Strange behavior of cloth simulation animation with hook #110902

New Issue

Guusebumps · 2023-08-07T19:50:08+02:00

Guusebumps commented

2023-08-07 19:50:08 +02:00

System Information
Operating system: Windows-10-10.0.19044-SP0 64 Bits
Graphics card: NVIDIA GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.25

Blender Version
Broken: version: 3.6.1, branch: blender-v3.6-release, commit date: 2023-07-21 12:59, hash: c284cfcd0830
Worked: none that I know of.

Short description of error
I made a cloth which is controlled by a hook and animated it moving forward. Then I dropped another cloth on the first cloth original position. While the first cloth is already moving forward, the downward movement of the second cloth is somehow stuck exactly at the original position of the first cloth while there is actually no object there.

Exact steps for others to reproduce the error
[Please describe the exact steps needed to reproduce the issue]
[Based on the default startup or an attached .blend file (as simple as possible)]

Create a cylinder with no caps with 32 vertices.
In edit mode select the bottom vertices and scale them up a little bit on X and Y axis so that it resembles a cone with a hole on top.
Loop cut the wall of the cylinder by 10.
Select the top vertices and make a vertex group. Name the vertex group as pin.
In object mode, create an empty with the shape of a sphere.
Move Empty upwards on Z axis until the middle of the sphere is about the position of the top of the cylinder.
Select the cylinder and add a hook modifier with Empty as the object and pin as the vertex group.
Add collision physics to the cylinder with Thickness outer at 0.001.
Add cloth physics to the cylinder. Set quality to 10. Name the cache as Cloth1. Set pin group to pin. Set collision quality to 5. Turn on object collision and set the distance to 0.001. Turn on self collision and set the distance to 0.001.
Select Empty and insert position keyframe at frame 60. Go to frame 80 and move the Empty quite distance forward on Y axis. Insert position keyframe.
Add a plane object and make it slightly larger than the cylinder.
in Edit mode subdivide the plane to 20.
in object mode move the cylinder upwards on Z axis so that it has quite a distance to the cylinder.
Add cloth physics to the plane. Set quality to 10. Name the cache as Cloth2. Set simulation start at 70. Set collision quality to 5. Turn on object collision and set the distance to 0.001. Turn on self collision and set the distance to 0.001.
Bake all the simulation (cylinder and plane).

**System Information** Operating system: Windows-10-10.0.19044-SP0 64 Bits Graphics card: NVIDIA GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.25 **Blender Version** Broken: version: 3.6.1, branch: blender-v3.6-release, commit date: 2023-07-21 12:59, hash: `c284cfcd0830` Worked: none that I know of. **Short description of error** I made a cloth which is controlled by a hook and animated it moving forward. Then I dropped another cloth on the first cloth original position. While the first cloth is already moving forward, the downward movement of the second cloth is somehow stuck exactly at the original position of the first cloth while there is actually no object there. **Exact steps for others to reproduce the error** [Please describe the exact steps needed to reproduce the issue] [Based on the default startup or an attached .blend file (as simple as possible)] 1. Create a cylinder with no caps with 32 vertices. 2. In edit mode select the bottom vertices and scale them up a little bit on X and Y axis so that it resembles a cone with a hole on top. 3. Loop cut the wall of the cylinder by 10. 4. Select the top vertices and make a vertex group. Name the vertex group as pin. 5. In object mode, create an empty with the shape of a sphere. 6. Move Empty upwards on Z axis until the middle of the sphere is about the position of the top of the cylinder. 7. Select the cylinder and add a hook modifier with Empty as the object and pin as the vertex group. 8. Add collision physics to the cylinder with Thickness outer at 0.001. 9. Add cloth physics to the cylinder. Set quality to 10. Name the cache as Cloth1. Set pin group to pin. Set collision quality to 5. Turn on object collision and set the distance to 0.001. Turn on self collision and set the distance to 0.001. 10. Select Empty and insert position keyframe at frame 60. Go to frame 80 and move the Empty quite distance forward on Y axis. Insert position keyframe. 11. Add a plane object and make it slightly larger than the cylinder. 12. in Edit mode subdivide the plane to 20. 13. in object mode move the cylinder upwards on Z axis so that it has quite a distance to the cylinder. 14. Add cloth physics to the plane. Set quality to 10. Name the cache as Cloth2. Set simulation start at 70. Set collision quality to 5. Turn on object collision and set the distance to 0.001. Turn on self collision and set the distance to 0.001. 15. Bake all the simulation (cylinder and plane).

Cloth test.blend

985 KiB

Guusebumps added the

labels 2023-08-07 19:50:09 +02:00

Tomás Cejas commented

2023-08-08 00:23:10 +02:00

This might me working as intended given the order of the modifiers, or only partially, since the deformation of the falling rectangular cloth seems weird: The falling cloth collides with the "skirt" as if it occupied at the same time all spaces it went through. I'm not sure if that's how it should work given that modifier order (Hook>Collision>Cloth)

For the simulation to do what you want, you have to place the collision modifier after (below) the cloth modifier, and the issue is solved. (Hook>Cloth>Collision).

This might me working as intended given the order of the modifiers, or only partially, since the deformation of the falling rectangular cloth seems weird: The falling cloth collides with the "skirt" as if it occupied at the same time all spaces it went through. I'm not sure if that's how it should work given that modifier order (Hook>Collision>Cloth) ![image](/attachments/856567ac-b736-48e3-986e-0a7c98f56e38) For the simulation to do what you want, you have to place the collision modifier after (below) the cloth modifier, and the issue is solved. (Hook>Cloth>Collision).

image.png

62 KiB

Guusebumps commented

2023-08-08 04:32:26 +02:00

For the simulation to do what you want, you have to place the collision modifier after (below) the cloth modifier, and the issue is solved. (Hook>Cloth>Collision).

Thank you, yes it works. I am so used to Hook>Collision>Cloth because in my previous cases it works. I didn't realize this is the very thing that caused the problem.

Can you give me some explanation on why it doesn't work?

> For the simulation to do what you want, you have to place the collision modifier after (below) the cloth modifier, and the issue is solved. (Hook>Cloth>Collision). Thank you, yes it works. I am so used to Hook>Collision>Cloth because in my previous cases it works. I didn't realize this is the very thing that caused the problem. Can you give me some explanation on why it doesn't work?

Tomás Cejas commented

2023-08-08 06:07:24 +02:00

My general idea is that Hook>Cloth>Collision first calculates the updated positions of the vertices according to the cloth simulation, and then uses those updated positions to calculate the collisions. Notice the cloth does include the hook in its calculations because the hook modifier is above it.

Hook>Collision>Cloth, on the other hand, first calculates collisions and then moves the mesh. This doesn't paint the whole picture but that's all I know.

While I am not sure why the deformation looks how it does in the 2nd case, I think this issue should now be closed/marked as solved.

My general idea is that Hook>Cloth>Collision first calculates the updated positions of the vertices according to the cloth simulation, and then uses those updated positions to calculate the collisions. Notice the cloth does include the hook in its calculations because the hook modifier is above it. Hook>Collision>Cloth, on the other hand, first calculates collisions and then moves the mesh. This doesn't paint the whole picture but that's all I know. While I am not sure why the deformation looks how it does in the 2nd case, I think this issue should now be closed/marked as solved.

Pratik Borhade commented

2023-08-08 08:59:45 +02:00

Will close if error was due to order of modifiers. Thanks @TomasCJ :)
Don't hesitate to reopen the report if problem reappears.

Will close if error was due to order of modifiers. Thanks @TomasCJ :) Don't hesitate to reopen the report if problem reappears.

Pratik Borhade closed this issue

2023-08-08 08:59:47 +02:00

Blender Bot added

and removed