Page MenuHome

Child hair particles set to "Interpolated" causes jittering when groom on alembic cache
Closed, DuplicatePublic

Description

Blender 2.8x
Win10

Hi,
THE PROBLEM is that if you turn on "CHILD PARTICLES" and SET them TO "INTERPOLATED" MODE it causes JITTERING of them
acros surface of the geo from which roots.

When selecting "Simple" mode everything works fine - No jitter at all, perfectly sticked to the geo.

You can test it simply by switching from mode to mode to see this problem. Its already setuped for rendering using eevee but same problem using
Cycles renderer.

Note that I ve briefly tried to change export settings for Alembic during export of anim but nothing helped...still causes this problem.

Also note that when outputting alembic cache for animated object using transforms not armatures, it works ok. But when exporting alembic from armature animated
object this causes this problem.

please see this Zip archive with readme info and example scene....

Thank you,

Libor

Details

Type
Bug

Event Timeline

here is some rendered animation with close up of the hair particles jittering on my animated model (alembic cache)

pls when viewing change quality below video to highest resolution (1080p) or just download and playback at your desktop

Thx again.
L.

Philipp Oeser (lichtwerk) triaged this task as Confirmed, Medium priority.

Confirmed on first sight
(Looks like 2.79 has a similiar issue even)

@Sybren A. Stüvel (sybren): mind having a look?

@Libor Batek (lbatek) please put the description of the issue in this task, instead of hiding it in a ZIP file. Hiding it like this makes it invisible to the search engine, and makes it harder for developers to access as well.

@Sybren A. Stüvel (sybren) ok, sorry not doing it before....should be done right now

Have the same problem, here is a side by side basic test

Libor Batek (lbatek) added a comment.EditedApr 26 2019, 9:29 AM

I ve tested to render animation with mblur and hair on alembic cache and there is some problems too....not sure if it corelates to the above mentioned problems or its kind of new problem.

EXAMPLE_MOV (Apple ProRes)

still frame from animation with mblur...seems like hair particles doesnt refer to preceding frame position of the hair particles (not sure If I shouldnt use bake cache for particles??)

they spread from world origin

@Sybren A. Stüvel (sybren): any chance to have a look at this? [still wont consider this High prio -- but I know at least two other studios suffering from this as well, and having to go .mdd/.pc2 instead...]
(how did this work for Spring btw.?)

CC @Brecht Van Lommel (brecht)

Sybren A. Stüvel (sybren) lowered the priority of this task from Confirmed, Medium to Needs Information from User.Fri, Aug 9, 11:08 AM

@Libor Batek (lbatek) please provide us with a valid example file. The ABC file from the attached zip seems to be corrupt:

% abctree testMesh_anim.abc -a
ABC
 :--.childBnds
 `--Armature
     | :--.xform
     | :    :--.inherits
     | :    :--.ops
     | :    `--.vals
     :--visible
     `--TestMesh
         | :--.xform
         | :    :--.inherits
         | :    :--.ops
         | :    `--.vals
         :--visible
         `--TestMeshShape
             | `--.geom
             |    :--.selfBnds
             |    :--P
             |    :--.faceIndices
             |    :--.faceCounts
             |    | :--.userProperties
             |    | :    `--meshtype
terminate called after throwing an instance of 'Alembic::Util::v10::Exception'
  what():  ICompoundProperty::getPropertyHeader()
ERROR: EXCEPTION:
Out of range index in CprData::getPropertyHeader: 18446744073709551615
zsh: abort (core dumped)  abctree testMesh_anim.abc -a

I know at least two other studios suffering from this as well, and having to go .mdd/.pc2 instead...

It would be helpful if they could provide:

  • An as-simple-as-possible ABC file that shows this issue,
  • An MDD or PC2 file of the same source that does not show this issue,
  • The blend file used to generate those files, and
  • A blend file that imports & shows the problems.

(how did this work for Spring btw.?)

AFAIK we didn't put hair on top of Alembic cashes.

@Sybren A. Stüvel (sybren): I have updated T56408 with a file that has both abc and pc2 caches (pc2 renders fine, abc not) including the output sequences.

I assume these reports are really about the same issue (and could be merged?)

Philipp Oeser (lichtwerk) raised the priority of this task from Needs Information from User to Waiting for Developer to Reproduce.Fri, Aug 9, 12:46 PM