Bounding Boxes

tjbrunetto

I didnt realize the colliders in unity could not be deformed with the spine gameobjects. I need to add bounding boxes that utilize weights. The weights need to be exactly the same as the images that have been weighted so the bounding box follows the bends and trajectories of the players images. How can I do this quickly without manually adjusting every single weight? Can I copy and paste weights from the meshed images?

TIm

Edit: I have created three seperate bounding boxes for my player. One for the upper torso, minor issues with stretching the bounding box mesh when the arms move while aiming.

The other two meshes are really solid and I have used them for the left and right legs.

Going to implement into spine and then test the collision detection of the upper body with objects in my Unity scene to determine if the left and right arms need their seperate bounding box attachments. Also interested in the performance requirements of the bounding boxes as I know that polygon colliders in Unity use a lot more resources then do the box colliders etc.

Further Edit: Have now added Bounding Box follower scripts to my updated spineGameobject. Now it seems that they colliders are greyed out so they are not working. Also they seem to move in relation to one another. Assuming they are pushing on one another after being placed onto my character. I also now have four rigidbodies on my character one on the main character and three on all of the bounding box child objects. Not sure how to proceed but will keep fiddling. Any help appreciated.

Erika

Hi tjbrunetto, unfortunately you can't copy and paste weights, but afaik a bounding box should be a less complex version of your game character so hopefully with fewer vertices and therefore less difficult to weight.

I'm curious about their performance requirements too, I'll ask someone more knowledgeable to shed some light 😛

tjbrunetto

Hi Erikari,

The setup I have is exactly the same as the videos you have posted about spineboy with the bounding box. I simply cannot access the polygon collider 2d. It is greyed out and when I delete all of the other components on my player I cannot get rid of the polygon collider 2d unless if I delete the entire gameobject.

I have ditched the large bounding box style for small bounding boxes that cover individual aspects of our limbs IE right forearm, head, left foot. The rotation from the bone follower script is working and they are 80 percent dynamic which is probably good enough for me. The only thing I am worried about here is that I am creating a ton of bounding box follower scripts, rigidbody 2d's, bone follower scripts, and polygon colliders (Which worries me the most). If I have about 15 body parts this adds up to a lot of extra information using the bounding box follower system with the bone follower system. Would using a skeleton hierarchy be more efficient?

Pharan

Hi, this question is sort of related, I think: Modifying a BoundingBox each frame
I'll just quote my answer there:

Note that Box2D (the 2D physics system that Unity uses) does not support deforming physics shapes.
Updating the shape of a PolygonCollider2D causes a destruction and rebuild on the underlying simulation, which they don't recommend, is not performant and likely results in unexpected behavior in many use cases.

However, if you know how to do manual hit and collision detection through your own code, it would work just fine. It just wouldn't be incorporated into Unity's physics.

If you need to work with Unity physics, you either need to respect how it works, or understand it enough to hack what you need into it even if it's not how it was meant to be used.

If you are following a lot of bones that pretty much follow the hierarchy, a hierarchy has better performance.
If you only need to follow one or two bones, you're better off with BoneFollower.

tjbrunetto

I have a different question. Now that I know that I cannot deform colliders; what about using the bounding boxes to determine the translation of the skeleton mesh. Then the box collider2d could follow the slot / bones in the skeleton utility and produce not exact but still rather accurate results of when my character would be hitting into a wall, enemy, etc. What would be the best way to go about achieving this.