Particle system baking messes particle dietime (that in turn messes up particle textures) #107447

New Issue

Open

opened 2023-04-28 18:23:05 +02:00 by E-13 · 9 comments

E-13 commented

2023-04-28 18:23:05 +02:00

System Information
Operating system: Windows-10-10.0.19044-SP0 64 Bits
Graphics card: NVIDIA GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15

Blender Version
Broken: version: 3.5.0, branch: blender-v3.5-release, commit date: 2023-03-29 02:56, hash: 1be25cfff18b
Worked: (newest version of Blender that worked as expected)

Short description of error
If texture with color ramp is used for animating particle size through lifetime, in ~10 last frames of simulation all particles PERMANENTLY getting smaller until totally disappear lol, and it occurs only if using BAKE.
this is not animation, this is 100% bug, looks like particle size animation is applied to WHOLE particle system respectively/relatively to all scene timeline, and is overriding every single particle age at the end of timeline.

Exact steps for others to reproduce the error
File uploaded.
Particle systems in blender are SOOOO broken, the way how it works is so facepalmable, particularly "bake" and "textures".
looks like in my case they met lol

**System Information** Operating system: Windows-10-10.0.19044-SP0 64 Bits Graphics card: NVIDIA GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15 **Blender Version** Broken: version: 3.5.0, branch: blender-v3.5-release, commit date: 2023-03-29 02:56, hash: `1be25cfff18b` Worked: (newest version of Blender that worked as expected) **Short description of error** If texture with color ramp is used for animating particle size through lifetime, in ~10 last frames of simulation all particles PERMANENTLY getting smaller until totally disappear lol, and it occurs only if using BAKE. this is not animation, this is 100% bug, looks like particle size animation is applied to WHOLE particle system respectively/relatively to all scene timeline, and is overriding every single particle age at the end of timeline. **Exact steps for others to reproduce the error** File uploaded. Particle systems in blender are SOOOO broken, the way how it works is so facepalmable, particularly "bake" and "textures". looks like in my case they met lol

bugreport.blend

17 MiB

E-13 added the

labels 2023-04-28 18:23:06 +02:00

Iliya Katushenock added the

Nodes & Physics

label 2023-04-28 18:28:28 +02:00

Richard Antalik commented

2023-04-29 06:02:43 +02:00

Member

What are the steps to make incorrect bake? I have re-baked particles and could not reproduce the issue.

What are the steps to make incorrect bake? I have re-baked particles and could not reproduce the issue.

Richard Antalik added

Needs Information from User

and removed

labels 2023-04-29 06:02:55 +02:00

E-13 commented

2023-05-02 13:27:06 +02:00

Author

What are the steps to make incorrect bake? I have re-baked particles and could not reproduce the issue.

select particle system in scene ("particles.120" e.g.), go to "textures" and switch ON "SIZE" influence

now in the end of scene it looks like this —

> What are the steps to make incorrect bake? I have re-baked particles and could not reproduce the issue. select particle system in scene ("particles.120" e.g.), go to "textures" and switch ON "SIZE" influence now in the end of scene it looks like this —

22 KiB

168 KiB

E-13 commented

2023-05-04 17:47:55 +02:00

Author

blender particle system is SO FUCKING BROKEN
this is one particle system, no constrains, baked
left is viewport, right is render window
wtf is this lmao? facepalm

blender particle system is SO FUCKING BROKEN this is one particle system, no constrains, baked left is viewport, right is render window wtf is this lmao? facepalm

219 KiB

Iliya Katushenock commented

2023-05-04 18:30:23 +02:00

Member

@E-13 Please be calm and give an example file to reproduce issues

@E-13 Please be calm and give an example file to reproduce issues

Richard Antalik commented

2023-05-09 19:23:10 +02:00

Member

What are the steps to make incorrect bake? I have re-baked particles and could not reproduce the issue.

select particle system in scene ("particles.120" e.g.), go to "textures" and switch ON "SIZE" influence

now in the end of scene it looks like this —

In provided file texture had already size influence enabled. So I can't reproduce that. Are you sure this is what is causing baking to be incorrect?

> > What are the steps to make incorrect bake? I have re-baked particles and could not reproduce the issue. > > select particle system in scene ("particles.120" e.g.), go to "textures" and switch ON "SIZE" influence > > now in the end of scene it looks like this — In provided file texture had already size influence enabled. So I can't reproduce that. Are you sure this is what is causing baking to be incorrect?

E-13 commented

2023-05-15 14:21:50 +02:00

Author

In provided file texture had already size influence enabled. So I can't reproduce that. Are you sure this is what is causing baking to be incorrect?

switch size influence OFF
delete particle physics bake
create bake again
delete bake
switch size influence ON
create bake
(repeat all this steps several times maybe, just to create NEW bake)

now it should be broken

i am pretty sure that bake+textures influence produces broken bake, i tryed enough times all of this

> In provided file texture had already size influence enabled. So I can't reproduce that. Are you sure this is what is causing baking to be incorrect? switch size influence OFF delete particle physics bake create bake again delete bake switch size influence ON create bake (repeat all this steps several times maybe, just to create NEW bake) now it should be broken i am pretty sure that bake+textures influence produces broken bake, i tryed enough times all of this

Richard Antalik commented

2023-05-23 06:08:07 +02:00

Member

Still can't reproduce unfortunately.

Still can't reproduce unfortunately.

Richard Antalik added

and removed

Needs Information from User

labels 2023-05-23 06:08:18 +02:00

Philipp Oeser commented

2023-05-26 16:09:26 +02:00

Member

I can confirm the thing and looked a bit at it abit (but wont continue doing so because after all the particle system is EOL, glad we have geometry nodes with simulation to move forward... -- and yes @E-13 , old particles can be broken at various places)

So just for reference here are some findings:

a particle has a dietime (the frame on which it will die)
determining the coordinate on the particle texture relies on the particles birthtime and dietime and current frame
now to the part where my findings get fuzzy (but will still continue)
-- when we bake, particles that are outside their lifetime are skipped (see ptcache_particle_write)
-- reading back the bake (see ptcache_particle_read) stored times are not read for some reason (not sure why that is, but
-- resetting particles (see reset_particle), it happens for many particles (at least in this example), that psys_get_dietime_from_cache will set the dietime to our scenes endframe -- but really, that particles dietime should be later... that will then result in the texture coordinate being much more to the right, reducing the size where it shouldnt.

If anyone wants to check further, here is what I have been looking at:

diff --git a/source/blender/blenkernel/intern/particle.cc b/source/blender/blenkernel/intern/particle.cc
index 4fddd8241a6..510c10712c2 100644
--- a/source/blender/blenkernel/intern/particle.cc
+++ b/source/blender/blenkernel/intern/particle.cc
@@ -4478,6 +4478,7 @@ void psys_get_texture(
         case TEXCO_PARTICLE:
           /* texture coordinates in range [-1, 1] */
           texvec[0] = 2.0f * (cfra - pa->time) / (pa->dietime - pa->time) - 1.0f;
+          printf("cfra %f pa->time %f pa->dietime %f texvec[0] %f\n", cfra, pa->time, pa->dietime, texvec[0]);
           if (sim->psys->totpart > 0) {
             texvec[1] = 2.0f * float(pa - sim->psys->particles) / float(sim->psys->totpart) - 1.0f;
           }
diff --git a/source/blender/blenkernel/intern/particle_system.c b/source/blender/blenkernel/intern/particle_system.c
index 7060983c281..e4d134ef20e 100644
--- a/source/blender/blenkernel/intern/particle_system.c
+++ b/source/blender/blenkernel/intern/particle_system.c
@@ -1114,6 +1114,7 @@ void reset_particle(ParticleSimulationData *sim, ParticleData *pa, float dtime,
       (sim->psys->pointcache->mem_cache.first))
   {
     float dietime = psys_get_dietime_from_cache(sim->psys->pointcache, p);
+    printf("reset_particle : got psys_get_dietime_from_cache %f (would have been %f instead)\n", dietime, pa->dietime);
     pa->dietime = MIN2(pa->dietime, dietime);
   }
 
diff --git a/source/blender/blenkernel/intern/pointcache.c b/source/blender/blenkernel/intern/pointcache.c
index de3c7256bdb..890adbb36ae 100644
--- a/source/blender/blenkernel/intern/pointcache.c
+++ b/source/blender/blenkernel/intern/pointcache.c
@@ -286,6 +286,7 @@ static int ptcache_particle_write(int index, void *psys_v, void **data, int cfra
       const int pa_sfra = (int)pa->time - step;
       const int pa_efra = ((int)pa->dietime - 1) + step;
       if (!(cfra >= pa_sfra && cfra <= pa_efra)) {
+        printf("ptcache_particle_write skip %f\n", pa->dietime);
         return 0;
       }
     }
@@ -295,6 +296,8 @@ static int ptcache_particle_write(int index, void *psys_v, void **data, int cfra
   times[1] = pa->dietime;
   times[2] = pa->lifetime;
 
+  printf("ptcache_particle_write write %f\n", pa->dietime);
+
   PTCACHE_DATA_FROM(data, BPHYS_DATA_INDEX, &index);
   PTCACHE_DATA_FROM(data, BPHYS_DATA_LOCATION, pa->state.co);
   PTCACHE_DATA_FROM(data, BPHYS_DATA_VELOCITY, pa->state.vel);
@@ -356,6 +359,7 @@ static void ptcache_particle_read(
     pa->time = times[0];
     pa->dietime = times[1];
     pa->lifetime = times[2];
+    printf("ptcache_particle_read: pa->dietime %f\n", pa->dietime); // this never happens
   }
 
   if (boid) {

I can confirm the thing and looked a bit at it abit (but wont continue doing so because after all the particle system is EOL, glad we have geometry nodes with simulation to move forward... -- and yes @E-13 , old particles can be broken at various places) So just for reference here are some findings: - a particle has a dietime (the frame on which it will die) - determining the coordinate on the particle texture relies on the particles birthtime and dietime and current frame - now to the part where my findings get fuzzy (but will still continue) -- when we bake, particles that are outside their lifetime are skipped (see `ptcache_particle_write`) -- reading back the bake (see ptcache_particle_read) stored times are not read for some reason (not sure why that is, but -- resetting particles (see `reset_particle`), it happens for many particles (at least in this example), that `psys_get_dietime_from_cache` will set the dietime to our scenes endframe -- but really, that particles dietime should be later... that will then result in the texture coordinate being much more to the right, reducing the size where it shouldnt. If anyone wants to check further, here is what I have been looking at: ```Diff diff --git a/source/blender/blenkernel/intern/particle.cc b/source/blender/blenkernel/intern/particle.cc index 4fddd8241a6..510c10712c2 100644 --- a/source/blender/blenkernel/intern/particle.cc +++ b/source/blender/blenkernel/intern/particle.cc @@ -4478,6 +4478,7 @@ void psys_get_texture( case TEXCO_PARTICLE: /* texture coordinates in range [-1, 1] */ texvec[0] = 2.0f * (cfra - pa->time) / (pa->dietime - pa->time) - 1.0f; + printf("cfra %f pa->time %f pa->dietime %f texvec[0] %f\n", cfra, pa->time, pa->dietime, texvec[0]); if (sim->psys->totpart > 0) { texvec[1] = 2.0f * float(pa - sim->psys->particles) / float(sim->psys->totpart) - 1.0f; } diff --git a/source/blender/blenkernel/intern/particle_system.c b/source/blender/blenkernel/intern/particle_system.c index 7060983c281..e4d134ef20e 100644 --- a/source/blender/blenkernel/intern/particle_system.c +++ b/source/blender/blenkernel/intern/particle_system.c @@ -1114,6 +1114,7 @@ void reset_particle(ParticleSimulationData *sim, ParticleData *pa, float dtime, (sim->psys->pointcache->mem_cache.first)) { float dietime = psys_get_dietime_from_cache(sim->psys->pointcache, p); + printf("reset_particle : got psys_get_dietime_from_cache %f (would have been %f instead)\n", dietime, pa->dietime); pa->dietime = MIN2(pa->dietime, dietime); } diff --git a/source/blender/blenkernel/intern/pointcache.c b/source/blender/blenkernel/intern/pointcache.c index de3c7256bdb..890adbb36ae 100644 --- a/source/blender/blenkernel/intern/pointcache.c +++ b/source/blender/blenkernel/intern/pointcache.c @@ -286,6 +286,7 @@ static int ptcache_particle_write(int index, void *psys_v, void **data, int cfra const int pa_sfra = (int)pa->time - step; const int pa_efra = ((int)pa->dietime - 1) + step; if (!(cfra >= pa_sfra && cfra <= pa_efra)) { + printf("ptcache_particle_write skip %f\n", pa->dietime); return 0; } } @@ -295,6 +296,8 @@ static int ptcache_particle_write(int index, void *psys_v, void **data, int cfra times[1] = pa->dietime; times[2] = pa->lifetime; + printf("ptcache_particle_write write %f\n", pa->dietime); + PTCACHE_DATA_FROM(data, BPHYS_DATA_INDEX, &index); PTCACHE_DATA_FROM(data, BPHYS_DATA_LOCATION, pa->state.co); PTCACHE_DATA_FROM(data, BPHYS_DATA_VELOCITY, pa->state.vel); @@ -356,6 +359,7 @@ static void ptcache_particle_read( pa->time = times[0]; pa->dietime = times[1]; pa->lifetime = times[2]; + printf("ptcache_particle_read: pa->dietime %f\n", pa->dietime); // this never happens } if (boid) { ```

Philipp Oeser added

Nodes & Physics

and removed

Nodes & Physics

labels 2023-05-26 16:10:08 +02:00

Philipp Oeser commented

2023-05-26 16:12:24 +02:00

Member

Marking as Known Issue, since the old particle system will likely not see fixes from core developers (community patches are still welcome of course).

Reminder to check out geometry nodes, what this file wants to achieve should already be possible there with much more flexibility

Marking as Known Issue, since the old particle system will likely not see fixes from core developers (community patches are still welcome of course). Reminder to check out geometry nodes, what this file wants to achieve should already be possible there with much more flexibility

Philipp Oeser changed title from ~~particle system texture+bake issue~~ to Particle system baking messes particle dietime (that in turn messes up particle textures)

2023-05-26 17:17:04 +02:00

Philipp Oeser referenced this issue

2023-06-08 13:01:56 +02:00

Lifetime-Keyframes for particle system are ignored when rendering a baked particle simulation #108729

Philipp Oeser referenced this issue

2023-06-08 16:11:44 +02:00

Bake All Dynamics doesn't include all when Lifetime of particle system gets animated #83802

Philipp Oeser referenced this issue

2023-06-29 09:51:33 +02:00

Particle system bug with texture influence #109450

Sign in to join this conversation.

No Label

Animation & Rigging

Asset Browser Project Overview

Automated Testing

Blender Asset Bundle

Dependency Graph

Development Management

EEVEE & Viewport

Images & Movies

Motion Tracking

Nodes & Physics

Pipeline, Assets & IO

Platforms, Builds & Tests

Render & Cycles

Render Pipeline

Sculpt, Paint & Texture

Video Sequencer

Virtual Reality

Blender 2.8 Project

Milestone 1: Basic, Local Asset Browser

Good First Issue

Animation & Rigging

Development Management

EEVEE & Viewport

Nodes & Physics

Pipeline, Assets & IO

Platforms, Builds & Tests

Render & Cycles

Sculpt, Paint & Texture

Needs Info from Developers

Needs Information from User

No Milestone

No project

No Assignees

4 Participants

Notifications

Due Date

The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: blender/blender#107447

No description provided.