Oblivion - Alternate Ending - WIP - LW 2020

Planeguy

Member
I would suspect you never really cleared cache properly after one simulation setting, and then changed and rendered out something you thought would be the new setting, while it just rendered what was the old cached simulation.

Just guessing on that.
What I did was this, I went to the folder where the VDB cache is, and deleted all the files. Is there another folder for cache, or just one?
 

prometheus

REBORN
What I did was this, I went to the folder where the VDB cache is, and deleted all the files. Is there another folder for cache, or just one?
But why don´t you clear the cache from within the vdb tools/gas solver?..and the last tab is Cache where you have cache path and clear cache.
 
CDsgNgH.png
 

prometheus

REBORN
Ok, I think this looks good enough guys. I'm done tweaking this thing! Im having VDB nightmares now.


I don´t think there´s a good way to slow down simulation other than what erikals said about Frame rate, not within the simulation, feels the gas solver and vdb Is too limited there.

For TFD ..as I recall I think there was a time remapping, but I can´t recall for sure.

In blender you have a time scale adjustment field to change simulation speed.

fluids time scale.jpg

here.
 

Planeguy

Member
I don´t think there´s a good way to slow down simulation other than what erikals said about Frame rate, not within the simulation, feels the gas solver and vdb Is too limited there.

For TFD ..as I recall I think there was a time remapping, but I can´t recall for sure.

In blender you have a time scale adjustment field to change simulation speed.

View attachment 150637
here.
Nice, that's exactly what I needed, a time remap or "slow it down" button. I'm starting to see the limitations you were talking about in earlier replies regarding LW's VDB system. It shouldn't require so many steps to reduce the speed. Oh well, however the latest sim seems to work ok for me. For future sims I might have to start re-studying Turbulence or give blender's a try.
 

prometheus

REBORN
Nice, that's exactly what I needed, a time remap or "slow it down" button. I'm starting to see the limitations you were talking about in earlier replies regarding LW's VDB system. It shouldn't require so many steps to reduce the speed. Oh well, however the latest sim seems to work ok for me. For future sims I might have to start re-studying Turbulence or give blender's a try.

by default time scale is 1, you would have to lower scale to slow down the simulation in blender, and of course that means you need to add more frames to simulate the whole fluid sequence.
 
Yes, TurbulenceFD has time remapping.

But back to Lightwave OpenVDBs:
What scale is your scene? The simulation is of course size-dependent.

VDB_Test_005.gif


The left simulation is 5 times larger than the right one. I then enlarged the finished simulation 5 times to make both the same size. Both have the same settings and the same number of voxels (these are also 5 times larger in the left simulation). So apart from other settings, the size has a key effect on the speed of the smoke.

One more note on the previous discourse about the time it takes to set up a fluid simulation in different software:
In my experience, these differences don't really matter in practical work for real projects. It doesn't matter if I need 10 seconds or 1 minute to set up the simulation once.
After all, I spent hours or days on a simulation to get everything right.

For playing around and for newcomers, a preset is of course great, even if you have to change almost all settings x times afterwards anyway. And it can't hurt. So: OD Tool offers some solutions to speed up setup for Lightwaves OpenVDBs.

ciao
Thomas
 

prometheus

REBORN
Yes, TurbulenceFD has time remapping.

But back to Lightwave OpenVDBs:
What scale is your scene? The simulation is of course size-dependent.

View attachment 150641

The left simulation is 5 times larger than the right one. I then enlarged the finished simulation 5 times to make both the same size. Both have the same settings and the same number of voxels (these are also 5 times larger in the left simulation). So apart from other settings, the size has a key effect on the speed of the smoke.

One more note on the previous discourse about the time it takes to set up a fluid simulation in different software:
In my experience, these differences don't really matter in practical work for real projects. It doesn't matter if I need 10 seconds or 1 minute to set up the simulation once.
After all, I spent hours or days on a simulation to get everything right.

For playing around and for newcomers, a preset is of course great, even if you have to change almost all settings x times afterwards anyway. And it can't hurt. So: OD Tool offers some solutions to speed up setup for Lightwaves OpenVDBs.

ciao
Thomas

I can agree about most of that, apart from I think it would differ individually and what you do exactly for performing fluid simulations, and how that project is looking, wether or not you care about proper display of voxels in opengl, how easy it is to set up and paint weight maps for, how easy it is to set forces etc.

As for the simulation comparative and changing speed, that is not exactly the same in fluid behaviour, which is exactly what you get when changing simulation speed with time scaling, while it is similar, the actual sim is still looking different in behavior.

Thanks for confirming the time scale option in TFD, and as such I would recommend planeguy to stop messing with current Native gas solver (til it matures) and instead utilize what he already has, since it is faster with gpu to some level, and can time scale it and much much more.

What is it exactly in the VDB tools gas solver that brings something special to the table for you, which TFD cant?
 
Last edited:
I can agree about most of that, apart from I think it would differ individually and what you do exactly for performing fluid simulations, and how that project is looking, wether or not you care about proper display of voxels in opengl, how easy it is to set up and paint weight maps for, how easy it is to set forces etc.

I wasn't talking about the basic operation of each fluid simulation tool but the basic setup. I'm not a big fan of node based solutions unless it really brings more than a traditional GUI.
For fluid simulation I prefer non-node based applications in terms of usability. Lightwave Gas Solver is the easiest of the three applications mentioned. But that also means that there are not as many possibilities as in Blender or TFD.

As for the simulation comparative and changing speed, that is not exactly the same in fluid behaviour, which is exactly what you get when changing simulation speed with time scaling, while it is similar, the actual sim is still looking different in behavior....

My question is precisely aimed at the fact that any fluid simulation is scale-dependent. It is very difficult to make a 20 cm explosion look like a 100 meter explosion.
We're talking about simulations that try to recreate reality. In doing so, we are almost always in a conflict between realistic and freely designed.
But, of course, it's not the same as changing timeing.

What is it exactly in the VDB tools gas solver that brings something special to the table for you, which TFD cant?

Nothing.
Planeguy uses the Lightwave Gas Solver and I tried to give him some hints and tips.
I prefer to use TFD for such tasks, because it is much more powerful.

ciao
Thomas
 

prometheus

REBORN
Nothing.
Planeguy uses the Lightwave Gas Solver and I tried to give him some hints and tips.
I prefer to use TFD for such tasks, because it is much more powerful.

ciao
Thomas

Well he uses the gas solver right now, but he has TFD..so why not use that? apart from an interest to learn the native new system that is, which is understandable but it may get in the way of what you can have as finished output if you have other options, TFD I don´t have any purchased license for...so different for me.

As for gas solver and TFD, understood as for what it brings to you.
In my case, I see perhaps mostly one funcion of using particles as velocity force, I am not certain on this, but I do not think TFD has that option?
I am not talking about particles as fluid emission force that also accounts for velocity ..that is sort of included, but can´t be separated as the gas solver does by using velocity force from particles exclusively, which I showcased here on one video.
Don´t think blender has that option either, the result is better in a way since there isn´t any fluid emission taking place based on a particle size and thus any blobs can be avoided.
 

prometheus

REBORN
I wasn't talking about the basic operation of each fluid simulation tool but the basic setup. I'm not a big fan of node based solutions unless it really brings more than a traditional GUI.
For fluid simulation I prefer non-node based applications in terms of usability. Lightwave Gas Solver is the easiest of the three applications mentioned. But that also means that there are not as many possibilities as in Blender or TFD.

My question is precisely aimed at the fact that any fluid simulation is scale-dependent. It is very difficult to make a 20 cm explosion look like a 100 meter explosion.
We're talking about simulations that try to recreate reality. In doing so, we are almost always in a conflict between realistic and freely designed.
But, of course, it's not the same as changing timeing.



ciao
Thomas
Who said anything about realistic?
I thought it was about getting the sim to look as you want it to look like, and not specificly realistic if that isn´t the aim that is, and as such I have nothing to argue against what you say really.
The rest of the argument, well ..I think we are agreeing but not following eachother fully here.
 

Planeguy

Member
by default time scale is 1, you would have to lower scale to slow down the simulation in blender, and of course that means you need to add more frames to simulate the whole fluid sequence.
But this sounds so much easier to do than LW. I wish they would keep developing this. I lke having it all in one system instead of using third party software.
 

Planeguy

Member
Yes, TurbulenceFD has time remapping.

But back to Lightwave OpenVDBs:
What scale is your scene? The simulation is of course size-dependent.

View attachment 150641

The left simulation is 5 times larger than the right one. I then enlarged the finished simulation 5 times to make both the same size. Both have the same settings and the same number of voxels (these are also 5 times larger in the left simulation). So apart from other settings, the size has a key effect on the speed of the smoke.

One more note on the previous discourse about the time it takes to set up a fluid simulation in different software:
In my experience, these differences don't really matter in practical work for real projects. It doesn't matter if I need 10 seconds or 1 minute to set up the simulation once.
After all, I spent hours or days on a simulation to get everything right.

For playing around and for newcomers, a preset is of course great, even if you have to change almost all settings x times afterwards anyway. And it can't hurt. So: OD Tool offers some solutions to speed up setup for Lightwaves OpenVDBs.

ciao
Thomas
Thank you! This is a big tip I didn't know. I don't remember the size of my scene, and I already got something that works for this animation. So, if I understood correctly, if I change the size of my scene can I get a slower simulation? Is this correct? What size was your scene in this animation?
 
Last edited:

Planeguy

Member
Well he uses the gas solver right now, but he has TFD..so why not use that? apart from an interest to learn the native new system that is, which is understandable but it may get in the way of what you can have as finished output if you have other options, TFD I don´t have any purchased license for...so different for me.

As for gas solver and TFD, understood as for what it brings to you.
In my case, I see perhaps mostly one funcion of using particles as velocity force, I am not certain on this, but I do not think TFD has that option?
I am not talking about particles as fluid emission force that also accounts for velocity ..that is sort of included, but can´t be separated as the gas solver does by using velocity force from particles exclusively, which I showcased here on one video.
Don´t think blender has that option either, the result is better in a way since there isn´t any fluid emission taking place based on a particle size and thus any blobs can be avoided.
I'm planning to give it a try for future animations if I can't work with LW's native vdb system. But for now, the sim I got works fine for the animation. I'm just ready to move on to more exciting and adventurous things...such as rendering rocks and skies. ;)

Check this animation I did years ago with TFD. It wasn't a high res simulation though and I was in the early stages of learning to use it. The exhaust smoke effects were made with hypervoxels and image sequences placed as alpha transparencies in the exhaust ports. Rendered in LW10.

(ignore the amateur acting, those were kids from our church)
 

prometheus

REBORN
But this sounds so much easier to do than LW. I wish they would keep developing this. I lke having it all in one system instead of using third party software.

It is, and yes..you wish, I wish, everyone wishes..except perhaps for those at the wheel wishing for something else, I mean, it was promising ..a bit crude, had they focused on polishing that immediately, we could have had someting much more useful natively.
 

prometheus

REBORN
I'm planning to give it a try for future animations if I can't work with LW's native vdb system. But for now, the sim I got works fine for the animation. I'm just ready to move on to more exciting and adventurous things...such as rendering rocks and skies. ;)

Check this animation I did years ago with TFD. It wasn't a high res simulation though and I was in the early stages of learning to use it. The exhaust smoke effects were made with hypervoxels and image sequences placed as alpha transparencies in the exhaust ports. Rendered in LW10.

(ignore the amateur acting, those were kids from our church)

Yes, I can see it was a bit too low in resultion and turbulence undulations, it got a bit better in the end but it also lacked temperature ignition I would say.
Always a bit of a fun to see amateur acting and such..what to expect with a low low budget :)
 

Planeguy

Member
Yes, TurbulenceFD has time remapping.

But back to Lightwave OpenVDBs:
What scale is your scene? The simulation is of course size-dependent.

View attachment 150641

The left simulation is 5 times larger than the right one. I then enlarged the finished simulation 5 times to make both the same size. Both have the same settings and the same number of voxels (these are also 5 times larger in the left simulation). So apart from other settings, the size has a key effect on the speed of the smoke.

One more note on the previous discourse about the time it takes to set up a fluid simulation in different software:
In my experience, these differences don't really matter in practical work for real projects. It doesn't matter if I need 10 seconds or 1 minute to set up the simulation once.
After all, I spent hours or days on a simulation to get everything right.

For playing around and for newcomers, a preset is of course great, even if you have to change almost all settings x times afterwards anyway. And it can't hurt. So: OD Tool offers some solutions to speed up setup for Lightwaves OpenVDBs.

ciao
Thomas
So I gave this a go, and went and resized the grid size of my scene. Cleared the cache of the simulation. The fluid rate is still the same speed. What am I doing wrong?
 

Planeguy

Member
Yes, I can see it was a bit too low in resultion and turbulence undulations, it got a bit better in the end but it also lacked temperature ignition I would say.
Always a bit of a fun to see amateur acting and such..what to expect with a low low budget :)
Thanks! I completely forgot what settings I used in TFD for this scene, but it definitively needed tuning. We were on a very tight schedule, and a very, very tight budget. Had to finish one episode each month.
 
Top Bottom