nicked on 15/10/2016 at 21:37
This is more out of idle curiosity that any particular mission plans, but has anyone ever made a hostile, combat-worthy rat?
I was messing about with it, and the best effect I got was to make it a non-exploding frogbeast with an Act/React source for a bash stim of 1, on contact. It's not very effective though, and doesn't seem to be able to damage other AI at all.
I tried making it work like a small spider too, but he tends to bounce around about 7 feet off the ground.
I guess to really work, it would need a rat AI mesh model made to conform to an existing skeleton, but that's easier said than done.
Nameless Voice on 15/10/2016 at 22:03
You could check how the grubs are done in SS2. Those are also jointed models rather than meshes.
They have a "grub" AI which probably isn't in Thief, though.
R Soul on 15/10/2016 at 22:10
A contact stim won't work well because the AI/Garrett phys model is two spheres, with the lowest just touching the ground, and the rat's phys model is tiny.
I think the best approach would be to have it fire an invisible projectile, but with a very low 'ideal distance', which would simulate biting.
ZylonBane on 16/10/2016 at 04:12
Grub damage in SS2 is a contact stim. Their physics is a single sphere of dimension 1, Z offset 0.9.
nicked on 16/10/2016 at 08:25
Thanks guys - spent a bit more time messing around with it.
I also took idle curiosity to it's logical conclusion and made a rat monster to try and fit the frogbeast skeleton:
Inline Image:
http://i857.photobucket.com/albums/ab140/nickdablin2009/ratmonster_zpskicjooo0.jpgUnfortunately, the 3DS for the frogbeast that I got from Shadowspawn's site doesn't seem to work. Even exporting it directly with the original frog model gets 3DS to bin throwing a wobbly. I get a huge number of "vertex [number] [coords] not assigned to seg" errors. Is that something I can troubleshoot or are frogs a bit broken generally?
Nameless Voice on 16/10/2016 at 18:25
Sounds like the limit planes are missing?
nicked on 16/10/2016 at 19:48
It's got all the limit planes and joint boxes present in the model; maybe they're labelled wrong from what 3ds to bin is expecting or something.
LarryG on 16/10/2016 at 21:09
It also depends on where the polygon vertices are with respect to the limit planes. If you miss-position the vertices with respect to the limit planes you will get lots of errors. One approach is to create a small box and then move it to the vertex location (x, y, z) indicated in an error. That should let you isolate the troublesome vertices and move them until their errors go away. Repeat until all the errors are resolved.
Alternatively, you can randomly try tweaking the limit planes and maybe get lucky.
I think Jason Otto Schwaa had a (
http://thiefmissions.com/lpg/ImportMesh.html) ImportMeshTut with a small troubleshooting section. You can download it (
http://thiefmissions.com/lpg/Tut/ImportMeshTut.zip) here.
nicked on 17/10/2016 at 05:56
If it was one or two, yeah, but I'm not doing that 400+ times. Pretty sure that is all, or nearly all, the vertices in the whole original frogbeast model, which certainly looks like everything is in more-or-less the right place. Thanks for that link though, that's a clear explanation of the problem if I come across it again in smaller doses.
Anyway, not to worry, maybe I'll just make a giant humanoid rat instead. :ebil:
nicked on 17/10/2016 at 18:01
Converting meshes full stop seems a bit janky - it seems that 3ds to bin will fail if any of the limit planes' edges end within the model's geometry. Which I could understand - except half the original Thief models break that rule.
