3ds Max and collision

[pepperman35]

I recently experienced a problem with simple collisions. For example, I made a simple cylinder (Column) and then made a simple cylinder for its collision (c_Column). The collision is linked to the column and havok applied using cylinder as the shape. Exporting worked fine; however, when trying to open the NIF I get the following error:

 

encountered unknown block (bhkCylinderShape)

 

I repeated the above process using a box and sphere. Both of these worked fine. The cylinder however continues to be problematic. Pretty sure I have used a cylinder before so I am a bit perplexed as to why it all of sudden doesn't want to cooperate.

 

Anyone seen or experienced this before?

[nateneurotic]

I've never experienced that with havok meshes before. Have you tried just changing the type to mesh? Or making a convex hull for the cylinder shape instead of modeling a cylinder? I haven't investigated, but there's a chance that "bhkCylinderShape" isn't actually a think that nifskope or F04 recognizes, but I know that it does recognize custom meshes or convex hulls.

[pepperman35]

Yes, changing the type to mesh usually always works for everything. Unfortunately, it is the most computational option. But it will work. Thanks for the response.

[PJMail]

I have used cylinders and had no problems.

The fact that you have a bhkCylinderShape node type suggests that you are not attaching the collision to to an existing node/material object (so it has created a new node).

Is that possible or am I completely off the mark?...

[pepperman35]

I am not in front of that computer at the moment, but I am pretty sure I had it linked to the object it represent. I did a simple test afterward with a dummy, a square, a sphere and a cylinder to further test. Textures applied to all.

 

Dummy

SquareObject

c_SquareObject

SphereObject

c_SphereObject

CylinderObject

c_CylinderObject

 

For each of the object pairs (i.e., SquareObject, c_SquareObject) I did the rigid body with proxies route. Then selected the Dummy node to export. It exported with no problem. I tried opening it without going through Elrich to verify the collisions, and experienced the error. I will double check everything was properly connected in a bit.

[pepperman35]

Okay, so I believe I solved this. Using the above will create the error if you look at it before running Elrich. However, after running it through Erich it appears as it should and seems to have it collision intact. Didn't test it in game but it opens in NIFskope and it show the corresponding collision in the file structure.

[PJMail]

Interesting. I never looked at it in nifscope before running Elrich.

[pepperman35]

Looking at it in nifskope before Elrich always you to visually see the collision. Once Elrich does its thing, the collision can no longer be seen in nifskope.

[PJMail]

I never knew that - thanks! I don't suppose there is a reverse of that is there?

I would like to see the actual shapes of collision objects that make up a complex pre-existing nif...

I know I can see the outline in CK, but not what it was built from.

 

Thinking about it - the answer must be NO otherwise everyone would convert collision back to atomic shapes, manipulate it however they want, then convert it back.

Everyone would be happy to make collision updates - and I know they aren't...