PDA

View Full Version : Krakatoa coming to C4D, how about lightwave



geo_n
12-18-2013, 02:51 AM
http://www.thinkboxsoftware.com/news/2013/12/16/thinkbox-software-announces-krakatoa-for-maxon-cinema-4d.html

My mentor uses krakatoa for 3dmax a lot and I always wanted something like this for lightwave.
Seeing that hypervoxels have not been updated in ages, maybe its better to integrate than to create.
Lw 12 needs to seriously update its particle engine imho.

saranine
12-18-2013, 02:55 AM
I worry about Lightwave missing the boat. Sigh.

jwiede
12-18-2013, 03:27 AM
As with how Vray was ported to C4D, the Krakatoa port to C4D seems driven by a third-party company (Katachi/Navie, C4D developer of Effex, a fluids/liquids sim pkg) stepping in, doing the legwork, and basically putting C4D on Thinkbox's radar by investing development effort to do so.

The main reason LW isn't being "supported" seems to be because a LW third-party plugin developer hasn't funded and developed a Krakatoa bridge for LW.

prometheus
12-18-2013, 12:00 PM
me wantīs it too, but I donīt think that will happen, I think and recall sligthly that I have mailed the creators of krakatoa, and the answer was simply ..no there was no plans for lightwave, but things can
change of course.
Cinema4d just got it so who knows..and cinema4d gets more and more attractive everyday.

And yes we do indeed need particle enhancements, the list is so huge though, that I wonder if they ever will catch up or just fall behind more and more.
I recal Dstorms jetstream allowing for fast openGL render of particles, along with effects to double and fine tune particles, but that is dead in the water.

We did get the 1 000 000 limit removed not long ago, so we can render over 8 millions (havenīt tried over that) of sprite particles, but reaching 3 millions of particles will get your system to cook the particle settlement and at the same time halt your system for a while, so a better particle "cooking" and handling would be nice, sprites renders pretty fast...but have no idea on how that compares to krakatoa which I believe renders quite fast in a volume mode, krakatoa also has some type of partitition handling, and you simulate with lesser amount of particles and then render with higher amounts...I think.

Apart from that, Lightwave needs a boost in vortex,noise fields...and then we havenīt started to talk about bullet intergration:)
I keep my thumbs up for the lightwave group to enhance lw 12 with enhanced particle systems, and hypervoxels.

363 000 particles smoke style low res...
http://www.youtube.com/watch?v=8AUguci1lPY


Texture driven particles on the velocity channel, 3 mil particles.

http://www.youtube.com/watch?v=OYpvZyikRqY


Michael

short223
12-18-2013, 01:21 PM
Out of curiosity, what is the difference between this and Turbulence (other than stability?)

geo_n
12-18-2013, 10:42 PM
jwiede - how long did it take for the connection to do?

short223 - krakatoa is more advance than any pyro toolkit afaik. Some of the stuff I've seen personally can't be done on fumefx or similar. Krakatoa is more versatile like hypervoxels but to another quality level.

This should be interesting for newtek since most lw work involves creating dust and particle effects for any given tv show. Here's an interesting review about lightwave 11.6 and how its mentioned that particles need updating soon
http://www.cgchannel.com/2013/11/review-lightwave-11-0-to-11-6/
at the bottom of the article
Features that should be added in LightWave 12: a personal wish list

prometheus
12-19-2013, 02:26 AM
Out of curiosity, what is the difference between this and Turbulence (other than stability?)

Well ...I donīt have that much insight on it, but maybe think of It this way, krakatoa is the renderer of particles from various particle sources and donīt do any simulation by itself, And fluids does that with special dedicated eqvations, so you would have to simulate the particles in houdini,maya,max,cinema4d thinking particles for example and then employ the renderer.
This also means a type of particles that are not bound to containers or voxel resolution as fluids are, and thus krakatoa is more of advantage when working with large areas that need have particle simulations, and also faster render speed.

Then again since krakatoa doesnīt simulate, you are limited to what different other particle system can simulate and thus fluid behavior needs to be translated by simply using fumefx,mayafluids,houdini fluids and use a particle advection tool that letīs you push the particles in a fluid behavior, and at the end render them with the particle renderer called krakatoa.

So it depends on the look of the effect, since particles can have a more distinct look and different attributes than fluids, and it also depends on how complex and large scale the shot will be, but it
goes on to even more complex stuff regarding particle emission,collisions etc that fluids might not be suitable for.

I would say and think the krakatoa renderer is way faster than rendering fluids as well.
So since krakatoa doesnīt do any simulation, that is way we se fumefx, or naiad fluids that pushes particles, but then employs krakatoa for rendering the volumetrics.
So even if we could get krakatoa for Lightwave, we will lack some advanced particle simulation tools(thinking particles) and advanced noise fields and perhaps full node control over it.

We do have TurbulenceFD, but that is not compliant to work with particles yet(particle advection) so you cant perform fluid behavior on the particles and then render with krakatoa if we were to have krakatoa anyway.

To Sum it up, we need...
1. krakatoa for lightwave
2. particle advection in turbulenceFD
3. more advanced particle control in lightwave, thinking particles, vorticles, new noise/vortex fields, node controllable particles etc....Then we are in the game again:)

Edit...Reservation,this description might not be completly true or technical correct.

Michael

jwiede
12-19-2013, 03:48 AM
If LW3DG doesn't improve LW's particles soon... :devil:

Rendering ~4M particles in X-Particles 2.x (C4D) (http://vimeo.com/76465279)

More fun with X-Particles 2.x:
https://www.youtube.com/embed/YssqlckYR54?rel=0

X-Particles 2.5 grinds through surfacing and rendering millions of particles without breaking much of a sweat. LW's Pfx/Hypervoxels were groundbreaking for their time, but they're really showing their age of late.

geo_n
12-19-2013, 05:42 AM
Thinking particles probably is not needed for lw because lightwave has a flexible node system already in place. Max can use krakatoa with even just using the native particle flow in max which is like the node editor in lw. So only a connection would be needed.

fun with vray, krakatoa

http://www.youtube.com/watch?v=5v_hfVsUlnk&feature=youtu.be

prometheus
12-20-2013, 12:02 AM
Thinking particles probably is not needed for lw because lightwave has a flexible node system already in place. Max can use krakatoa with even just using the native particle flow in max which is like the node editor in lw. So only a connection would be needed.



Uhhm...particle flow in Lightwave with nodes? havenīt seen anything of that, particle emitters donīt have any node editor, except a very limited procedural texture node editor that could be called upon, in textured channels, but I fail to see and havenīt seen any particle control that way.

Do you have any samples of controling particles with nodes? I must have missed it in such case.

Michael

50one
12-20-2013, 12:27 AM
Uhhm...particle flow in Lightwave with nodes? havenīt seen anything of that, particle emitters donīt have any node editor, except a very limited procedural texture node editor that could be called upon, in textured channels, but I fail to see and havenīt seen any particle control that way.

Do you have any samples of controling particles with nodes? I must have missed it in such case.

Michael

I think what geo said is 'groundwork is in place we need the connection between node editor and particles"...


But the thing is, it's not only the connection that is needed but the usability factor that counts.
If I'll get a node editor with nodes such as vectors, multiply, add etc.with only three people smart enough to use it, then cheers, I'll go to a package that got nodes that describe what they do, like particle birth rate, speed etc etc.


That's the current problem with Lw nodes for me, but maybe it's the lack of proper documentation, not only few lines what the node do, but a small i age also describing the principles, but that's a different story...

Anyway, I'll see what 12 will bring in this and other fields as I'm more and more convincing myself to jump onto Houdini platform.

prometheus
12-20-2013, 12:57 AM
I think what geo said is 'groundwork is in place we need the connection between node editor and particles"...


But the thing is, it's not only the connection that is needed but the usability factor that counts.
If I'll get a node editor with nodes such as vectors, multiply, add etc.with only three people smart enough to use it, then cheers, I'll go to a package that got nodes that describe what they do, like particle birth rate, speed etc etc.




That's the current problem with Lw nodes for me, but maybe it's the lack of proper documentation, not only few lines what the node do, but a small i age also describing the principles, but that's a different story...

Anyway, I'll see what 12 will bring in this and other fields as I'm more and more convincing myself to jump onto Houdini platform.

Donīt know what he ment? to me he said lw has a flexible node system in place, and that I can not see if it hasnīt been adapted to work with particles so far to this date.
And the particle flow in max doesnīt have any counterpart in lightwave nodes, itīs completly different states.
If he said connection between nodes and particles..then I understand, but to me it sounded like he ment connection between krakatoa and lightwave.
But I might have misinterpreted...Only Geo_n can say what he really ment, have to wait for that.)

I agree with you too, controlling it all with nodes are tricky, but then I donīt follow you if you want to jump in to houdini, that will be even trickier with that kind of node architecture.
modo particles really came along fast and with a lot of node control, impressing..at the same time just applying an image map in modo for particle birth is much trickier to setup since you have to go the node
connection approach, where in lightwave you do that much easier/faster with the texture channel.

Maybe compund nodes with asseblemed node control behavior could work nicely, so you pack the node flow of the particles in a particle noce compound node, and name it accordingly to the behavior..now if they
work on a preset system..you could simply open a particle preset tab, and drag your preset node to the emitter ..and voila you got a nice predetermined flow.


Michael

50one
12-20-2013, 01:29 AM
Donīt know what he ment? to me he said lw has a flexible node system in place, and that I can not see if it hasnīt been adapted to work with particles so far to this date.
And the particle flow in max doesnīt have any counterpart in lightwave nodes, itīs completly different states.
If he said connection between nodes and particles..then I understand, but to me it sounded like he ment connection between krakatoa and lightwave.
But I might have misinterpreted...Only Geo_n can say what he really ment, have to wait for that.)

I agree with you too, controlling it all with nodes are tricky, but then I donīt follow you if you want to jump in to houdini, that will be even trickier with that kind of node architecture.
modo particles really came along fast and with a lot of node control, impressing..at the same time just applying an image map in modo for particle birth is much trickier to setup since you have to go the node
connection approach, where in lightwave you do that much easier/faster with the texture channel.

Maybe compund nodes with asseblemed node control behavior could work nicely, so you pack the node flow of the particles in a particle noce compound node, and name it accordingly to the behavior..now if they
work on a preset system..you could simply open a particle preset tab, and drag your preset node to the emitter ..and voila you got a nice predetermined flow.


Michael

Was talking about houdini as I would really like to learn something that will help me to get some more skill and maybe change my job a little. Last time I used Houdini was version 8 and as far as I can tell a lot has changed.

My bigget problem with nodes is their usability, in Modo When you want to move item with effector and have some falloff applied you just go to node editor and choose nodes you need, yes you can add heavy math but it's much easier to do it when you got the basic setup ready and node layout seems to be more logical.

Have a look at the video I've done in Modo 601(this is not particles related btw. But nodes in general) - http://vimeo.com/58212744 even tho it's quite simple I'm still struggling to recreate this in LW, due to DP move and all that multiply/add mumbo-jumbo:) one thing is certain Swampy could do this in less than 20minutes lol

Agree with you on Modo, but since it's the very first iteration and it works great, I'm sure there will be lot's of updates in upcoming 801 to the particle engine.

creacon
12-20-2013, 02:12 AM
I agree that the particle systems in LW need an upgrade, but HV is far from being slow, in fact it is the fastest "isosurface" renderer I know of. Combined with VPR it could be one of the best FX systems around. Be patient I am working on it ;-)

creacon

creacon
12-20-2013, 02:16 AM
My system does that and a lot more. But LW has some nodal possibilities built in, though I agree it's not really user friendly and not everything is possible.

creacon


Uhhm...particle flow in Lightwave with nodes? havenīt seen anything of that, particle emitters donīt have any node editor, except a very limited procedural texture node editor that could be called upon, in textured channels, but I fail to see and havenīt seen any particle control that way.

Do you have any samples of controling particles with nodes? I must have missed it in such case.

Michael

geo_n
12-20-2013, 02:59 AM
Prom - what 50one said. 3dmax and particleflow can be enough. Thinkingparticles is not absolutely needed.
Now with lightwave, it has very strong nodal base even better than 3dmax currently. With lw 11 they introduced a lot of nodal based tools like instancing, flocking(particle subset), nodal edgeshading. Its probable lightwave has more nodal power hidden that could easily be applied to new nodal particle engine. With that krakatoa can be connected. Lw 12 should offer enough time to develop a new nodal particle engine and maybe new hypervoxels. Considering lw is used in quick vfx shots this should be one the top priority feature for lw 12.

prometheus
12-20-2013, 03:20 AM
I agree that the particle systems in LW need an upgrade, but HV is far from being slow, in fact it is the fastest "isosurface" renderer I know of. Combined with VPR it could be one of the best FX systems around. Be patient I am working on it ;-)

creacon


My system does that and a lot more. But LW has some nodal possibilities built in, though I agree it's not really user friendly and not everything is possible.

creacon
Your system? would like to see screenshots of some sample of that, and maybe animation sample if you could share please.


HV slow or fast? it all depends on how it is set up, you canīt run two seperate hv groups (two nulls) and have them intersect with a fairly large scale, it will halt the system when rendering enourmously.
Modo is said to be faster with deep shadow maps.
and houdini cloudfx...that is depending on resolution in fact realtime, in openGL. with geometry objects in full volumetrics, however..it isnīt that fast to render finally..compared to hypervoxels.

and I indeed would like to see Item/ or implicit surface volumetrics allowing for us to use geometric shapes as voxels, such as modo and houdini has, not only particles or vertices, volumedic sort of does it but for
additional costs, and who donīt remember old dynamite which could do that ..sort of, Now...I have lost some beliefs in that we will see anything happening soon unfortunatly, so what geo_n said about it being a top priority feature for lw 12, and regarding itīs history for quick vfx shot...I really have my doubts we will se anything in the first lw versions, Ivé been so dissapointed in the neglection of Hv improvements from 9 -to lw 11.6..unfortunatly.

enhancements were made for hv volume mode with blending that is poorly implemented compared to the old dynamite plugin blending mode, some enhancements and feature improvements thanks to dpont, and a procedural node hack..where you can cut off the falloff, and change shapes and distortions in hypervoxels..but since it is a hack..the tweaking updates arenīt updating automaticly, as it does when changing hvīs in main properties, it is utterly slow in workflow to click update vpr to see every tiny change in frequencie, scaling or something alike that within the node editor textures for hypervoxels when they donīt autoupdate.
I hope node editing of hypervoxels becomes more efficient and reacts to auto update in vpr as it does when tweaking the main parameters for hypervoxels.

creacon
12-20-2013, 06:36 AM
http://www.youtube.com/channel/UCuv590EM3dC52M0w6gFbuZQThe old stuff is here, but I am a lot further than that right now.
Have been working more on the liquids than on the dumb particles, but the dumb particles have a node editor which allows you to drive velocity, color, size of the particles using nodes.

There's a chance that I am going to use the PhysX dumb particles, so they interact with solids and forces too.
But first things first: Get a good particle caching implemented.

creacon

prometheus
12-20-2013, 07:11 AM
I agree that the particle systems in LW need an upgrade, but HV is far from being slow, in fact it is the fastest "isosurface" renderer I know of. Combined with VPR it could be one of the best FX systems around. Be patient I am working on it ;-)

creacon


http://www.youtube.com/channel/UCuv590EM3dC52M0w6gFbuZQThe old stuff is here, but I am a lot further than that right now.
Have been working more on the liquids than on the dumb particles, but the dumb particles have a node editor which allows you to drive velocity, color, size of the particles using nodes.

There's a chance that I am going to use the PhysX dumb particles, so they interact with solids and forces too.
But first things first: Get a good particle caching implemented.

creacon


Keep up the work, looks interesting for liquid stuff.

Off topic sort of, I really would like to direct particles in preconstructed paths, that is..letīs say you create a tree line poly skelegon as for example, using dp_verdure tree tool, and create a tree branch in skelegon mode..then port it to layout and let the particles flow from the beginning of that tree branch and the particle flows along and divides and follows along the other branches, if we could do that.. I can imagine some really awesome partlicle flow stuff. sort of converting a mesh to velocity vector flow.
the only thing that comes close today is to use a collision object set to event and move it through a two point poly string and let it emitt along the way, but thatīs not the same as flowing along such branch.

Im not sure if modo can do that, the implemented particle to mesh, and then resculpting the mesh guide and then convert back to particle path is cool, but not sure if you could extract a premade mesh, or splinepaths and then convert
to particle vector velocity flow in modo?

Right now we do only have the dynamic wind path to direct particles in predefined paths where particles flow along one defined path(I hate setting it up with nulls and move etc) or we can push and direct particles in velocity channels
of the particles motion or wind direction, but then you canīt direct is exactly as you want, merely use procedurals to drive it.

So a tool to construct mesh skelegon paths, or spline paths that can then convert to particle velocity flow and use the mesh/guideīs different branch paths in a modeled layer would be awesome.

Michael

geo_n
12-20-2013, 08:03 AM
Impressive
http://www.youtube.com/watch?v=iHulEPSEVUo
Liquid in lw!. Wonder if the blobby look from hypervoxels are eliminated.

creacon
12-20-2013, 08:19 AM
Not at all, you'll have to use HV to render this.

creacon

jwiede
12-30-2013, 07:12 AM
I agree that the particle systems in LW need an upgrade, but HV is far from being slow, in fact it is the fastest "isosurface" renderer I know of.

Have you ever used Navie's Krakatoa's renderer, or even V2.5+ X-Particles' render materials? Given the rate they crank through tens of millions of particles and more with full GI, even with (f.e.) translucent, illuminating particles, I can't really agree with your assessment. VrayForC4D's 1.8.x volumetrics are quite impressively performant now, as well.

jwiede
12-30-2013, 07:32 AM
My system does that and a lot more.

Of course, their capabilities and performance are likely to improve in the future as well, so it isn't really fair to make claims about your as-yet-unreleased product compared against their currently available products.