sekow
-
Posts
139 -
Joined
-
Last visited
-
Days Won
5
sekow's Achievements
-
yeah, it should be applied locally where needed. which is possible with the gas particle separation dop as we can apply it to a group or pop stream.
-
to correct myself, that would result in added surface tension. as those points go towards their average. and not spread out
-
Considering the uniform distribution. Are you just averaging the "about to rip" areas and blend that with @P with the freeMask as bias? I am trying some like that right now, but struggle with the right distribution formula.
-
just did this today.. cached out vel and piped it back into a dop net with upres solver. but build it up from scratch
-
HQueue Substep caching of geometry NOT WORKING
sekow replied to Poetsofthefall's topic in General Houdini Questions
yep, hqueue does not support subframes write out. has been a long standing rfe. quoting sidefx support: This is a known issue. HQueue at the moment only works with integer frames. It's a pretty big sweep to fix, I will not expect it to happen very soon. -
I dont why I assumed that an hda or scripts folder inside $JOB would be recognized automatically, but I fear that one has to take care for those folders to show up in scanned asset libs..
-
$JOB/hda, $JOB/scripts wont work for you? EDIT: scratch that.. otls are not recognized..
-
How to get better at creating complex effects for shots?
sekow replied to magneto's topic in Lounge/General chat
break it down. like every problem is better to disect it into smaller problems. start with to one you most comfy with. build up from there. identify whats important and whats not. what comes first? (thats maybe the hardest part) most times the next step is pretty obvious, and I bet you will have a good picture of progression pretty soon sketch out in your mind. thats why houdini is great at that kind of work. your effect can contain on bigass setup or, which is way better, is coming to live thru individiual blocks/modules. -
please do and let support know!
-
I cant really observe that here. dual 12 core xeons. with a usage of around 50 % (of all 24). I think the slow down comes from that vdb loading out of sops. Did you observe ram peaks? 64 should be enough, but Houdini on windows aint that good with flushing ram, so its does accumulate in time. and if your is full windows starts to swap, and that will result in just a core working.
-
thinking of that a second more, one could hack the grain constraints to achieve a better result and even attach that to something... interesting task indeed
-
had to try that out myself, here is my attempt: snot.hip
- 31 replies
-
- 2
-
-
- constraint
- attach
- (and 3 more)
-
while we are at it. Ive never understood that proxy thing. When we volume sample vdbs from sops, isnt that redundant? In the end dops transfer the sdf into a dop object out of houdini volumes.
-
cant really help without the alembic file. but I assume that dops does not know the center of gravity. you could fill out the pivot parameter in the initial state tab with the centroid() expression. centroid(opinputpath("/obj/sphere_object1/dopimport1", 0),D_X), centroid(opinputpath("/obj/sphere_object1/dopimport1", 0),D_Y), centroid(opinputpath("/obj/sphere_object1/dopimport1", 0),D_Z), good luck
-
what you need to check is how rbds work in dops. when you import your animated geo into dops, dops assume the first simulation frame as inital state of the geo and wont refresh the geo. In your case you want you should do some reading in the docs. In a nutshell: RBDs are set to active by default, so the solver know that object needs to be solved. But you have the possibility to set it to deactive (i@active == 0). And when using bullet your are able to set an attribute "@animated" to 1 to make dops aware that it should refresh the animated geo. Check the file for a crude example animated_rbd.hip