Maniphest T53628

Particle children loosing their randomness
Closed, Archived
Actions

Assigned To
Germano Cavalcante (mano-wii)
Authored By
Acrivec Cevirca (Acrivec)
Dec 25 2017, 6:25 AM
Tags
  • BF Blender
  • Tracker Curfew
Subscribers
Acrivec Cevirca (Acrivec)
Germano Cavalcante (mano-wii)
Philipp Oeser (lichtwerk)

Description

System Information
Windows 10, 980Ti, 6700k

Blender Version
Broken: 2.79, release version

Short description of error
Particle children are suddenly loosing their randomness. Pictures will say much more:
Left and right is the same project, but in different stadiums.
Left is old save, right is current.
Please note, all particle settings are exacly the same.

  1. https://i.imgur.com/YtBsIa6.png

As you can see left one is cool, random. Right one has a pattern, no randomness. They are just slightly moved.

  1. https://i.imgur.com/u1iPsjZ.png

Changing segments on the left project from 5 to 4: still looks OK.

  1. https://i.imgur.com/ZWtdYGf.png

Changed segments back to 5 from 4 (original value). Suddenly: no randomness!
I can change those segments again and again and the randomness will not come back.
However (!) undo option does fix that!
But after two changes it will get broken again.

Exact steps for others to reproduce the error
I can't attach my .blend file due to copyrights, unfortunately.
I also don't know how many steps it takes to screw it up. But something is bad. I can't leave render overnight due to this problem. :(

Side note: changing the seed does not help, they are moved in different positions but the pattern is still clearly visible.

Also, this gets somehow fixed.
Look, https://i.imgur.com/z5Lxsaa.png <- bugged (I haven't noticed it then), the next day: https://i.imgur.com/mhn6NOC.png - fixed! Without touching.

Related Objects

Event Timeline

Acrivec Cevirca (Acrivec) created this task.Dec 25 2017, 6:25 AM
Acrivec Cevirca (Acrivec) updated the task description.Dec 25 2017, 5:07 PM
Philipp Oeser (lichtwerk) lowered the priority of this task from 90 to 30.Jan 3 2018, 8:31 PM
Philipp Oeser (lichtwerk) added a subscriber: Philipp Oeser (lichtwerk).

I have a hard time reproducing this unfortunately [though it is apparent theres something wrong]

couple of questions:

  • these are 'simple' (not 'interpolated') children, right?
  • what randomizations are we talking about? just the orientations? or rough?, kink?
  • no hair dynamics or particle edit mode, right?

It would be very helpful if you could try recreating the problem in a simple .blend that you can share, otherwise we might be stuck here.
Marking as inclomplete for now...

Acrivec Cevirca (Acrivec) added a comment.Jan 3 2018, 9:38 PM

Yes, simple children.
https://i.imgur.com/LX4Zf9I.png
No hair dynamics, about the second one I'm not sure what are you talking about.

I think it should be better if I'll send you this .blend file by private message, so it won't get public?

Philipp Oeser (lichtwerk) added a comment.Jan 4 2018, 8:16 AM

Yes, can you mail me [info at graphics-engineer dot com] on how to get the file?

Acrivec Cevirca (Acrivec) added a comment.Jan 5 2018, 12:04 AM

I've sent you the .blend.

Philipp Oeser (lichtwerk) added a comment.Jan 12 2018, 8:13 PM

Sorry to get back so late, I've never recieved the .blend. Could you upload somewhere with a password or mail me how we can solve otherwise? [I'm afraid my mailbox cannot accept large files...]

Acrivec Cevirca (Acrivec) added a comment.Jan 12 2018, 8:17 PM

It was ~11MB, so not this big.
I've sent second mail, with link , without the rest of the mail.

Philipp Oeser (lichtwerk) claimed this task.Jan 12 2018, 8:23 PM

@Acrivec Cevirca (Acrivec) : sorry for the noise, found the blend in the mail, will have a look shortly...

Philipp Oeser (lichtwerk) raised the priority of this task from 30 to 50.Jan 17 2018, 7:54 PM

OK, looking further there seems to be a problem with the dupli_group Use Count option.
So (if I observe correctly) what looks like children 'loosing their randomness' is actually the dupli group always picking the same element in the group (instead of a different element for each child)

So (as a workaround) I can successfully change segments in case I also enable Pick Random.
I can also (as a workaround) leave Pick Random disabled but change the count of an element.
Both these workarouds will restore the randomness.

That said, this really seems to be a bug, will look at actual code shortly.

Philipp Oeser (lichtwerk) added a comment.Jan 17 2018, 8:59 PM

on first sight, this is NOT the same element displaying again and again, needs further investigation...

Acrivec Cevirca (Acrivec) added a comment.Jan 17 2018, 9:45 PM

I'm also experiencing problems with children's size (quite extreme, for example hair length is 1cm and children are 50cm), and random rotation - but I guess that would be a separate bug report.

Philipp Oeser (lichtwerk) removed Philipp Oeser (lichtwerk) as the assignee of this task.Feb 4 2018, 10:23 PM

Unfortunately I'll only be able to look at this in April again, so let others take over....

Philipp Oeser (lichtwerk) mentioned this in T54257: group objects and their counts become mismatched in particle system "use count" list.Apr 10 2018, 1:42 PM
Dalai Felinto (dfelinto) added a project: Tracker Curfew.Dec 23 2019, 4:36 PM
Germano Cavalcante (mano-wii) closed this task as Archived.Jan 21 2020, 7:51 PM
Germano Cavalcante (mano-wii) claimed this task.
Germano Cavalcante (mano-wii) added a subscriber: Germano Cavalcante (mano-wii).

I created this file to try to reproduce this problem, but without success:

bug_hair_segments.blend552 KBDownload

Unfortunately the scenario described is too time consuming for us to track down, we require the bug reporter to narrow down the problem.

Normally .blend files can be simplified by removing most objects and disabling settings, until the problem reveals itself more clearly.

If there are many steps to reproduce the problem, please attach a .blend file (simplified) saved right before the bug occurs, so only a few steps are needed to trigger the bug.