VDB for collisions

   48399   30   10
User Avatar
Member
373 posts
Joined: March 2009
Offline
If you are using a standard houdini volume and use “volume sample” and proxy volume, it just fetches the volume, so there's no need to to save them in dops. No benefit afaik. You can use vdb and do convertvdb before sending it to dops.

On the other hand, and this may be a bug, if you are using vdb and use volume sample, then it will actually sample the volume to the resolution specified on the static object. It'd be nice to get this fixed / make it more consistent. Allow file to use bgeo, add a check box for proxy volume to “use volume directly” instead of sampling it…
Ian Farnsworth
User Avatar
Member
948 posts
Joined: April 2008
Offline
Solitude
It'd be nice to get this fixed / make it more consistent. Allow file to use bgeo, add a check box for proxy volume to “use volume directly” instead of sampling it…

That would be awesome
https://vimeo.com/user2163076 [vimeo.com]
User Avatar
Member
471 posts
Joined: Nov. 2013
Offline
hi i have problem in VDB:
when i create teapot and add VDB to it, part of my object disappear!!
if the voxel size parameter lower than 0.1 this problem happening!
how can i fix that?
thanks.

Attachments:
VDB Problem.jpg (77.2 KB)

User Avatar
Member
948 posts
Joined: April 2008
Offline
I would try to make that particular object a closed polygon shape instead to try to use the thin shell that it is.

I hope that helps.
https://vimeo.com/user2163076 [vimeo.com]
User Avatar
Staff
2540 posts
Joined: July 2005
Offline
To add, the teapot is not the best candidate object for direct conversion to a volume. It is comprised of many bezier surfaces just like the real original teapot.


Just append an Exploded View SOP to the teapot geometry to see all the various pieces, not a one is manifold (water tight closed geometry) so constructing a volume, collision or density is not going to work very well.

It is common to take input geometry and wrangle it a bit in order for it to construct a decent volume representation.

With the teapot, you can create a good version that will construct a good volume:
- convert to polygons, 3x3 divisions is good
- isolate the body, rim, lid and lid top geometry (select option, hit 4 for prims and 0 for connected prims helps)
- fuse the teapot body selection aggressively (0.01 works ok)
- append a Group SOP set to edges and only set edges unshared to group
- append a PolyCap SOP and feed in the above group.

Now you should have a decent set of three pieces for your teapot: body nozzle and handle that are all kinda manifold. See if that creates a decent volume.
There's at least one school like the old school!
User Avatar
Member
1391 posts
Joined: Dec. 2010
Offline
NimaNolan
hi i have problem in VDB:
when i create teapot and add VDB to it, part of my object disappear!!
if the voxel size parameter lower than 0.1 this problem happening!
how can i fix that?
thanks.

Also this problem occur when you use PointFromVolume SOP for teapot !
I think that's because teapot shape has some open separated parts (handle,door,body,pipe).

To solve that you have two option :

1. Cap all opened parts , Of course you should fuse them before Capping.
2. Add different VDB for each part and merge them with together.


I think you take better result with second method :wink:

Attachments:
VDB.jpg (94.6 KB)

https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com
User Avatar
Member
1391 posts
Joined: Dec. 2010
Offline
And this is a simple project about these methods my friend :wink:

Attachments:
JKVDB.hip (81.0 KB)

https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com
User Avatar
Member
1391 posts
Joined: Dec. 2010
Offline
jeff
To add, the teapot is not the best candidate object for direct conversion to a volume. It is comprised of many bezier surfaces just like the real original teapot.


Just append an Exploded View SOP to the teapot geometry to see all the various pieces, not a one is manifold (water tight closed geometry) so constructing a volume, collision or density is not going to work very well.

It is common to take input geometry and wrangle it a bit in order for it to construct a decent volume representation.

With the teapot, you can create a good version that will construct a good volume:
- convert to polygons, 3x3 divisions is good
- isolate the body, rim, lid and lid top geometry (select option, hit 4 for prims and 0 for connected prims helps)
- fuse the teapot body selection aggressively (0.01 works ok)
- append a Group SOP set to edges and only set edges unshared to group
- append a PolyCap SOP and feed in the above group.

Now you should have a decent set of three pieces for your teapot: body nozzle and handle that are all kinda manifold. See if that creates a decent volume.

Sorry , I didn't see Jeff's reply , He explain this problem more clearly :roll:
https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com
User Avatar
Member
471 posts
Joined: Nov. 2013
Offline
thanks really thanks my friends.
i need this for create safe border for collision with pyro fx and fluid sim and i completely understand.
User Avatar
Member
4 posts
Joined: Dec. 2014
Offline
Hi guys,

Every time a try to connect a volumesource to the flipsolver in the DOP (Initialize: collision), the flipsovolver ignores the fliptank limits. And all the particles go on the ground.

Is anyone experiencing the same?

Thanks!
User Avatar
Member
2 posts
Joined: Nov. 2016
Offline
mzigaib
Is there a way to use an OUTPUT DOP to save the simdata collision from inside the dop object?

I hope it makes sense.

Thanks.



I just found this supper useful threat… well, I guess you probably found the solution for this issue after all this years since this was posted in 2014… but for people out there who might come across this the way I get the particles out of the DOP network is by putting a dopimportfields outside the DOP network. In the dopimportfields node you add the DOP network path and in the Default Object you use your flip object.
Edited by AlexandraLópez - Oct. 14, 2017 00:05:47
  • Quick Links