Found 423 posts.
Search results Show results as topic list.
Technical Discussion » Crowd: using non looping animation clips
-
- cwhite
- 794 posts
- Offline
Have you tried the Agent Clip Properties SOP? That allows you to set up some information that the crowd solver can use when playing back clips, such as whether to do some blending while looping around
Houdini Indie and Apprentice » Crowd state transition
-
- cwhite
- 794 posts
- Offline
Technical Discussion » Agent layers from FBX with multiple variations of heads.
-
- cwhite
- 794 posts
- Offline
Yeah, that's the correct menu - the default shape names produced by the Agent SOP/ROP are just based on the node names in the source character rig / FBX file.
If Minimal Nodes is turned on, nodes that aren't visible and aren't bones/joints that influence skin geometry will be skipped, so you could try turning that off. Otherwise, you could send in the file to support so that we can take a look at it.
If Minimal Nodes is turned on, nodes that aren't visible and aren't bones/joints that influence skin geometry will be skipped, so you could try turning that off. Otherwise, you could send in the file to support so that we can take a look at it.
Technical Discussion » Agent layers from FBX with multiple variations of heads.
-
- cwhite
- 794 posts
- Offline
It should be able to - do those variations show up in the Shape menu when adding a new shape binding?
Technical Discussion » Pops - document for to explain targetv , implicit drag - how and when to use them
-
- cwhite
- 794 posts
- Offline
Technical Discussion » Crowd Queries
-
- cwhite
- 794 posts
- Offline
1) Not currently, there's RFE #73640 to add that (agents have a very different code path in the viewport from any other primitives)
2) There isn't anything yet built-in for that, as you found, but it basically consists of blending into the start pose of the recovery animation from the settled ragdoll. Some appropriate pin constraints on the ragdoll might be one way to approach that.
3) You could use multiple POP Steer Avoid DOPs instead of the embedded one in the crowd solver, and then apply different settings for different groups of agents
4) pscale is multiplied into the gait speed, which affects the agents' motion if retiming and/or limiting the particle speed to the gait speed range is enabled - otherwise you'd have very noticeable foot sliding. This could be made optional - are there cases where you don't want this behaviour?
2) There isn't anything yet built-in for that, as you found, but it basically consists of blending into the start pose of the recovery animation from the settled ragdoll. Some appropriate pin constraints on the ragdoll might be one way to approach that.
3) You could use multiple POP Steer Avoid DOPs instead of the embedded one in the crowd solver, and then apply different settings for different groups of agents
4) pscale is multiplied into the gait speed, which affects the agents' motion if retiming and/or limiting the particle speed to the gait speed range is enabled - otherwise you'd have very noticeable foot sliding. This could be made optional - are there cases where you don't want this behaviour?
Technical Discussion » APF Files from Massive
-
- cwhite
- 794 posts
- Offline
Google doesn't turn up much documentation on the apf file format, but you'd probably need to write an importer in Python / VEX
Technical Discussion » how to limit rbd velocities?
-
- cwhite
- 794 posts
- Offline
Technical Discussion » Möbius strip crowd sim
-
- cwhite
- 794 posts
- Offline
In 15.5 there are a couple new parameters, and various solver improvements, that make it easy to get agents adjusting their orientation to follow the terrain normal - the ‘FollowTerrain’ example file added in build 489 demonstrates this.
If agents are veering off the terrain because of avoidance forces or something like that, you may want a path force to keep them heading in the direction you want.
If agents are veering off the terrain because of avoidance forces or something like that, you may want a path force to keep them heading in the direction you want.
Technical Discussion » how to limit rbd velocities?
-
- cwhite
- 794 posts
- Offline
yeah, just wire it in before the Bullet solver. Or, if you're using the Rigid Body Solver DOP, use the pre-solve input
Technical Discussion » how to limit rbd velocities?
-
- cwhite
- 794 posts
- Offline
For Bullet, the speedmin and speedmax point attributes are supported. These match the standard POP attributes (e.g. as created by the POP Speed Limit DOP)
Houdini Indie and Apprentice » Crowd, POP & Steer Seek
-
- cwhite
- 794 posts
- Offline
You'll want to either turn off the Limit Particle Speed to Gait Speed Range on the Crowd State DOP, or increase the Allowed Variance parameter accordingly. This is what was preventing the agents from slowing down to a stop.
Then, on the crowd solver you can either decrease the minimum speed at which orientation changes are allowed to occur, and / or turn off the option to constrain the velocity by the max turn rate.
Then, on the crowd solver you can either decrease the minimum speed at which orientation changes are allowed to occur, and / or turn off the option to constrain the velocity by the max turn rate.
Houdini Indie and Apprentice » Question about crowds and orientation/heading
-
- cwhite
- 794 posts
- Offline
Marc Lauper
My first problem is that the agents (fish) can change their heading only in
the xz plane. Since the swarm is not constrained to a ground plane,
how can I allow the fishes heading (controlled by velocity) to change in all planes?
I think you want to allow the solver to modify the agent's up vector - this is controlled by the Adjust Up Vector parameter on the crowd solver.
Marc LauperI think it would probably be useful to see an example file for that issue
The second question:
- I set up a crowd using “Populate” and “Simulate” from the shelf.
- Then set “Initial Velocity”, “Heading”, “Up” to 0 in the “Crowdsource” node.
> As soon as I add a POP node (ie pop force) in the simulation, I also get an
initial force on the Z axis. Why is that and where does it come from?
Houdini Learning Materials » Simulate Crowd
-
- cwhite
- 794 posts
- Offline
For the other error, you should update to at least 15.5.493 - there were some issues in Apprentice due to the Channel ROP being unavailable
Technical Discussion » Issue setting up agent transition graph. help please?
-
- cwhite
- 794 posts
- Offline
Technical Discussion » Issue setting up agent transition graph. help please?
-
- cwhite
- 794 posts
- Offline
Without a file it's hard to tell what the issue is. Did you look at the example file provided in the docs?
Houdini Indie and Apprentice » Multiple Glue Strengths or Multiple Glue Networks
-
- cwhite
- 794 posts
- Offline
On the glue network geometry, you can create a ‘strength’ primitive attribute to control the glue strength for each constraint. In general, this can be done for any constraint parameter.
Technical Discussion » Randomize the default state of the agents
-
- cwhite
- 794 posts
- Offline
You can also just use regular SOP tools to set the “state” point attribute however you'd like.
Technical Discussion » [DOPS] packed rbd - impact point primuv
-
- cwhite
- 794 posts
- Offline
Since the Bullet solver doesn't directly use the source geometry for collision detection, the impact points might not actually be on the surface of the source geometry (for example, a convex hull collision shape could be used with geometry that is concave). Furthermore, recording UV coordinates for an impact against a packed primitive would be much more complex, since you would need to record the primitive number for every level of nesting (packed primitives can be nested arbitrarily deep).
You may want to consider using the Ray SOP to project the impact points onto the (unpacked) geometry and find the UV coordinates there.
You may want to consider using the Ray SOP to project the impact points onto the (unpacked) geometry and find the UV coordinates there.
Technical Discussion » Crowd not transitioning to proper animation
-
- cwhite
- 794 posts
- Offline
Your agents only have a walk clip (you can see this by looking at the ‘agentclipcatalog’ primitive intrinsic in the geometry spreadsheet). The Agent SOP is loading all of the clips matching ${HIP}/agents/walk/*.bclip, which is only a single clip in your case.
You probably want to set the Agent Name parameter to the same value (e.g. ‘mocapbiped2’) on each Agent ROP so that the clips are all in the same directory. Then, adjust the Agent Name parameter on the Agent SOP to load the agent definition from that directory.
You probably want to set the Agent Name parameter to the same value (e.g. ‘mocapbiped2’) on each Agent ROP so that the clips are all in the same directory. Then, adjust the Agent Name parameter on the Agent SOP to load the agent definition from that directory.
-
- Quick Links