PDA

View Full Version : Deep Rising FX v1.1 Update Released



darkChief
02-20-2017, 04:05 PM
Deep Rising FX v1.1 Update!

Check out the demo :D

www.deepfxworld.com (http://www.deepfxworld.com/index.php?route=product/product&path=59&product_id=50)

New Features:

Simulation Preview
- You can now preview the simulation progress in the viewport.

Particle Preview settings
- New preview particle primitive.
- You can change preview particle size.
- You can change the color gradient that is based on the particles velocity.
- Adjust preview peak velocity for color gradient.
- Can reactivate old point preview primitive. Which is faster to render.

Simulation Button
- New simulation button for the Lightwave menus. Shortcut to solvers simulate button.

Hyper Voxel Support
- You can now attach a particle system to any cache. This will allow you to use hypervoxels on the particles.

Circle & Square Emitter Randomness
- You can now randomise particles birthed from emitters.


https://www.youtube.com/watch?v=ZFrLep7A0tY

samurai_x
02-26-2017, 09:22 PM
https://www.youtube.com/watch?v=KOH_x0ax8VI
Dragon 2m X 1m X 3m.
Tank 8.5m X 4m X 4m
125,000 fluid particles

Looks cool!

darkChief
02-27-2017, 03:15 PM
Thanks

jwiede
02-27-2017, 07:26 PM
Thanks

Aw! Should have replied "Tanks!" ;D

samurai_x
02-28-2017, 04:09 AM
jwlede made a joke in nt forums. Unbelievable :D


OT.

Would you ever consider porting Redshift renderer to lightwave seeing as you are a very capable developer and made deeprising so quick?
These companies are always looking for third party devs to port their renderers.

darkChief
02-28-2017, 05:47 AM
Well rendering is not my thing at the moment, would take some time to accumulate the required knowledge. Also considered making a renderer at some point :D but that market is obviously saturated.

Also compared to other feature requests the demand for Redshift seems low (from my observations here anyway).

I will be peddling dynamics and solvers for now;)

samurai_x
02-28-2017, 09:44 AM
Well rendering is not my thing at the moment, would take some time to accumulate the required knowledge. Also considered making a renderer at some point :D but that market is obviously saturated.

Also compared to other feature requests the demand for Redshift seems low (from my observations here anyway).

I will be peddling dynamics and solvers for now;)


Yeah its saturated right now.

But just wait, Redshift will be the next vray and all these other renderers will be "rendered" obsolete. "D
It will be as common as vray for anybody doing 3d that wants gpu speed. Lightwave will be left out again sadly.

Looking forward to other plugins you create for lightwave :D Its developing at a good pace.

darkChief
03-04-2017, 09:10 AM
Another Deep Rising FX test, this time a simple flood test.


https://www.youtube.com/watch?v=_7Ai4Xrv04k

ianr
03-04-2017, 09:36 AM
Yeah its saturated right now.

But just wait, Redshift will be the next vray and all these other renderers will be "rendered" obsolete. "D
It will be as common as vray for anybody doing 3d that wants gpu speed. Lightwave will be left out again sadly.

Looking forward to other plugins you create for lightwave :D Its developing at a good pace.


Dude, Let D.C. Mambo do his thang on his wet stuff.

(nice try though!)

i'll p.m. you soon.

hrgiger
03-04-2017, 09:44 AM
Looks good dark chief.

darkChief
03-04-2017, 12:15 PM
Thanks

darkChief
03-13-2017, 06:01 AM
https://www.youtube.com/watch?v=5O1LEPDlAQU

prometheus
03-13-2017, 11:41 AM
keep em coming.

I notice that there is often a sort of step issue, often initially when something is poured etc, can something be done to improve on that..and what is the cause of it? maybe lack of subframe calculation or is it the blending? resolution?

darkChief
03-13-2017, 04:29 PM
Ummmm not sure exactly which part you're talking about.

jboudreau
03-18-2017, 07:23 PM
keep em coming.

I notice that there is often a sort of step issue, often initially when something is poured etc, can something be done to improve on that..and what is the cause of it? maybe lack of subframe calculation or is it the blending? resolution?

Hi

If you change the Particle radius it gets rid of that stepping you are talking about

Thanks,
Jason

Schwyhart
03-18-2017, 08:44 PM
The particle rate of speed seems too slow for the faucet. When I turn on my bathroom faucet, the stream looks white and you can't see through it. If I don't turn it full blast, it is a smaller stream and also comes out slower, making it more transparent, but not like the example shown.
I know these are tests/examples. Just thought I'd let you know :)

darkChief
03-19-2017, 04:41 AM
Thanks for the feedback.

prometheus
03-19-2017, 07:55 AM
Hi

If you change the Particle radius it gets rid of that stepping you are talking about

Thanks,
Jason

Thanks for the info, currently though..I donīt even have the demo, so I woulnīt know how that would react and look like, but wouldnīt increasing particle radius make the liquids to big in other areas ..a general too large particle radius, thatīs where I thought the lack of subframe calculation would be an issue, but maybe I am wrong on that, and would be the same with real flow?

Michael

jboudreau
03-19-2017, 09:25 AM
Thanks for the feedback.

Hi darkChief I was playing around with the demo of Deep Rising FX 1.1, which is quite awesome by the way, Excellent job :)

I came across a bug that pretty much will give you a crash every time. Here are the steps to take

- Create a box in layout or in modeler
- triple the polygons if you made it in modeler layout already triples them for you
- click the deep rising domain button and add the box as a fluid
- click the fluid in the list
- now change the particle spacing from .04 to say .02
- now press the resample mesh button.
- Sometimes it will work but most of the times if you go back and forth between .01 and .04 it will cause a crash or a complete lock up of layout

I have a scene that I can send you that will crash every time after just changing the particle spacing and hitting the resample mesh button

Hope this helps

UPDATE: Now after getting these crashes fluids will no longer show up or create using your plugin. I click on the item and add it to the domain and choose fluid and nothing happens, there are no blue dots at all. If I click on resample mesh immediate crash


Thanks,
Jason

darkChief
03-19-2017, 09:53 AM
UPDATE: Now after getting these crashes fluids will no longer show up or create using your plugin. I click on the item and add it to the domain and choose fluid and nothing happens, there are no blue dots at all. If I click on resample mesh immediate crash


Thanks,
Jason

Hi,

Thanks for the update, I can reproduce the first bug, and it will be fixed immediately. You shouldn't use a particle setting lower than 0.038 for fluids, sadly that can't be enforced in the Ui. But it shouldn't crash if it goes lower.

I will watch out for the second problem. I can't reproduce it at the moment.

How big is your cube?

jboudreau
03-19-2017, 10:00 AM
Hi,

Thanks for the update, I can reproduce the first bug, and it will be fixed immediately. You shouldn't use a particle setting lower than 0.038 for fluids, sadly that can't be enforced in the Ui. But it shouldn't crash if it goes lower.

I will watch out for the second problem. I can't reproduce it at the moment.

No problem. I think I have completely broke your plugin lol. I tried deleting all my configs and rebooting my machine. I re-installed the plugin and still for some reason I can't get it to create fluids at all anymore. It just won't work no more blue dots at all just an empty wireframe box. Very strange. I'm wondering if something is still left running in the task managers processes or something.

I'll keep you posted on my progress of getting this back working again.

UPDATE: It seems to be the volume is no longer working. Surface now gives me the little blue dots but volume does not but it did in the past.

Thanks,
Jason

darkChief
03-19-2017, 10:05 AM
Hi,

I see, interesting, that shouldn't be possible. The plugin stores nothing on your system. Once the process dies everything should be back to normal.

Let me know if you are able to figure something out.

UPDATE: It seems to be the volume is no longer working. Surface now gives me the little blue dots but volume does not but it did in the past.

[/QUOTE]

Bad particle spacing maybe? What happens if the spacing is 0.038?

jboudreau
03-19-2017, 10:18 AM
Hi,

I see, interesting, that shouldn't be possible. The plugin stores nothing on your system. Once the process dies everything should be back to normal.

Let me know if you are able to figure something out.


Okay I figured out the problem.

Steps taken:

- I create a box in layout
- I add the selected box to the domain and choose fluid
- No fluids are created. No little blue dots are made
- I have to change it from volume to surface and then click resample mesh
- this creates the little blue dots again
- When I change it back to volume the little blue dots dissappear.
- If I press resample mesh the blue dots are still gone
- The only way I can get the blue dots to appear with volume set is I have to resize my box.
- Once I resize the box and then press the resample mesh button the blue dots reappear when I'm in volume mode.

After I do that everything works as it should even after I remove all from the domain and start again.

The issue though I have to go through all the above steps once I close down layout and re-open it.

Hope this helps figure out what the issue might be


UPDATE: So it looks like the size of the fluids is staying in memory because if I do the following:

- Create a box 1m x1m x1m (Default) from layout
- Add the box to the domain and choose fluid. No fluids are created
- But if I resize the box then click on resample mesh it works the fluid particles are created.

So this gave me an idea what if I create the box without those 1m x 1m x 1m Default size.

- I created a box 2m x 1m x 1m from layout
- I then added the box to the domain and choose fluid. The Fluids created this time in volume mode.

So for some reason that default size won't work but anything else does

Hope this helps

Thanks,
Jason

darkChief
03-19-2017, 10:21 AM
Okay I figured out the problem.

Steps taken:

- I create a box in layout
- I add the selected box to the domain and choose fluid
- No fluids are created. No little blue dots are made
- I have to change it from volume to surface and then click resample mesh
- this creates the little blue dots again
- When I change it back to volume the little blue dots dissappear.
- If I press resample mesh the blue dots are still gone
- The only way I can get the blue dots to appear with volume set is I have to resize my box.
- Once I resize the box and then press the resample mesh button the blue dots reappear when I'm in volume mode.

After I do that everything works as it should even after I remove all from the domain and start again.

The issue though I have to go through all the above steps once I close down layout and re-open it.

Hope this helps figure out what the issue might be

Thanks,
Jason

Hi,

Yeah, that's bad particle spacing, you can tell if 0 particles are generated. You don't have to go through all those steps. A particle spacing of 0.038 should work. If not let me know.

Regards,

jboudreau
03-19-2017, 10:25 AM
Hi,

Yeah, that's bad particle spacing, you can tell if 0 particles are generated. A particle spacing of 0.038 should work. If not let me know.

Regards,

UPDATE: So it looks like the size of the fluids is staying in memory because if I do the following:

- Create a box 1m x1m x1m (Default) from layout
- Add the box to the domain and choose fluid. No fluids are created
- But if I resize the box then click on resample mesh it works the fluid particles are created.

So this gave me an idea what if I create the box without those 1m x 1m x 1m Default size.

- I created a box 2m x 1m x 1m from layout
- I then added the box to the domain and choose fluid. The Fluids created this time in volume mode.

So for some reason that default size won't work but anything else does

UPDATE: I tried changing it to .038 and pressing the resample mesh and it worked but when I went back to .04 and hit resample mesh it crashed

Hope this helps

Thanks,
Jason

darkChief
03-19-2017, 10:30 AM
Hi,

Don't resize the box. The default particle spacing works for most irregular models. change the particle spacing to 0.038. That's one of its main uses, to get perfect sampling.

jboudreau
03-19-2017, 10:32 AM
Hi,

Don't resize the box. The default particle spacing works for most irregular models. change the particle spacing to 0.038. That's one of its main uses.

Okay but why can't I create a box 1m x 1m x 1m and use a particle spacing of .04? It use to work but now it will not. .038 works but only after I resample the mesh but if I go back to .04 immediate crash

Is .038 now the default because it keeps reverting back to .04 which won't work for me but has in the past for some reason.

Hope this helps

Thanks,
Jason

darkChief
03-19-2017, 10:38 AM
Hi,

To avoid bad voxelizaton (like hanging particles, jagged sampling etc) After the first sweep, the algorithm cleans the voxelizaton, if its not perfect it won't generate anything. This way any type of mesh can be voxelized without artifacts.

jboudreau
03-19-2017, 10:47 AM
Hi,

To avoid bad voxelizaton (like hanging particles, jagged sampling etc) After the first sweep, the algorithm cleans the voxelizaton, if its not perfect it won't generate anything. This way any type of mesh can be voxelized without artifacts.

So if you have to use a .038 all the time how do you add more particles in your fluid container? Also why aren't you able to reshape your fluid container?


Thanks,
Jason

darkChief
03-19-2017, 10:58 AM
Hi,

I'm sure this is mentioned in the documentation. Unlike packages like Realflow, the particle resolution is fixed. You can resize your mesh, I never said you couldn't. You said there were too many steps to sample the 1x1x1m cube, I gave you one step to solve it. I think you will find that only cubes may need adjusting most of the time. However if you are only interested in cubes, changing particle spacing may be necessary.

The workflow to add more fluids is to scale your scene. I know you have both options in something like Realflow, here things are simplified a little.

jboudreau
03-19-2017, 11:11 AM
Hi,

I'm sure this is mentioned in the documentation. Unlike packages like Realflow, the particle resolution is fixed. You can resize your mesh, I never said you couldn't. You said there were too many steps to sample the 1x1x1m cube, I gave you one step to solve it. I think you will find that only cubes may need adjusting most of the time. However if you are only interested in cubes, changing particle spacing may be necessary.

The workflow to add more fluids is to scale your scene. I know you have both options in something like Realflow, here things are simplified a little.

Okay just read the documentation regarding this. It says if you are having any issues to change your particle spacing down and hit the resample mesh button which as I stated above crashes most of the time but you said you were going to fix this which is great. Sorry you had said above not to resize the box.

I think it would be nice if when resize the particle spacing maybe a message pops up saying that the spacing is out of spec or something instead of just crashing. How do other packages besides RealFlow get around this limitatoin because I understand what you are saying if you set it to a fixed .04 particle spacing without being able to change it then some object like the 1m x 1m x 1m box would not work unless you changed it to .038 etc.

Anyway thanks for all your help. Hopefully you can fix that crashing bug.

Thanks,
Jason

darkChief
03-19-2017, 11:17 AM
Hi,

No problem, glad to help. The bug should be fixed in the next build. Small issue.

Regards,

Mambo

jboudreau
03-19-2017, 11:18 AM
I really think something is wrong regarding cubes. I just created a .5m x 1m x .5m cube and added it to the domain and choose fluid and immediate crash.

Why would this be?

Hopefully it's part of the same bug that you fixed

Thanks,
Jason

jboudreau
03-19-2017, 11:32 AM
Hi

So I created this simulation, and exported the mesh sequence. Used the proxy object and brought the mesh sequence back into layout where I surfaced it. Everything worked out great except when I move my timeline to another frame I loose my surface completely it reverts back to the default surface. Is this just a limitation with the Demo?

If so is there a document stating what limitations the Demo has?

Thanks,
Jason

darkChief
03-19-2017, 11:43 AM
Hi,

It's in the documentation. This is one of the workflows I'm not happy about, but unavoidable. You texture your proxy object, and save.

But if you want to take a short cut, save the current object in the sequence after updating the mesh surface. That will overwrite the proxy object.

If you run into a limitation, a message will pop up.

Apart from that bug, I can't reproduce the bug with the cube.

Regards,

jboudreau
03-19-2017, 12:26 PM
Hi,

It's in the documentation. This is one of the workflows I'm not happy about, but unavoidable. You texture your proxy object, and save.

But if you want to take a short cut, save the current object in the sequence after updating the mesh surface. That will overwrite the proxy object.

If you run into a limitation, a message will pop up.

Apart from that bug, I can't reproduce the bug with the cube.

Regards,

Okay so every time you make a change to the surface you need to either save it in modeler or do a save all objects in layout. Yeah that's a bit of a pain. Hopefully you can get around this in later versions.

Thanks,
Jason

darkChief
03-19-2017, 12:45 PM
Hi,

Unfortunately yes, the workflow is awkwards. If you have a really big model, and and the mesh sequence updates during rendering sometimes Lightwave doesn't clear that memory (mainly 32bit version). It also helps if you have a really huge scene and ram counts, instead of having two models in memory, the sequencer replaces it with a proxy (so best to have a small proxy), before creating new geometry.

jboudreau
03-19-2017, 12:51 PM
Hi,

Unfortunately yes, the workflow is awkwards. If you have a really big model, and and the mesh sequence updates during rendering sometimes Lightwave doesn't clear that memory (mainly 32bit version). It also helps if you have a really huge scene and ram counts, instead of having two models in memory, the sequencer replaces it with a proxy (so best to have a small proxy), before creating new geometry.

okay cool thanks for the tip. It's funny that when you scrub the timeline the mesh updates and animates. But when you play the timeline it does not. I guess you have to create a preview to see your animation.

Is there anyway to turn off the bounding box when making previews? It's quite annoying seeing the bounding box in your preview

Will lightwave NEXT have these limitations since it will now handle OpenVDB files. I'm guessing the object replacer will be obsolete then (No need to use it)

Thanks,
Jason

darkChief
03-19-2017, 01:23 PM
I don't know how to get rid of the bounding box.

I don't know what Lightwave Next will have, but hopefully it will have something to replace this system. Openvdb is mainly for storing volume data, not meshes. Alembic fluids would be the best, would eliminate the need for a sequencer.

Regards,

Mambo

jboudreau
03-19-2017, 02:12 PM
I don't know how to get rid of the bounding box.

I don't know what Lightwave Next will have, but hopefully it will have something to replace this system. Openvdb is mainly for storing volume data, not meshes. Alembic fluids would be the best, would eliminate the need for a sequencer.

Regards,

Mambo

I think I understand why users are saying the speed of the particles are not correct. I think you have something wrong with your gravity settings. I have tried Realflow, Lightwave and some other particle systems and when I change the gravity from -9.8 to say +2.0 or 10.0 the particles fly upwards in the air. In your simulation system the particles are hardly affected and just drop straight to the ground. Give it a try and you will see what I mean.

Make an emitter and set your gravity to +2 in the y-axis the particles just fall straight down.

The particles only get affected by the gravity once they hit their collision object.

I set the emitter to point down to the floor and emitted some particles I set my gravity to 10 in the x-axis the particles dropped straight down the gravity had no effect on them until they hit the floor they they started flying off in the x -axis direction.

I have to set it all the way up to 1000 (this does not seem right) in the x-axis for the particles to fly in the x - direction.


I would say something is wrong with the math because you can't use -9.8m real world specs for gravity and then have to go all the way to 1000 to make it work. -9.8m has absolutely no affect on your particles you would have to change that to be closer to -400 on the Y-axis to get accurate (-9.8m real world gravity)

Hope this makes sense

Thanks,
Jason

darkChief
03-19-2017, 04:41 PM
Hi,

Unless you're doing extreme or unorthodox things with the emitter it should work. I've posted an example below, gravity is working for me. Particles shouldn't even collide with objects if gravity is 1000, by the time they reach terminal velocity the time step would be to big to account for the displacement. And the laws of physic do apply, how close is the object to the emitter? A particle will fly in a straight line until an external force is strong enough to move it.



https://www.youtube.com/watch?v=51IWuY_DDe4

jboudreau
03-19-2017, 04:46 PM
Hi,

Unless you're doing extreme or unorthodox things with the emitter it should work. I've posted an example below, gravity is working for me. Particles shouldn't even collide with objects if gravity is 1000, by the time they reach terminal velocity the time step would be to big to account for the displacement. And the laws of physic do apply, how close is the object to the emitter? A particle will fly in a straight line until an external force is strong enough to move it.



https://www.youtube.com/watch?v=51IWuY_DDe4

Must be with the demo then. I have the same setting as you and my particles go straight to the ground. The only way I can get them to do what yours are doing in the video is to set my gravity to 300 or more. I will post a video very shortly to show you.

Test the demo and see if it does the same thing?

Thanks,
Jason

darkChief
03-19-2017, 04:59 PM
Let me check the demo.

jboudreau
03-19-2017, 05:01 PM
As you can see in the video my settings are exactly the same as yours, but my particles go really far down the screen before the gravity takes any affect on them. I'm not sure why your's is different results unless their is something wrong with the Demo.


https://youtu.be/g7X6PWSI2n0

Thanks,
Jason

darkChief
03-19-2017, 05:12 PM
Ah sorry, I forgot to show the emitter strength, I had set it to 5 or 6.

jboudreau
03-19-2017, 05:18 PM
Ah sorry, I forgot to show the emitter strength, I had set it to 5 or 6.

oh okay that was the issue. See I'm going by the default settings. and the default is set to 10 for strength. Shouldn't the default settings give you similar results to how most particle simulations software works including lightwave own particle system because I have to change quite a few settings to get it to be the same as lightwave's particles.

Thanks,
Jason

darkChief
03-19-2017, 05:24 PM
They are very different engines, for one particles can't overlap (SPH in general). Maybe it is coincidence that Lightwave and Realflow have similar default settings that give the results you talk about.

jboudreau
03-19-2017, 10:03 PM
So I've been playing around with the Deep Rising FX plugin and it's great!! Dark Chief has done an amazing job!!

on another note though I see some limitations that may or may not be addressed in future versions.

- Having the solver work for everything limits you from having two different emitters in the scene. (For example you can't set a different gravity or step count... for another emitter. Every emitter uses the same settings)

- You can't change the surface of another emitter because no matter how many emitters you have in the scene they all use the same one surface.

- When you use meshing it meshes all emitters in the scene (no way to mesh a specific emitter only) You may want to mesh one emitter different from the other (for example you might want one emitter to have gaussian blur and one to have dilate)

- It would be great if you could mix too emitters together and have the surfaces blend into one another (like mixing tow different colors of paint together)

Maybe you can add some of these as future feature requests.

Thanks
Jason

darkChief
03-20-2017, 04:09 AM
Hi,

I logged all these as feature requests. It's only at version 1.1 so anything is possible.

I can promise to do my best to get these implemented in the future.

Regards,

Mambo

Markc
03-20-2017, 12:54 PM
Mac version pretty please :thumbsup:

darkChief
03-27-2017, 08:17 AM
Rendering this out, here's a still in the mean time. This is going to be the main feature in the next free update, Bullet physics integration, and two-way coupling. There will be a host of other feature requests also implemented.

I'll be busy getting that done, so I won't be posting that many updates.

https://preview.ibb.co/cbqwRF/rigid_2.jpg (https://ibb.co/fwSzfa)

darkChief
03-27-2017, 04:12 PM
And the video.

https://www.youtube.com/watch?v=U3F1mRrbKPc

darkChief
04-10-2017, 02:44 AM
More rigid body dynamics tests! An ice cube!


https://www.youtube.com/watch?v=pEJOTc1uuB8

mav3rick
04-10-2017, 05:58 AM
almost there :) should float.... imho

darkChief
04-10-2017, 06:18 AM
Depends on the mass and coupling stiffness. Floating objects easily done;)

mav3rick
04-11-2017, 04:12 AM
Depends on the mass and coupling stiffness. Floating objects easily done;)

btw what is about stability of hard body objects/fluids at the end of calc?... is it possible to make it less vibrating and look more natural... is that somethin that deals with substeps or calc resolution

darkChief
04-11-2017, 05:27 AM
btw what is about stability of hard body objects/fluids at the end of calc?... is it possible to make it less vibrating and look more natural... is that somethin that deals with substeps or calc resolution

Just in time for a new clip;)

I like the little jiggle at the end, since there are forces working inside the fluids (like vorticity):D But you can tone it down or smooth it out completely. That's what the coupling stiffness does. When stitching physics engines together its easy to forget forces won't be equal (different integrators or iterative solvers that converge differently), so the stiffness lets you adjust how strongly the fluid forces affect bullet dynamics .

Timestep and substep automatically synchronised.

Also I was having trouble with Bullets default impulse based constraint solver, hard to to tune and synchronize in my opinion. Luckily its a very modular platform and they have different solvers that can be plugged in. I will be using the MCLP solver, because its more accurate, like pure SPH.

And you asked about floating objects, check out the clip. The surface area also affects how things float or sink.


https://www.youtube.com/watch?v=W7H-yOQlRE4

mav3rick
04-11-2017, 08:56 AM
nice :) now how it works with surface/weight compensation/penetration.. i mean if you have wood cube falling in water it will sink a little until weight is compensate with water density and pull it up on surface... i can still feel bullet detection is too much sensitive.. need bit more smoothout :) sorry for being picky just want to help you getting it believable.

darkChief
04-11-2017, 09:35 AM
The MCLP solver is very impressive, have you tried it out?
It's meant for traditional offline simulations, and not real time. The other physics engine I'm very familiar is Physx, you would get similar results I think.

Well a coupling stiffness of 0 will completely kill all forces. So it's completely tunable.

Technically there's no collision detection between the two engines. Euphorik only sees boundary particles, not polygons. Pressure is symmetric at the boundaries, so it creates forces for Bullet. Long conversation, there's a lot literature on the topic if you're interested in the details.

I've tested the cube in a shallow fluid simulation, it will sink and rise to the top. However the rigid coupling solution I've encountered so far won't give you a simulation you can use in a realworld engineering project of course, these methods were not designed for that (maybe SPH is an exception).

The coupling stiffness also allows a lot of artistic freedom.

darkChief
04-15-2017, 11:08 AM
https://www.youtube.com/watch?v=I1wmkQWbJAY

Farhad_azer
04-15-2017, 01:21 PM
Wow.
This seems to be revolutionary stuffs for LW.

mav3rick
04-15-2017, 01:29 PM
https://www.youtube.com/watch?v=I1wmkQWbJAY

nice one dark...
as i asked previously on other plugin thread :) is it possible to precalc fluid sim until rest state and set it as new default state .. for example... fill up glass ... set last frame of rest fluid to become default 0 frame and than recalc with ice cubes falling....

also this could also be used to fill up objects shape with fluids set it frame 0 and than simulate splash .. like poking water baloon with bullet

darkChief
04-15-2017, 01:49 PM
as i asked previously on other plugin thread :) is it possible to precalc fluid sim until rest state and set it as new default state .. for example... fill up glass ... set last frame of rest fluid to become default 0 frame and than recalc with ice cubes falling....

I just implemented this last week actually while testing rigid body stuff. You can make any cached frame into the initial state. Simple stuff. Available in the next update.

https://image.ibb.co/nk29mQ/Capture.jpg (https://imgbb.com/)


also this could also be used to fill up objects shape with fluids set it frame 0 and than simulate splash .. like poking water baloon with bullet

Soft bodies are possible, but that won't be in the next update. I will be implementing that paper later on, the foundation is already set. See the video bellow, based of the same paper, that's Bullet. Not sure about poking, could be easy, but not sure.


https://www.youtube.com/watch?v=kKU5M0cU5Ek

mav3rick
04-15-2017, 02:17 PM
niiiceeee :)

jwiede
04-16-2017, 05:02 PM
Mac version pretty please :thumbsup:

Agreed!

jwiede
04-16-2017, 05:08 PM
And you asked about floating objects, check out the clip. The surface area also affects how things float or sink.


https://www.youtube.com/watch?v=W7H-yOQlRE4

There's something odd going on with the "floating block" closer to the center of the screen, note how its bottom actually rises up above the water for a while (balancing on a corner, it appears) before making contact with liquid surface -- no matter how light it is, it should be maximizing area of contact w.r.t. displacement (due to surface tension, etc.). "Balancing on a corner" in the fluid with most of a side parallel to the liquid surface shouldn't be a sustainable position (beyond the initial contact, anyway, and in this case it continued well after initial contact). The whole side should immediately come to rest in the liquid, but instead there's a visible gap for a bit between the rest of the side and the liquid surface even after rebound.

Also, not sure what density you've set for those cubes, but even styrofoam would displace more than that. They appear to be literally resting on the liquid surface without displacing any liquid. Are they set to zero density (or unrealistically low value)? That might also explain the other issues as well, if so.

(edit) Hmm, are you accounting for surface tension interactions between liquid boundary and Bullet objects? If not, might explain, and if so, might want to double-check that part.

darkChief
04-16-2017, 06:11 PM
That's the way the coupling model is. If the objects are really light, they will hardly displace the liquid. Even in the technical paper videos, light objects barley displace the liquid. And the coupling stiffness is very low in the video. It's not aimed to be a completely physically accurate model, but it is a step up from previous methods (direct forcing), because it uses boundary particles, and symmetric forces. Very tunable artistically though, you can get the look you want.

Don't expect to use real world units.

darkChief
05-04-2017, 01:49 AM
https://www.youtube.com/watch?v=8eIQfvrbgU4

darkChief
05-04-2017, 01:30 PM
https://www.youtube.com/watch?v=3y7Qs0zT9yA

mav3rick
05-04-2017, 03:01 PM
nice nice!