So, with a final suggestion by Rune I've sorted some problems with the
particle system and now proudly present another trial scene which tests the
stability of the system with objects being inside curved objects etc. Enjoy!

Regards,
Tim

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>
Email: tim.nikias (@) nolights.de

Post a reply to this message

Attachments:
Download 'parsys.mpg' (415 KB)

Looking good, looking good... :)

Rune
--
3D images and anims, include files, tutorials and more:
rune|vision:  http://runevision.com **updated Jan 29**
POV-Ray Ring: http://webring.povray.co.uk

Post a reply to this message

Tim Nikias v2.0 wrote:
> So, with a final suggestion by Rune I've sorted some problems with the
> particle system and now proudly present another trial scene which tests the
> stability of the system with objects being inside curved objects etc. Enjoy!

I think you need a bit more random variation of the particle starting 
velocities, they all follow exactly the same path at the beginning.

I wonder if you could post the scene setup and the simulation parameters 
- i would like to try how mechsim compares in results and speed.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 11 Jan. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

Christoph Hormann wrote:
> I wonder if you could post the scene setup and the
> simulation parameters - i would like to try how
> mechsim compares in results and speed.

Heh, I could compare it to my system too in that case... :)

By the way - all the sample scenes in my system are available...
http://runevision.com/3d/include/particles/

Rune
--
3D images and anims, include files, tutorials and more:
rune|vision:  http://runevision.com **updated Jan 29**
POV-Ray Ring: http://webring.povray.co.uk

Post a reply to this message

I've posted the scene-file with some comments (and without the running code
for my system) to p.b.s-f. Let me see the results! :-)

Regards,
Tim

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>
Email: tim.nikias (@) nolights.de

Post a reply to this message

Tim Nikias v2.0 wrote:
> I've posted the scene-file with some comments (and without the running code
> for my system) to p.b.s-f. Let me see the results! :-)
> 

It seems you used much less particles than would result from your 
description in the source.  Some first results are here:

http://www.tu-bs.de/~y0013390/files/particle1.mpg
http://www.tu-bs.de/~y0013390/files/particle2.mpg

The first is with quite strong damping during collisions - as you can 
see in the other one i had difficulties making the particles fall 
through the torus with less damping.  With so many particles it looks 
quite confusing - you can't really follow the path of a single particle.

Concerning performance - since there are no collisions between the 
particles and i used the fastest environment method (object based, 
method 2) it simulates pretty fast, 200s for the whole animation while 
another render was running on the computer.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 11 Jan. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

> It seems you used much less particles than would result from your
> description in the source.

Uh? Now that you mention it, you're right. I'm using the source with the
mentioned particle amount here, but I must have posted one of the early
animations where I was using less particles. Sorry for that.

> The first is with quite strong damping during collisions - as you can
> see in the other one i had difficulties making the particles fall
> through the torus with less damping.  With so many particles it looks
> quite confusing - you can't really follow the path of a single particle.

What my system does is subtract 25% of the energy that is parallel to the
surface normal the particle has hit last. You're right about the amount of
particles, too many make it hard to discern one certain particle, but that's
mostly part of the idea of basic particle system, right? :-) That aside: how
about letting the particles start in that spherical area I mentioned in the
source? You could easily generate one with a random-stream which is based
off the frame-number, just place a particle with
<rand(R),rand(R),rand(R)>*.2-.1 for a 0.1-radius-sphere and add <-2,3,0> for
the center (you should know that, but I thought I'd just make it explicitly
clear what I'm talking about).

> Concerning performance - since there are no collisions between the
> particles and i used the fastest environment method (object based,
> method 2) it simulates pretty fast, 200s for the whole animation while
> another render was running on the computer.

Not sure what 200seconds refers to now. Parsing total? Parsing plus
rendering, and if so, with which resolution? How many frames total? And what
kind of PC do you have? And how does the system work? From frame to frame,
or independant of frames in given intervals?

I'm curious how Rune's system performs, since his is POV-SDL like mine.
Yours is compiled along into the Pov-Source, right? There just has to be
some speed-up with precompiled scripts.

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>
Email: tim.nikias (@) nolights.de

"Christoph Hormann" <chr### [at] gmxde> schrieb im Newsbeitrag
news:ftp### [at] tritonimagicode...
> Tim Nikias v2.0 wrote:
> > I've posted the scene-file with some comments (and without the running
code
> > for my system) to p.b.s-f. Let me see the results! :-)
> >
>
> It seems you used much less particles than would result from your
> description in the source.  Some first results are here:
>
> http://www.tu-bs.de/~y0013390/files/particle1.mpg
> http://www.tu-bs.de/~y0013390/files/particle2.mpg
>
> The first is with quite strong damping during collisions - as you can
> see in the other one i had difficulties making the particles fall
> through the torus with less damping.  With so many particles it looks
> quite confusing - you can't really follow the path of a single particle.
>
> Concerning performance - since there are no collisions between the
> particles and i used the fastest environment method (object based,
> method 2) it simulates pretty fast, 200s for the whole animation while
> another render was running on the computer.
>
> Christoph
>
> -- 
> POV-Ray tutorials, include files, Sim-POV,
> HCR-Edit and more: http://www.tu-bs.de/~y0013390/
> Last updated 11 Jan. 2004 _____./\/^>_*_<^\/\.______
>

Post a reply to this message

Tim Nikias v2.0 wrote:
> 
> What my system does is subtract 25% of the energy that is parallel to the
> surface normal the particle has hit last. You're right about the amount of
> particles, too many make it hard to discern one certain particle, but that's
> mostly part of the idea of basic particle system, right? :-) That aside: how
> about letting the particles start in that spherical area I mentioned in the
> source? You could easily generate one with a random-stream which is based
> off the frame-number, just place a particle with
> <rand(R),rand(R),rand(R)>*.2-.1 for a 0.1-radius-sphere and add <-2,3,0> for
> the center (you should know that, but I thought I'd just make it explicitly
> clear what I'm talking about).

I do so, just the area is somewhat smaller than you describe - things 
get very random otherwise.

> 
> Not sure what 200seconds refers to now. Parsing total? Parsing plus
> rendering, and if so, with which resolution? How many frames total? And what
> kind of PC do you have? And how does the system work? From frame to frame,
> or independant of frames in given intervals?

200s simulation time for the whole animation (450 frames), megapov 1.1 
has separate time statistics for the simulation.  That's about 1/10 of 
the render time (Athlon XP 2600+ with other tasks running).

Here is another version using function based and method 1 (force based) 
environment.  Simulation time is about 10 times of the previous ones.

http://www.tu-bs.de/~y0013390/files/particle3.mpg

Complete description of the techniques can be found on:

http://megapov.inetart.net/manual/global_settings.html#mechsim

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 11 Jan. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

> 200s simulation time for the whole animation (450 frames), megapov 1.1
> has separate time statistics for the simulation.  That's about 1/10 of
> the render time (Athlon XP 2600+ with other tasks running).

Ah, I see. My latest run through the scene took 0.2 seconds at the beginning
and 0.5 seconds at the end per frame, mainly because when the particles slow
down, they tend to always have contact with a surface, thus more
calculations need to be done for interaction. Parsing total for 450 frames
was 221 seconds, so I guess that's not too bad for interpreted POV-SDL
compared to your compiled script. :-)
Parsing time will increase though when I add some more features that are as
of now still 100% missing.

> Here is another version using function based and method 1 (force based)
> environment.  Simulation time is about 10 times of the previous ones.
>
> http://www.tu-bs.de/~y0013390/files/particle3.mpg

Looks good, though some damping might be required for the animation to stop.
Another thing I noticed: one particle sticks to the middle platform
underneath the first torus. Its like glued to the right side, but very
slowly drops, and I guess if you added a few more frames, the particle would
finally drop to the floor. My question is: what's the reason for this
stickyness?

> Complete description of the techniques can be found on:
>
> http://megapov.inetart.net/manual/global_settings.html#mechsim

I've had a look at the Mechsim-Docs some time ago when you posted lots of
nice animations and made the patch available. Looks good and useful, but I'm
the "Do-It-Yourself"-Type, so... And scripting is fun! :-)

-- 
"Tim Nikias v2.0"
Homepage: <http://www.nolights.de>
Email: tim.nikias (@) nolights.de

Post a reply to this message

Tim Nikias v2.0 wrote:
>>200s simulation time for the whole animation (450 frames), megapov 1.1
>>has separate time statistics for the simulation.  That's about 1/10 of
>>the render time (Athlon XP 2600+ with other tasks running).
> 
> 
> Ah, I see. My latest run through the scene took 0.2 seconds at the beginning
> and 0.5 seconds at the end per frame, mainly because when the particles slow
> down, they tend to always have contact with a surface, thus more
> calculations need to be done for interaction. Parsing total for 450 frames
> was 221 seconds, so I guess that's not too bad for interpreted POV-SDL
> compared to your compiled script. :-)

You should realize a few points though:

- my system is not optimized for such fast running simulations, reading 
and writing the simulation data takes a good amount of the time.
- i used 200 steps per frame which is probably far more than necessary 
with method 2 (i used the same 200 steps for particle3.mpg which is much 
more critical in that concern).

>>http://www.tu-bs.de/~y0013390/files/particle3.mpg
> 
> Looks good, though some damping might be required for the animation to stop.
> Another thing I noticed: one particle sticks to the middle platform
> underneath the first torus. Its like glued to the right side, but very
> slowly drops, and I guess if you added a few more frames, the particle would
> finally drop to the floor. My question is: what's the reason for this
> stickyness?

It is probably not visible in the strongly compressed mpeg but that 
particle is not sticking but hopping with hardly any sideward movement. 
  It will either fall off the platform after some more seconds or it 
will come to lie on the platform due to the collision damping.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 11 Jan. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message