xbox bsp editing

[zero_cool]

bump, come on guys this is huge stuff, lets see some input!

[mmmjoy]

In the NTSC Bloodgulch offset list, the structure BSP is listed to be at 0x023492B8, does anyone know if this is the visual/collision?

Also, whichever it is does the other come directly after or are they at seperate locations?

Does anyone know the sizes of these? It doesn't seem to be listed in the offset list and objects aren't listed sequentially.

and (finally) does anyone know if bloodgulch.map is included in the HaloCE download or if it uses the PC map (and can I get to it out without installing stuff?)

thanks

[Talin64]

grenadiac said

Ok, the way the bsp header is located is a little strange, but here is how it works.

At the end of the scenario, there is a section that contains the bsp offset and bsp magic (there can be more than one set). If you subtract the bsp magic from that offset (this is done for you in Sparkedit - the dump offsets function), you can locate the next clue.

This location is not the bsp header. It is the offset that points to the bsp header. Like this:

Scenario_bsp_info -> header_locator -> header

There is a ton of stuff in the header (probably 20 plus reflexives), of which I have only identified a few parts (weather, visible mesh info, etc). However, one of those reflexives points to the BSP Collision header.

The collision header contains 8 reflexives, of which I've identified 3 parts: nodes, planes, and vertices. The other 5 I am not really sure about, but I think that one is probably leaves.

The problem with modifying this stuff (which BOLL and I did a long time ago), is that the BSP tree gets screwed up because you need to update the planes. But if you update the planes, then the ordering of the tree gets messed up. The only real way to deal with it (I think) is to make a bsp compiler to rebuild the tree properly. I could be wrong, there may be a hack around this problem. I haven't spent much time other than updating the vertices section. (By the way, that verts section is the 2nd one that you were referring to, the first section is for the visible stuff).

[mtxboxfreak]

anything further on this?

no posts here for some time... did u all give up?

[xorange]

anything further on this?

no posts here for some time... did u all give up?

Well, no...I, personally, haven't given up on the concept yet.
It's just that what we were doing wasn't quite working.

We were focusing our efforts on accurate re-calculation of BSP planes, based on modified vertices.
We've achieved this, and it's quite a good step...but, as could've been expected, just recalc'ing the planes isn't enough.
...or rather, it works perfect for the planes, but modifying the verts & planes causes other parts of the BSP to require recalculation.

Now that we have the HEK I don't think putting together a "BSP calculator" is really the way to go.
Considering the obvious complexity of the BSP tree, and the fact that we now have tool to compile BSP's for us, I think our time would be better spent working on a method of succesfully importing new BSPs into Xbox maps.
...or something even more amazing, like a complete map converter or Xbox map compiler.

Anyway, my main focus right now (& probably for a long time) is on transfering HaloCE content to Xbox maps.
I'm not presently working on the scenario or BSP, but I'm sure I will be eventually.

As far as I'm concerned, this thread doesn't need to be stickied.
It's a very interesting thread with a lot of good information for anyone interested, but it doesn't really lead anywhere...and I don't think it will.
...but that's just my opinion, & Talin64 might have a different one. Idk.

If it makes a difference I'd much rather have the
HaloCE content on Xbox thread pinned,
because I think it's going to be a much more informative & active thread.
That's where I'll be actively posting my progress anyway, & I hope other people will too.