PDA

View Full Version : Particles>How do I fix this problem>



4dartist
09-17-2009, 02:20 PM
Anyone know how to fix this? When you emit fast particles from a fast moving object or from a static object with fast wind, you get this nasty looking stair stepping effect. I use particles to make fire from control motors on some NASA rockets and well.. they go fast, and the fire moves fast.. So I'm always having to slow stuff down or cheat things to mask or get rid of this stair stepping.

Anyone know what causes it? Is there something you can do to solve the motion of particles sub-frame?

Thanks

Attached is a scene file I setup in about 15 seconds just to illustrate the problem. Also an image of the issue.

prometheus
09-18-2009, 04:16 AM
do you really need that high value of 400 m on the directional winds x vectors, that is really destroying it, 50m and rising the y to 10 m perhaps..
but I donīt know if that slows it down to much for you?

Michael

4dartist
09-18-2009, 08:02 AM
Well I exaggerated the values to show what I am talking about. But many times, the emitter is a child of a rocket that is moving pretty fast and it emits fire/smoke, and smoke stops almost instantly when it hits moving air, so the same effect happens. And honestly when I watch reference footage of the fire I'm trying to emulate its the furthest thing from a camp fire, it shoots out as a violent high pressure flame, so they do need to travel fast out of the nozzle too.

I guess it's just a limitation I'm going to have to live with...

Otterman
09-18-2009, 08:32 AM
Oo nasty...I dont use particles very often but i played with every kinda setting possible and it still doesnt want to play ball. Short of motion blur in post im stumped....yet another workaround to lw limitations hey!

Wish i could help ya out...i owe you several ben!

prometheus
09-18-2009, 09:31 AM
well yeah..fast rocket fire travels fast, but I donīt think you should divert it with winds that way, I donīt think rocket fire thrusts are diverting from itīs initial velocity power thrust direction, except for a very short burst close
to the rocket thruster wich can be spread by vibration etc.

Im sure it can be done properly in other ways to get that motion your after..do you have a link to a videoref?

Michael

4dartist
09-18-2009, 10:08 AM
I'm sort of being confusing, so I have to apologize for that.

I was trying to not get into the technical reasons for the use of emitters like this haha.

I only used the wind effector to show the artifact, normally the emitter is just traveling fast and there is no wind effector. Also, I'm not talking about the main rocket engine, I'm talking about small motors that fire from the sides or in the reverse direction of travel to slow jettisoned parts during rocket staging as it travels to orbit.. It's still in enough atmos to blow back the smoke.

Lets just assume I know what I'm doing with the emitters/particles/speeds. It doesn't matter if I have the rocket actually traveling in position or it's sitting still at 0,0,0 and I use wind effectors to blow back the smoke to make it look like it's traveling fast. Either scenario if particles change direction quickly LW creates a stair step that is annoying.

I get the feeling it's just something I have to deal with, which is fine, I just had to check in the off chance I was missing something.

dee
09-18-2009, 11:11 AM
Try adjusting Parent Motion, maybe it helps.

4dartist
09-21-2009, 10:10 AM
Well I found a work around. Scale all keys by 2x and render twice as many frames then time scale in after effects. Basically gets lightwave to calculate twice as many times for particles.. Downsides are rendering 2 x the frames. Also, I'm not sure how it effects the particles motions, considering they are still shooting out same speed, but the vehicle is moving 2x slower. :(

I hope core has a particle calculate detail level or something. Subframe calculations.. I dunno.

bazsa73
09-21-2009, 11:49 AM
try to add some vibration, maybe that cancels the step crap

4dartist
09-21-2009, 12:14 PM
It helps to a tiny degree, but not as much as you would hope, also it's fine to add vibration to a point, but then you get away from the look you want. Adjusting parent motion also reduces the effect from emitting from a moving object, but quickly changes the angle the particles leave the emitter, becoming unrealistic and not what you want for your effect.

Castius
09-21-2009, 01:17 PM
Without subframe emission there are only a few ways to deal with speed issue. Motion blur is one way. Particle scaling based on velocity.

When i had to deal with this in XSI. I made a mesh that trailed behind the object. I stretched the object based on it's velocity. As I emitted from this mesh the scale would compensate for the emission rate space stepping.

4dartist
09-21-2009, 01:27 PM
Particle scaling works pretty darn good when you are emitting from a slow or static object. I use that a lot. But when the object they are emitting from is moving fast, it's useless. Mainly because if you scale hypervoxels based on velocity, it uses (world) velocity, thus each voxel scales in the direction of travel more than along the fire path. Which is fine for main rocket thrust, but totally useless for side motors. Or anything that shows a strong arc in particle paths.

Local coordinate hypervoxel scaling would be great to have.

Motion blur does help.

Thanks for the suggestions so far. I wonder if there is a program out there that does calculate subframe..