Jump to content
Sign in to follow this  
MaxTux Wonder

Problems with physics and mesh

Recommended Posts

 

screenshot24.png

When i try to get a physic shape of my model i obtain red spots on the mesh preview, i guess it's becouse the point are too closer each other (this is what i guess but not sure about it), so i tried to upload a physic mesh i made it cutting the tree branches:

screenshot25.png

 But this is what i get the physic mesh is deformed, and i've no cue why it is. the mesh is the same, and if i try to upload the .dae as a mesh (and not as a physic mesh), it loads fine.

 

I'm using SecondLife-i686-2.7.4.233473 on ubuntu linux, mesh exported with blender 2.58 

Share this post


Link to post
Share on other sites

The physics shape mesh (like the low LOD meshes) is stretched to fit the bounding box of the high LOD mesh. So you need to keep at least the most extreme vertex in each direction for each axis to keep it the right size and in the right place.

That apart, you need something very much simpler for the physics shape, whether you are going to decompose it or not. As you have it, the physics weight vwill be horrendous. I would suggest a single six-sided cylinder for the trunk ans a single flattend square pyramid for ceach major branch, making sure something touches each face of the bounding box. The principle for physics shape should be the simplest possible compatible with function, for performance reasons if not for cost.

If that is the gigh LOD mesh, that is going to be a very expensive tree. It loooks like it could be drastically simplified without losing much essential ndetail.

Share this post


Link to post
Share on other sites

thank you very much Drongle for your reply, but i need a very complex physic shape for my model, so i think i will try to add cubes at each vertex of the bounding box reference just for preserve the proportion of the physic shape.

* Edit:

Anyway, the point to upload a physic mesh is to create an exact shape were the avie can move, the physic mesh can be different from the original shape for many reason, for example i can create a mesh with details that i will not include in my physic shape. for an experienced 3D modeller it's easy to work around with a boundig box references.

Anyway a bounding box reference is an artifact, not a optimal way to work, consider to fix this bug. Added on Jira:

https://jira.secondlife.com/browse/CTS-661?

 

Share this post


Link to post
Share on other sites

Yes. It is quite annoying. There is one trick you might find useful. If you use flat planes with normals facing towards each other, the solid decomposition will take them into account for fitting the physics mesh to the bounding box, but it will omit them from the actual physics shape. This is the trick I used to get an offset door hinge without having an unwanted physical obstruction. It is described in one of the files with the sample door on the wiki. It still works as of last week.

 

  • Like 1

Share this post


Link to post
Share on other sites


Drongle McMahon wrote:

Yes. It is quite annoying. There is one trick you might find useful. If you use flat planes with normals facing towards each other, the solid decomposition will take them into account for fitting the physics mesh to the bounding box, but it will omit them from the actual physics shape. This is the trick I used to get an offset door hinge without having an unwanted physical obstruction. It is described in one of the files with the sample door on the wiki. It still works as of last week.

 

Thanks for the information Drongle.  I looked for the sample door on the wiki and could not locate it.  Could you please give more information to where this is located.  Ive been wanted to learn this trick for a while and have not found any information until you mentioned it here.  Thanks in advance.

Yeso

Share this post


Link to post
Share on other sites


Yeso Kidd wrote:


Drongle McMahon wrote:

Yes. It is quite annoying. There is one trick you might find useful. If you use flat planes with normals facing towards each other, the solid decomposition will take them into account for fitting the physics mesh to the bounding box, but it will omit them from the actual physics shape. This is the trick I used to get an offset door hinge without having an unwanted physical obstruction. It is described in one of the files with the sample door on the wiki. It still works as of last week.

 

Thanks for the information Drongle.  I looked for the sample door on the wiki and could not locate it.  Could you please give more information to where this is located.  Ive been wanted to learn this trick for a while and have not found any information until you mentioned it here.  Thanks in advance.

Yeso

Well, with some more searching I found the information.  I just searched for sample door in wiki search.  Thanks so much.

Share this post


Link to post
Share on other sites

i know a way to fix this issue, when you upload a "visual mesh" you can consider the highest LOD to create a physic mesh right?
If you allow to upload a "physic mesh" the same way i can upload a "visual mesh" i can simple use the highest LOD of the uploaded "physic mesh" (uploaded as a visual)!
Do not really need to fix a bug, just permit to use the highest LOD of a physic mesh uploaded as a visual mesh...

Share this post


Link to post
Share on other sites

i've found your simple door object on wiki: but the flat plane linked to the door has just one face, you can look at the normals directions, and actually i don't think i can create a flate plane with two normals facing each other in blender... i wrong?

screenshot41.png

Share this post


Link to post
Share on other sites

Sorry, I guess I wasn't very accurate about that. The door only has the one flat plane, and I think it works whichever way it faces. Actually, I think I should have said you have to avoid the normals pointing away from each other. That's when the decomposer will fill in - when they look to it like the inside of a box. So that's a bit less restrictive than saying the have to point towards each other. I only used two planes in one set of experiments, so I'm not absolutely certain about this. That still means the planes can't intersect, as intersecting planes will always have normals pointing away from each other somewhere.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...