Bounding Spheres ....er ?

[Bwian]

Alpaca ... this is part of what is confusing me. Those are STOCK numbers from a stock unit. I have tried scaling up certain values, but I can't see any obvious changes. I need to be confident that the unit, when scaled up, is able to engage an enemy in the correct place, and is attacked in the roght way.

[Casuir]

0 on the z plane isnt ground level if thats what you're thinking. Just looked at one of the jannisarys and the bounding sphere looks to encompass the model in the base pose with the arms spread out, so i dont see how it'd have any effect other than basic collison detection.

edit, goms reply wasnt there when I hit the reply button, sorry for repeating him

@Gom, what should it encompass exactly, bwians model is fairly large with a fairly large weapon to boot, to get it all in you'd need a radius of about 7.

[GrumpyOldMan]

Hi Casuir

0 on the z plane isnt ground level if thats what you're thinking. Just looked at one of the jannisarys and the bounding sphere looks to encompass the model in the base pose with the arms spread out, so i dont see how it'd have any effect other than basic collison detection.

edit, goms reply wasnt there when I hit the reply button, sorry for repeating him

No problems, having someone confirm what I say just releases the pressure from me that I might be talking gibberish (again)

@Gom, what should it encompass exactly, bwians model is fairly large with a fairly large weapon to boot, to get it all in you'd need a radius of about 7.

If Bwian is using scaling to increase the size of his figures, I'm pretty sure that the bounding sphere would be scaled up as well, if so he should leave it at vanilla figure size. If he's creating a new figure it has to cover most of the body and the majority of the melee weapon. I don't think that the bounding sphere is just collision info because of the discrepancy between values for a swordsman and a pike man, and the increase in the z component for a mounted figure, maybe it's attack collision info?

Cheers

GrumpyOldMan

[Casuir]

Had a look at some of the figures, the x values seem to be reversed, I think I recall you saying something about medievals co-ords being reversed on the x axis somewhere gom, the converter catching this with regards to bounding spheres? The pikemen and lancer models spheres encompass all of their model including the weapon in the basepose, so the centre is going to be different for
every model. Quadrupling the values wont necessarily give you the best values. Some rough values for the bone giant bwian:
-0.244
-0.348
2.850
5.234

[KnightErrant]

The conversion between CA's convention and Milkshape is
change the sign of the x-component of vectors and reverse
handedness of triangles. The bounding sphere data is just
written to the model comment and read out again so the
numbers in Milkshape are still for CA's mesh convention.

[Casuir]

Thats not the case though if you make a new bounding sphere based on a new model, the x values on the model are reversed but not the sphere. Minor thing thats easy to do by hand though.

[alpaca]

I think the bounding sphere is important for arrow collision detection.
If you zoom in close and play on low speed you'll see that units get bloody but aren't actually hit by the arrows sometimes. That's also what I was hinted to by a dev, if I interpreted that correctly.

As for the sphere: Isn't it set in the comments, so it'd stay the same size when you just scale your unit's model?