Particle System Bounds too large - dynamic lights influence

I am finding the bounds calculated for a particle system is too large. I can view the bounds and the volume appears much to large for the particles drawn. I believe this causes a dynamic lighting issue with the shader used on the particles. Due to the large bounds some "important" lights override closer non important lights and cause the lighting of the particle to pop harshly. Because the bounds are too large the important lights radius doesn't reach the particles inside the bound, and so the important lights do not affect the final lighting of the particle.

Is there a way i can manually override the bounds of the system to reduce it?

 

I'll need to post the system later tonight, but i can see via wireframe the particles occupy a much smaller area the yellow bounds box that the system uses.

 

am i right in assuming the bounds are used to check which lights should affect a system?

 

here is a picture of the system, the bounds seems particularly large, can i over ride the values and set the bounds myself?

 

actually setting to world space seemed to have helped...but i'll lose some performance this way?

 

seems a shame to lose performance in order to get a smaller bounds : (...

 

Hi !

Would it be possible to disable this bounds calculation and set my own by script ?

 

I'm moving particle along a spline, and sometime the particles are not displayed while the camera is very near (inside the bounds)
But if I rotate the camera toward the origin of the particle emiter, than they show up again.
Since my spline is not moving I know the bounds, and I wanted to set this bounding box to the particle system once en for all and see if it could fix it... else recomputing the bounds to always be in front of the camera would be my second option.

But since it's not possible... I'll try to find what else could cause this problem.

 

It's not the clipping plane.

I have a first person controller on the camera.
When the particles are gone, I just have to rotate the camera to look away (in the direction of the particles transform origin), then I rotate back and they are visible again.

Note : this is occuring only when I stop the simulation and restart it on the fly with a new particles emission rate (+prewarm).

EDIT : if the transform origin of the particles emitter is in the camera fov, the bug is gone.

 

Thats it ! "Automatic" did the trick ! Thanks a lot
(it was on "Pause and catch up")

 

I tried 'Automatic' but the bounds did not scale. Particles are set in code using ParticleSystem.SetParticles, but the bounds do no represent the actual bounds of the particles. If the camera frustum leaves the PS bounds, the particles do not render.



If the bounds cannot be set through code, how does one get around this error?

 

Karl,
After some noodling, I discovered that the bounds would update if I enabled (then disabled) the emission module after calling SetParticles. Hope this helps.

Code (CSharp):

1. PS.SetParticles(particles, particles.Length);
2.  
3. var emis = PS.emission;
4. emis.enabled = true;
5. emis.enabled = false;

 

I took the time to pack up a nice clean project to demo the behavior, and submitted a bug report.
(Case 1273773)

 

Has this been fixed? As of what version?