Hair Dynamics: not working with non-uniform scaling #69265

New Issue

Imre Varga · 2019-08-28T21:37:11+02:00

Imre Varga commented

2019-08-28 21:37:11 +02:00

System Information
Operating system: Windows 10
Graphics card: RTX 2070 QMax

Blender Version
Broken: v 2.80
Worked: (optional)

Basic hair dynamics act in a weird, erratic way.

190828_blender_280_hair_dynamics_basics.flv

Exact steps for others to reproduce the error

open attached file
press play animation
observe erratic behaviour of hair

190828_blender_280_hair_dynamics_basics.blend

p.s.

I know that this is not the right place for whining and I appreciate that a project like Blender relies on user contribution.

I started my project in Blender 2.79. Overall I was impressed with the stability and reliability of the application. I knew about the limitations e.g. that cloth simulation was pretty useless, but at least the application acted in a dependable and predictable way.

Ever since I moved my project to Blender 2.8, I haven't done much else than writing bug reports, reproducing errors and finding workarounds. I can't even think of making progress with my project.

My system is brand new, the components like the i7 9750 and the RTX 2070 are bog standard. It works perfectly reliably and no other applications show problems.

The problem I am describing here is just the tip of the iceberg. In my original project I hit all these weird issues all the time.

Mind you, what I am saying just refers to physics simulation. Everything else works fine.

Could there be a fundamental problem with physics and my system?

I am beginning to lose faith. Please help!

System Information Operating system: Windows 10 Graphics card: RTX 2070 QMax Blender Version Broken: v 2.80 Worked: (optional) Basic hair dynamics act in a weird, erratic way. [190828_blender_280_hair_dynamics_basics.flv](https://archive.blender.org/developer/F7708098/190828_blender_280_hair_dynamics_basics.flv) Exact steps for others to reproduce the error - open attached file - press play animation - observe erratic behaviour of hair [190828_blender_280_hair_dynamics_basics.blend](https://archive.blender.org/developer/F7708104/190828_blender_280_hair_dynamics_basics.blend) p.s. I know that this is not the right place for whining and I appreciate that a project like Blender relies on user contribution. I started my project in Blender 2.79. Overall I was impressed with the stability and reliability of the application. I knew about the limitations e.g. that cloth simulation was pretty useless, but at least the application acted in a dependable and predictable way. Ever since I moved my project to Blender 2.8, I haven't done much else than writing bug reports, reproducing errors and finding workarounds. I can't even think of making progress with my project. My system is brand new, the components like the i7 9750 and the RTX 2070 are bog standard. It works perfectly reliably and no other applications show problems. The problem I am describing here is just the tip of the iceberg. In my original project I hit all these weird issues all the time. Mind you, what I am saying just refers to physics simulation. Everything else works fine. Could there be a fundamental problem with physics and my system? I am beginning to lose faith. Please help!

Imre Varga commented

2019-08-28 21:37:11 +02:00

Added subscriber: @imrevarga

Philipp Oeser commented

2019-08-28 22:38:09 +02:00

Added subscriber: @lichtwerk

Philipp Oeser commented

2019-08-28 22:38:09 +02:00

In this particular case, applying scale on the objects seems to fix the mess.
Still needs further investigation, cannot reproduce on non-uniformly-scaled objects in simple cases...

Guess we have to confirm this for now.
(Appending this back into 2.79 shows this pretty messed up as well though...)
Btw. agree that 2.80's weak spot are Physics [there are some dark spots left here -- we'll get there though :)]

@imrevarga : it would be really helpful to check at which point the dynamics actually start breaking [is it the scaling? is it the armature? is it the vertex groups?]
(finding the steps to reproduce from a working state to the broken state is key to getting this fixed sooner than later... would help tremendously if you could do some checking as well...)

In this particular case, applying scale on the objects seems to fix the mess. Still needs further investigation, cannot reproduce on non-uniformly-scaled objects in simple cases... Guess we have to confirm this for now. (Appending this back into 2.79 shows this pretty messed up as well though...) Btw. agree that 2.80's weak spot are Physics [there are some dark spots left here -- we'll get there though :)] @imrevarga : it would be really helpful to check at which point the dynamics actually start breaking [is it the scaling? is it the armature? is it the vertex groups?] (finding the steps to reproduce from a working state to the broken state is key to getting this fixed sooner than later... would help tremendously if you could do some checking as well...)

Imre Varga commented

2019-08-28 23:21:53 +02:00

Thanks for the quick response.

I applied scale to the two leftmost objects, and now it works. Funnily, if I then scale the objects without applying, it still works.

I made up the two objects on the left by inserting two spheres, scaling them independently and then joining them into one object. I made the second object by copyind the first one and deleting the armature. The hair acts up on that object as well, but in a different way.

The original purpose of this setup was to reproduce another problem related to physics and rendering.

So I can go on now to do just that and prepare my next bug report ...

Thanks for the quick response. I applied scale to the two leftmost objects, and now it works. Funnily, if I then scale the objects without applying, it still works. I made up the two objects on the left by inserting two spheres, scaling them independently and then joining them into one object. I made the second object by copyind the first one and deleting the armature. The hair acts up on that object as well, but in a different way. The original purpose of this setup was to reproduce another problem related to physics and rendering. So I can go on now to do just that and prepare my next bug report ...

Imre Varga commented

2019-08-29 07:23:23 +02:00

Pinpointed it further: Whenever you scale u n e v e n l y, the hair plays up. If you then apply scale, it is ok.

Philipp Oeser commented

2019-09-04 11:21:05 +02:00

Here is another one that fails:
#69265.blend
Marking as regression from 2.79

Here is another one that fails: [#69265.blend](https://archive.blender.org/developer/F7716166/T69265.blend) Marking as regression from 2.79

Philipp Oeser changed title from ~~basic hair dynamics not working~~ to Hair Dynamics: not working with non-uniform scaling

2019-09-04 11:22:01 +02:00

Garry R. Osgood commented

2019-09-04 13:36:11 +02:00

Added subscriber: @grosgood

Garry R. Osgood commented

2019-09-04 13:36:11 +02:00

@lichtwerk:
I'm a bit slow on the uptake, but I'm trying to figure out what it is I'm supposed to be seeing.
There is a sudden, abrupt transform that happens to the hair itself. Almost looks like a shear. Shot mpv-shot0001.jpg is the frame immediately before the transform, and mpv-shot0002.jpg is immediately after. Other than that, the hair dynamics looks pretty much as intended. See my vid_T69265.mp4. That abrupt transform - that's the manifestation of the issue, right?
vid_T69265.mp4

@lichtwerk: I'm a bit slow on the uptake, but I'm trying to figure out what it is I'm supposed to be seeing. There is a sudden, abrupt transform that happens to the hair itself. Almost looks like a shear. Shot mpv-shot0001.jpg is the frame immediately before the transform, and mpv-shot0002.jpg is immediately after. Other than that, the hair dynamics looks pretty much as intended. See my vid_T69265.mp4. That abrupt transform - that's the manifestation of the issue, right? [vid_T69265.mp4](https://archive.blender.org/developer/F7716273/vid_T69265.mp4) ![mpv-shot0001.jpg](https://archive.blender.org/developer/F7716274/mpv-shot0001.jpg) ![mpv-shot0002.jpg](https://archive.blender.org/developer/F7716276/mpv-shot0002.jpg)

Philipp Oeser commented

2019-09-04 15:19:30 +02:00

Cant see the video for some reason...

Garry R. Osgood commented

2019-09-04 17:36:41 +02:00

Yes. Sorry about that - the video was a fast load before I was off to my office and I probably screwed it up somehow. It doesn't play here either. Can fix it after I get back to my studio tonight (four hours from now, five allowing for domestic fiddle-faddle...)

Video is not the heart of the matter, but the two stills, one frame apart, about the moment Blender started running the animation timeline. I observed a kind of whip-lash transformation of the hair particles that they make themselves, sort of a (but not quite) a shear in the Y direction that the geometry itself doesn't make. I surmise that such a transformation happened to the geometry itself at some point - two or more separate and not the same transforms performed on the original default cube. Scaling is unity in all directions on the object itself and in the deltas, so I presume you cleaned these up by various applies, but the particle system exhibits a weird memory of them. I've been seeing that kind of latency for a long time now (maybe back to 2.78 or so) that is hard to track down.

And yes, I've heard the yada-yada about the particle system being replaced Real Soon Now. Recall chatter like that back in 2014. But the reality is that THIS particle/hair system is what is out in the world now and the new stuff won't be out for 2.82 probably ;); people are trying to use THIS particle/hair system now, and when it becomes too painful to thread that needle Blender takes a reputational hit. Kind of a death by a thousand cuts. Enough of this. I'll take the rest of this rant to devtalk sometime or other. Thank you all for the great work (and that is sincere...).

Yes. Sorry about that - the video was a fast load before I was off to my office and I probably screwed it up somehow. It doesn't play here either. Can fix it after I get back to my studio tonight (four hours from now, five allowing for domestic fiddle-faddle...) Video is not the heart of the matter, but the two stills, one frame apart, about the moment Blender started running the animation timeline. I observed a kind of whip-lash transformation of the hair particles that they make themselves, sort of a (but not quite) a shear in the Y direction that the geometry itself doesn't make. I surmise that such a transformation happened to the geometry itself at some point - two or more separate and not the same transforms performed on the original default cube. Scaling is unity in all directions on the object itself and in the deltas, so I presume you cleaned these up by various applies, but the particle system exhibits a weird memory of them. I've been seeing that kind of latency for a long time now (maybe back to 2.78 or so) that is hard to track down. And yes, I've heard the yada-yada about the particle system being replaced Real Soon Now. Recall chatter like that back in 2014. But the reality is that THIS particle/hair system is what is out in the world now and the new stuff won't be out for 2.82 probably ;); people are trying to use THIS particle/hair system now, and when it becomes too painful to thread that needle Blender takes a reputational hit. Kind of a death by a thousand cuts. Enough of this. I'll take the rest of this rant to devtalk sometime or other. Thank you all for the great work (and that is sincere...).

Philipp Oeser commented

2019-09-05 15:04:56 +02:00

I hear you :)
Will continue looking at it (as time permits), I also think we should at least iron out the worst bugs 2.8 has compared to 2.79...
[ squashed one yesterday :) e52ad1835a ]

I hear you :) Will continue looking at it (as time permits), I also think we should at least iron out the worst bugs 2.8 has compared to 2.79... [ squashed one yesterday :) e52ad1835a ]

Sybren A. Stüvel commented

2020-01-27 15:09:25 +01:00

Added subscriber: @dr.sybren

Sybren A. Stüvel commented

2020-01-27 15:09:25 +01:00

Changed status from 'Confirmed' to: 'Archived'

Sybren A. Stüvel closed this issue

2020-01-27 15:09:25 +01:00

Sybren A. Stüvel self-assigned this 2020-01-27 15:09:25 +01:00

Sybren A. Stüvel commented

2020-01-27 15:09:25 +01:00

Scaling in physics simulations is always iffy, and nonuniform scaling even more so. This is a known limitation, and also in 2.79 this gave issues:

Here you see two icospheres, the left one with scale (1, 1, 2), the right one with the scale applied.

Scaling in physics simulations is always iffy, and nonuniform scaling even more so. This is a known limitation, and also in 2.79 this gave issues: ![image.png](https://archive.blender.org/developer/F8304994/image.png) Here you see two icospheres, the left one with scale (1, 1, 2), the right one with the scale applied.

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Hair Dynamics: not working with non-uniform scaling #69265