Submerged Units

[TARevenger]

with the TASpring engine, will units on the water surface eg boats, hovercraft be able to move over top of submerged units if they are deep enough.

and being able to dive subs to the bottom would be nice

[Buggi]

Yeah, I would like to see this too. Not sure how they could do it programmatically.

[Sean Mirrsen]

They already do, unless you've noticed. If a unit's colision sphere does not touch another's, they won't bother each other, same height (or depth)or not.

[Cheery]

Also smaller infratry units should be able to move trough the bigger artillery and be able to organize better. Now it's just so embarrasing when your Kbot unit says: not able to reach destination.

[TARevenger]

sorry to burst the collision sphere bubble But I don't see units going overtop of submerged units
if anyone does a video of a boat or hovercraft going over a group of submerged amphibious tanks or underwater metal extractors (without going around them) then I'll be happy

and frankly the collision spheres on big units for example the sea dragon and black hydra just make the unit a much easier target to hit.

[Kuroneko]

Need to ditch collision spheres and go to boxes, and I mean plural.

[[K.B.] Napalm Cobra]

Keep the spheres and add per polygon.

[Sean Mirrsen]

Do you have any idea WHAT you are asking, NC? With THOUSANDS of units on a map, and up to a thousand (in the worst case) faces on a unit, we're talking millions of faces. Imagine checking for collision of all that stuff! Spring already bogs down some not-too-weak computers with more than two players active, imagine what happens with 10 players, the classic TA maximum! And if there will be AI, further increasing processor load?

Now, I'm not totally against per-polygon collision detection (contrary to what you might have thought ), but something has to be done to simplify it before it can be used in Spring. One solution would be to have ultra-low-res pieces of the model that are nontextured, and reside in the base piece. Since the baseplates are no longer needed, and most OTA units (that use the base piece for pieces of the actual model) are low-poly, this would be the best solution for collision detection.

and TARevenger - most of submerged units' collision spheres touch the surface, and most of naval units' collision spheres go underwater. They simply can't pass over each other. Ships usually move above wreckage in the water with little problems (if the wreckage is indeed submerged, ofcourse).

[zwzsg]

No needs for detailed unit collision shapes. A single box per units is enough. A sphere does not fit the shape of most units, but a box, with each of the three dimension resized to graze the faces without any pieces sticking out, would work well I think. However I have no idea about how hard is it to go from a simple "distance between units center" to detecting collision between boxes. Maybe even that would cost too much CPU. Would it save some CPU to first do a simple sphere collision, and then if the sphere are colliding switch to a more precise algorithm?

Anyway, the current collision system doesn't work that bad for me, I never noticed any problem ingame with collision, and I already played lots of games.

[SinbadEV]

dude...zwzsg...

Would it save some CPU to first do a simple sphere collision, and then if the sphere are colliding switch to a more precise algorithm?

I think thats a great idea, that sounds like it would work great. When 2 spheres overlap (weapon/unit, unit/unit, feature/unit) it would then go into more presise calc for those units. I wonder if anyone has thought of this before.

[SJ]

Collisions between ground units (including boat/subs) arent handled by the spheres but with what squares they take up on the groundBlockingMap (iirc) so no going above subs at the moment im afraid.

[el_muchacho]

Anyway, the current collision system doesn't work that bad for me, I never noticed any problem ingame with collision, and I already played lots of games.

I've seen units going into walls, though, in the MetalHeck map : when the slope is very steep (alnost like a wall), the units sometimes go into them.

[Sean Mirrsen]

SinbadEV - it would work great, although the spheres don't fit most units, so the same with boxes instead of spheres would be even better.

[zwzsg]

el_muchacho: It's not a collision issue. The issue is that units are slowed down, but never blocked by any slopes. I posted somewhere I thought this was to let them go out of any craters, and suggested to let them climb any slopes only if the original slope, of the pristine map, on that point, was already passable.

[AF]

I think that box collision over sphere is somethign that would at least give an increase in many possibilitys, suhch as bridges as features, platforms etc.

Also I ahve noticed that I dont need to click on the commander to select him, I can simply click 'near him' to select. The problem is 'near him' is not always what th eocmputer thinks, for example if I try to select the commander, I may end up havign the unit infront of him selected. Thus am I right in saying that the collision spheres have soemthing todo with selection and thus may cause mooe of a problem?

Also, spheres dont make for very good moddability when dealing with long thin objects such as barges or towers due to the large sphere they create, thus spring has further restraints to 'flatten' the game contrary to the new innovation of the Z axis to the AT community.

OTA used boxes and not spheres..........

[TARevenger]

Sean Mirrsen with the water depth set to 75 naval/hovercraft units wouldn't beable to move over submerged units (as in on the ocean floor)

setting the water depth to 100-150 should allow units to move over top of submerged units on the ocean floor but they can't. (I can't check the spheres underwater)
It's like they run into a brick wall and go around

and with subs at their current set depth you can't go over top of them
they should be able to make use of the 3d plane just as aircraft do.