not sure if I should just trim those out or include.

It would be awesome if your software could scan the whole document (looking for the 0000 0100 etc.) and find the blocks for us. especially if you can now find the offsets accurately.

Looking forward to where this is going!

[shakotay2]

yeah, seems I finally got the correct uvs. They were simply too small sized (divided by 10000?):

Eagle_SM4_and_vulture.JPG

Can't wait to create a Make_H2O project but it's on low priority.

btw: would be nice to have a texture - at least for the vulture. (Hope it will look cooler than Han Solo's Millennium Falcon. )

[eriger777]

I can get the textures no problem. It seems that you guys have made a huge step forward. Would this be useful for a converter?

[prefim]

I own the vulture so ripping not a problem and with eriger777's ripper method for getting the DDS's out, maybe we can combine the two.

Even if H2O just scanned the files, and offered up what it things are the blocks of verts and faces, we could simply have a yes no option based on how it looks. the software basically says 'does this look right?' and if you say yes, it saves it off and moves onto the next bit.

[Vindis]

Hi,

do anyone still researching the model files?
I'm fiddling with it, the models are pretty easy to find, and found some data aswell, but most of it still unknown.
Sometimes there are padding bytes between face indices and vertex indices, sometimes there isn't, sometimes the vblock is 20, sometimes it's 24 even in one file (but the suspected vblock size value says otherwise tho).

So do anyone found anything?

Ripping files is a tedious task, and didn't even remotely as fun as creating a converter

Plus I don't have imperial Eagle

[Vindis]

My converter is getting into shape \m/

Now just have to figure out texture files, and UV coordinates, and ovls

[shakotay2]

Hi,

do anyone still researching the model files?

not really/unfinished.
I used the exe file from my ForzaHor3 project (which has EagleDeb in its drop down list) - view link in my sig -
to create H2O files from Vulture-I0-00000000.dat then hex2obj-024d (menu File/SaveAs Mmesh) to get obj files.

Most of them are not valid (-> .objx) but some work:

Vulture-I0-00000000_dat-red.jpg