magneto Posted July 25, 2013 Share Posted July 25, 2013 Hi, I am watching some dynamics tutorials, and they are using standard Houdini volumes. If I use VDBs, would I gain a massive speed improvement when cooking? If so, is it simply a matter of dropping them in place of standard Houdini volumes so VDBs are created instead? I just don't know if all DOPs, etc work with VDBs without doing anything special. Thanks Quote Link to comment Share on other sites More sharing options...
Guest tar Posted July 25, 2013 Share Posted July 25, 2013 Do you mean smoke/pyro? Then the Gas microsolvers would need to support VDB. I don't know if any do. Quote Link to comment Share on other sites More sharing options...
Solitude Posted July 25, 2013 Share Posted July 25, 2013 You should use them when you can in sops -- they are generally faster to create, and you can also convert them to standard Houdini volumes if you need. We aren't using 12.5 yet, so we have the plugins compiled from the openvdb site, and I use it for pretty much all of my collision volumes, as it's much faster than using a standard isooffset for generating an sdf (even with high bandwidth values and converting to a Houdini volume), thanks to the multithreading. They are not used in dops for the actual simulations yet, but using them for sources and collisions works in 12.5 (afaik). 1 Quote Link to comment Share on other sites More sharing options...
magneto Posted July 25, 2013 Author Share Posted July 25, 2013 Thanks alot guys, yes I meant for pyro, smoke, etc. If they are not supported in those, do you know how will they be supported? I am wondering if you will have to specify the type of volume of if it will be automatically identified? I imagine auto method might be slower since it will try to cover both types? I will use them for sources and collisions then. Thanks again Quote Link to comment Share on other sites More sharing options...
Solitude Posted July 25, 2013 Share Posted July 25, 2013 (edited) In the future I would assume it would be as seamless as possible, much like it is now for some things. A quick example of 'seamless' would be that a node like a volume vop asks the volume primitive (whether it's houdini volume or openvdb), for some data at a given position or voxel index, and gets the data back from the primitive. It works for vdb OR a houdini volume seamlessly for us. I don't know if it's feesable or truly beneficial for them to use openvdb for the simulations themselves, there may be good reasons why a sparse grid doesn't work well for a pyro simulation, or maybe a situation where Houdini's volumes are better at certain things... I guess somebody from SESI would be able to provide better insight. Check out Johner's thread here btw: http://www.sidefx.co...t=vdb collision Edited July 25, 2013 by Solitude 1 Quote Link to comment Share on other sites More sharing options...
magneto Posted July 25, 2013 Author Share Posted July 25, 2013 Thanks alot Solitude, very helpful insight Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.