BerconMetaball 1.0

[MasterBercon]

Displacement is not possible because of the nature of how the potential field is calculated for single ray. You'd need to mesh it first and I still want to stay way from meshing.

It is however possible to add/multiply the density field with procedural (3D) texture, like dimo's vrayiso plugin or volume fog in sp3. Works for stills but for animation it would be fixed to object/world coordinates so the effect wouldn't move with the blobs.

[joconnell]

getcha - I'm looking at a cell division effect right now and I thought it'd be a nice way to add some more small organic detail. I'd imagine it'd be quite annoying to try and take the noise centre / transform from a particle?

[MasterBercon]

It wouldn't be a problem at all. However, it means you have to evaluate the texture map for each particle. Say you have few hundred particles, evaluating texmap for each particle for each step could easily mean doing it hundred of millions of times. Perlin noise becomes pretty slow with these numbers.

[joconnell]

Getcha - you're almost into raymarching type of stuff to see what's visible. either way I'm going to enjoy trying a realflow sim loaded into pflow rendered via this.

[MasterBercon]

Here is an example of multiplying the density field with texture map fixed to world coordinates: http://www.niksula.cs.hut.fi/~jylila...taballs_27.jpg

It was fairly easy to implement.

Again this is totally useless for animated stuff because it uses world coordinates and doesn't move with the blobs.

[duke2]

Here is an example of multiplying the density field with texture map fixed to world coordinates: http://www.niksula.cs.hut.fi/~jylila...taballs_27.jpg

It was fairly easy to implement.

Again this is totally useless for animated stuff because it uses world coordinates and doesn't move with the blobs.

Wow, that's pretty much VRayISO functionality! Coral reefs here we come! ;D

[instinct]

It's a different approach if you think about it. It is multiplies with the Implicit surface in world-space. Hence the blocking out with the checker map. Still a nice touch and great work Bercon!

Regards,
Thorsten

[MasterBercon]

If somebody throws me a nice function to measure distance to mesh object (a very very fast one) I might aswell do a quick replicate of dimo's iso plugin.

[joconnell]

This looks very interesting too - http://www.xsibase.com/forum/index.p...threadid=39689 - based on this - http://www.tessendorf.org/

Next challenge maybe

[MasterBercon]

Okey now motion blur for particles changing size works too.

I also go the motion blur samples working, here is an example about that: http://www.niksula.cs.hut.fi/~jylila...taballs_28.jpg

There is bug in the 1.0 version and the above example (even sample 2 one) would have been impossible to reproduce since it reads the velocity of the particles. The example uses Postion Object operator with "Lock to Object" mode, which doesn't change the velocity of the particles, while they obviously should inherit it from the parent (box in this case). The new version simply samples particle position and computes the velocity from those.

Motion blur should also work fine when particles are born/deleted inbetween frames. The system handles each motion blur sample individually so they are independent of each other. So if some particle happens to get deleted somewhere within frames, only those motion blur intervals (within the frame) of the particle are drawn where it exists.

So the motion blur should now work correctly no matter how you twist your particles around.

There is still some testing to do and I'll also have to implement the texture support properly and optimize it a bit. If there aren't any major glitches I should complete the 1.1 version during next week.

[percydaman]

this looks sweet. your a beast!

[MasterBercon]

Some spare time at work: http://www.niksula.cs.hut.fi/~jylila...taballs_29.jpg

[jacksc02]

nice Jerry.

[re:FORM]

Oh well, I guess I'll update it to support variable size and multiple motion blur steps.

Hi Master Bercon!
Thanks for this great tool. I don't suppose you've a beta version implimenting p-flow based blob-size yet?

[MasterBercon]

My versions under development only have some single feature working at any given time or none at all. Its not linear process of adding single feature while keeping everything else working. This is such a small scale project and since I'm the sole developer there is no point keeping stuff working while adding stuff. So when I develope stuff I'll just smash it to pieces and clue things back together.

http://www.niksula.cs.hut.fi/~jylila...taballs_29.jpg

Now all the features are there, its just the matter of fixing few things and it should be done. I'll probably release the 1.1 version during next weekend.

EDIT: Here is also picture of the final UI: http://www.niksula.cs.hut.fi/~jylila...taballs_30.jpg