PDA

View Full Version : DStorm unleashes Liquid Dynamics PhysX Plugin for LW9



Wickster
07-09-2007, 11:35 PM
Sorry I mean to post it here and not on the General Discussions page.



:EEK: OH MY GOSH...somebody has to bring this here.

Japanese Page:
http://www.dstorm.co.jp/products/plugins/liquidPack/

Babelfish Translation:
http://babelfish.altavista.com/babel...FliquidPack%2F

By the looks of it, it even uses an optional Agea PhysX card for faster simulation. Please bring it here! please show it SIGGRAPH!

Wickster
07-09-2007, 11:40 PM
There's a link to the demo at the bottom. But I think that demo is in Japanese though.

RedBull
07-10-2007, 02:48 AM
Very cool, nice to see Dstorm adding some LW stuff...

$240 USD for the Plugins (FluidParticle & NaminamiFX)
$455 USD for the Liquid Pack (2Plugins) with Hardware PhysX Processor.

It would mean LW is the first 3D software to support a Hardware PhysX implementation as XSI does not.

Seems to be 32bit only on the Demo, i'm not convinced by the renderings, but i will check out the Demo... I didn't think PhysX was really well suited to fluids as yet. Hope i'm wrong.

(YES Newtek should show it at Siggraph, i won't be there this year, but it would be good to show off some Nodal + Hardware PhysX fluids)

Dodgy
07-10-2007, 03:12 AM
Wow.... We want it for 9.5 :)

mav3rick
07-10-2007, 03:35 AM
great can't wait !!

im downloading demo...

colkai
07-10-2007, 04:25 AM
Yup, why is this not going to be in LW natively? Surely that would please a lot of folks?
From what I've seen so far, I don't need it, I really, really don't do anything that justifies buying it.

Which means of course, I want it! :p

Panikos
07-10-2007, 04:29 AM
If its similar like

http://graphics.stanford.edu/~fedkiw/

I will buy it. :thumbsup:

RedBull
07-10-2007, 06:54 AM
If its similar like

http://graphics.stanford.edu/~fedkiw/

I will buy it. :thumbsup:

LOL, no it's not CFD, this is basically just a PhysX implentation...

It seems they have geared this more at liquids, however it's really not ideal for them (as i suspected) however the plugin could be quite useful for faking CFD.

The main problem i see with this is that it relies on HV, JetstreamFX or Pixie Dust, and this really will lead to unrealistic blending, and velocity stretching etc that has hobbled HV for years.

Some of the demo's are far from what Blender can already do with it's real CFD. (Cup of water etc) But i could see it being used for some softbody and particle FX...

It is quite simple to use and setup, and it didn't crash and worked with Fprime, (except the volumetrics of course) LW now has a much closer working model of XSI's particles.

Perhaps with HVDeformer, Dynamite and FluidParticle you could do some nice stuff with enough time and patience.....

colkai
07-10-2007, 07:06 AM
The main problem i see with this is that it relies on HV, JetstreamFX or Pixie Dust, and this really will lead to unrealistic blending, and velocity stretching etc that has hobbled HV for years.

That was my thought too, i'm sure the dynamics and options are great, but HV's just end up looking "wrong" unless you've got a gazillion particles and even then, I've yet to see a non "gooey/blobby" effective HV liquid in LW.

Not to say I wouldn't love to play with this, but as HV's stand, I tend to think of them as a bottleneck in it's usefulness, rightly or wrongly.

Panikos
07-10-2007, 07:08 AM
Fat blobby balls again :(

Dodgy
07-10-2007, 07:12 AM
The demo is limited to 5000 points, per emitter, with more you could probably get some really nice results. The collision seems a lot more stable than PFX, it's really easy to setup, though there's some nodes built in, be interesting to see what you can do with them... Here's a quick HV animation.

Now we need a really fast fluid renderer :)

Sensei
07-10-2007, 06:42 PM
Dodgy, try animating template where fluids flows using f.e. Morph Mixer to check how it handles such more complex examples.. :)

lino.grandi
07-10-2007, 11:57 PM
This is a test I made:

http://www.youtube.com/watch?v=64S7-uLQiC4

It's absolutely fast. Collision are very precise, and even inter-collision among several different particle systems work great!

I must say I'm impressed. Love it, love it, love it!

Titus
07-11-2007, 12:49 AM
Excellent test!

Mr Rid
07-11-2007, 01:46 AM
Yup. This is why we need a global 'Distance Between Particle' parameter in the Size and Blend in HV that I have been asking for ever since LW6 messed this up (could do it in v5.6). Please go to https://secure.newtek.com/FogBugz/default.asp?pg=pgEditBug&command=new

Click 'New Case,' select 'Lightwave Feature Request' and ask for a global Distance Between Particle parameter in the Size and Blend.

The current axis specific gradients dont work in all the channels and are too tricky to layer and are missing from Size or Blend where they would do the most good.


That was my thought too, i'm sure the dynamics and options are great, but HV's just end up looking "wrong" unless you've got a gazillion particles and even then, I've yet to see a non "gooey/blobby" effective HV liquid in LW.

Not to say I wouldn't love to play with this, but as HV's stand, I tend to think of them as a bottleneck in it's usefulness, rightly or wrongly.

scratch33
07-11-2007, 01:58 AM
wow... Bravo Lino...awesome:thumbsup:

mav3rick
07-11-2007, 03:35 AM
great :) we got fluids:) for 1.0 it is impressive

RedBull
07-11-2007, 06:03 AM
I wouldn't really call it fluids really it's not CFD...
Have a look at the splash on any NamiNamiFX and notice there isn't any on the sides of the cup, when sloshed around. It's quite unrealistic really.

I think it should of been aimed at PhysX particles, rather than fluids as such,
as it's pretty cool but just not really for fluids. :)

But yes the particles themselves will behave a little better than LW's current FX. (also done by Ino) But shading them to make them look fluidy will be the same problem as before.

Ironically XSI's particle and fluids (PhysX) as been critisized for a while, with promises they will improve on the fluids it can achieve in future versions, but at the moment it suffers the exact same problems as this....

Not enough to stop me from using Blender or RF, but still it's capable of some really nice interaction and effects overall.

creacon
07-11-2007, 08:05 AM
I'm just guessing here, but wouldn't it be possible to change the form of the HV using Denis Pontonnier's node volume plugin? And combine that with the particle info that's in DStorm's plugins.

I just don't get why this node uses a scalar as input.

creacon




I wouldn't really call it fluids really it's not CFD...
Have a look at the splash on any NamiNamiFX and notice there isn't any on the sides of the cup, when sloshed around. It's quite unrealistic really.

I think it should of been aimed at PhysX particles, rather than fluids as such,
as it's pretty cool but just not really for fluids. :)

But yes the particles themselves will behave a little better than LW's current FX. (also done by Ino) But shading them to make them look fluidy will be the same problem as before.

Ironically XSI's particle and fluids (PhysX) as been critisized for a while, with promises they will improve on the fluids it can achieve in future versions, but at the moment it suffers the exact same problems as this....

Not enough to stop me from using Blender or RF, but still it's capable of some really nice interaction and effects overall.

alifx
07-11-2007, 08:50 AM
very neat system, and nice test lino

I wish to see lw dynamics improved on the next versions.

dstorm
07-12-2007, 03:55 AM
Hello

This is D-Storm. Thank you for testing Liquid Pack trial version!

This plugin's author is INO-san who was at NewTek. His contract with NewTek expired last year and had been developing Liquid Pack by himself since at the start of this year. Thus, NewTek has no plans to build Liquid Pack in LightWave next version for now.

Currently we D-Storm has no plans to release English version because of the language problem. However, if we will receive many requests for English version, we might change our plan.

If we will release Liquid Pack English version, we will need help from you to correct documents in our poor English.

Thank you. :)

wavk
07-12-2007, 04:12 AM
hi dstorm , can we see more animated tests?

mlon

RedBull
07-12-2007, 04:23 AM
Hello

This is D-Storm. Thank you for testing Liquid Pack trial version!

This plugin's author is INO-san who was at NewTek. His contract with NewTek expired last year and had been developing Liquid Pack by himself since at the start of this year. Thus, NewTek has no plans to build Liquid Pack in LightWave next version for now.

Currently we D-Storm has no plans to release English version because of the language problem. However, if we will receive many requests for English version, we might change our plan.

If we will release Liquid Pack English version, we will need help from you to correct documents in our poor English.

Thank you. :)

Thank you for your reply....

I'm not really worried about the documentation, as it's fairly easy to use...
However are there plans for a 64bit compile of the plugin, as i would be interested if i got both 32bit and 64bit...

alifx
07-12-2007, 04:25 AM
hey dstorm :) :)
nice to see you posting here.



If we will release Liquid Pack English version, we will need help from you to correct documents in our poor English.

we have very helpful people here :)
I hope they help to make the english version.

I wish to help but my english is not that much :D

JohnMarchant
07-12-2007, 04:28 AM
Hey anything that we can do to help D-Storm, let us know

creacon
07-12-2007, 04:51 AM
Well, the interface is already in english and a lot of the functionality can be found by experimenting.

What I have seen so far is impressive.
But what is really needed is a mesh that can be generated from the particles, like realflow does.

Great job.
But does this also mean that INO-san won't be working on fixing the particle systems that are in LW now? There are still a lot of bugs in there.

creacon


Hello

This is D-Storm. Thank you for testing Liquid Pack trial version!

This plugin's author is INO-san who was at NewTek. His contract with NewTek expired last year and had been developing Liquid Pack by himself since at the start of this year. Thus, NewTek has no plans to build Liquid Pack in LightWave next version for now.

Currently we D-Storm has no plans to release English version because of the language problem. However, if we will receive many requests for English version, we might change our plan.

If we will release Liquid Pack English version, we will need help from you to correct documents in our poor English.

Thank you. :)

Wickster
07-12-2007, 09:32 AM
Hi DStorm, glad to see you joining in on the discussion. The plugin is pretty is pretty easy to use, very powerful, a lot of potential and a something that a lot of LW users here have been waiting for. Making the plugin available for purchase outside of Japan would probably benefit everyone.

About the documentation, there are probably people here who would gladly help out on it. Maybe you can give us the roughly English translated documentation and someone will probably take the time to rephrase and rewrite it. Don't forget, this is the community who stood up, gave their time and gave their best to create a community reel to help launch LightWave 9.

Besides, the plugin is pretty straightforward that we might not need a documentation. If a new user would use your plugin would need help, all they have to do is ask here in the forum and all of us try our hardest to solve even the smallest of problems.

By the way, I visit your site everyday and I think you do a great job on covering everything LightWave internationally. :thumbsup:

creacon
07-12-2007, 09:44 AM
That's what I think too.

I was able to get some pretty neat stuff out of it in a few minutes time.
Stuff that took me days in realflow.
It's so fast that it allows you to test different settings and pick the best one.

The things I can't find out:

- Is the 5000 particles a limitation for the demo only?
- Is the GRID size (wave) limited to 100 X 100 for the demo only?
- Is the particle size global? When I change 1 emitter, the other emitter's particle size change too.
- Would it be possible to use different event objects that have different effect strengths (wave)?

It's a pitty that japanese takes so long to learn.


creacon







Hi DStorm, glad to see you joining in on the discussion. The plugin is pretty is pretty easy to use, very powerful, a lot of potential and a something that a lot of LW users here have been waiting for. Making the plugin available for purchase outside of Japan would probably benefit everyone.

About the documentation, there are probably people here who would gladly help out on it. Maybe you can give us the roughly English translated documentation and someone will probably take the time to rephrase and rewrite it. Don't forget, this is the community who stood up, gave their time and gave their best to create a community reel to help launch LightWave 9.

Besides, the plugin is pretty straightforward that we might not need a documentation. If a new user would use your plugin would need help, all they have to do is ask here in the forum and all of us try our hardest to solve even the smallest of problems.

By the way, I visit your site everyday and I think you do a great job on covering everything LightWave internationally. :thumbsup:

moussepipi2000
07-12-2007, 02:46 PM
this plug is really simple to learn, we dont need any documentation how to learn !! and its powerful. the only thing we can ask to ino san is to create metaball or mesh system it can fit on the particles. nami nami fx is really good too. it can make really good results in few times. good job INO and d-storm

このプラグインの作り方はとてもかんたんです、説明書は必要ありません。強大なプラグです!イノ様メタボー ルとかメシュシステムとか作って下さい。ナミナミFXもいいね。短期間ですごい結果があります 。 ありが とうイノ様とD-STORM・がんばってください 。

i tried to do my best in japanese. sorry for mistake and more

moussepipi2000
07-12-2007, 03:40 PM
this plug is really simple to learn, we dont need any documentation how to learn !! and its powerful. the only thing we can ask to ino san is to create metaball or mesh system it can fit on the particles. nami nami fx is really good too. it can make really good results in few times. good job INO and d-storm

このプラグインの作り方はとてもかんたんです、説明書は必要ありません。強大なプラグです!イノ様メタボー ルとかメシュシステムとか作って下さい。ナミナミFXもいいね。短期間ですごい結果があります 。 ありが とうイノ様とD-STORM・がんばってください 。

i tried to do my best in japanese. sorry for mistake and more


使い方だった。ごめん

PixelFarmer
07-12-2007, 04:06 PM
I'd be interested in an English version and could contribute to the docs if there is a team effort, but barring that, can we purchase the Japanese version?

GregMalick
07-13-2007, 09:19 AM
Yup. This is why we need a global 'Distance Between Particle' parameter in the Size and Blend in HV that I have been asking for ever since LW6 messed this up (could do it in v5.6). Please go to https://secure.newtek.com/FogBugz/default.asp?pg=pgEditBug&command=new

Click 'New Case,' select 'Lightwave Feature Request' and ask for a global Distance Between Particle parameter in the Size and Blend.

The current axis specific gradients dont work in all the channels and are too tricky to layer and are missing from Size or Blend where they would do the most good.

I'm not saying that NT shouldn't do this - but check out THREAD (http://www.newtek.com/forums/showthread.php?t=70798) Post #12.

specifically:

Yes, as you can see in the node setup here: http://perso.orange.fr/dpont/plugins...tml#NodeVolume (http://perso.orange.fr/dpont/plugins/nodes/nodes/Node_Editors.html#NodeVolume)
An Hypertexture is similar to the normal displacement scalar that you can find as input in the Surface node editor. If you want to change the shape of the HyperVoxels, you could also use the Texture node editor in Particle Size Texture Layer and the Partcle Info node, but don't forget to select the emitter in the node panel.

Here is an experimental win32 Particle Info node, with a "Between" output, distance to nearest particle, except itself (could be slow).
ParticleInfoNode_x32.zip (http://vbulletin.newtek.com/attachment.php?attachmentid=48091&d=1184259932)

Denis.

zatara
07-13-2007, 10:38 AM
I think Newtek should hire again Ino-San again.
I havenīt tryed this liquid pack, but it sounds awesome.

RedBull
07-13-2007, 04:18 PM
I'm not saying that NT shouldn't do this - but check out THREAD (http://www.newtek.com/forums/showthread.php?t=70798) Post #12.

specifically:

Yeah that's really not good enough, or goes far enough to improve HV...
That also only works with Volume not surface.

HV Deformer or (Waterpool) is a much better idea of what HV actually needs..
But at the moment, HV Deformer is way to much trouble to manually setup, this process needs to be automated among other things...

JamesCurtis
07-13-2007, 04:42 PM
Dstorm,

I would definately like to see an English version of the Liquidpack plugin for purchase, and i'm sure anyone else in the Lightwave community here will agree!

Lewis
07-14-2007, 02:12 PM
Nice to hear someone from Dstorm , thanks for posting here :). I thought Daisuke Ino is still working at NT - I was wrong obviously :(.

Mr Rid
07-14-2007, 06:56 PM
Yeah that's really not good enough, or goes far enough to improve HV...
That also only works with Volume not surface.

HV Deformer or (Waterpool) is a much better idea of what HV actually needs..
But at the moment, HV Deformer is way to much trouble to manually setup, this process needs to be automated among other things...

HV Deformer may as well not exist for all the practical use it is. I've only seen it used effectively for a still image. I could not get even a simple splash of liquid landing flat on a floor, let alone pouring over an object. Anyone have a good example of HVdeform being used for anything convincing?

I get tired of people pointing out little tidbit examples that work only in innocuous tests but not in practical reality. Its a waste of time to explore unless you can show how it works in a convincing example, and not just a couple of balls doing something.

A proper Distance Between Particle parameter would work wonders, particularly in the Blend of voxels. And the data is already there. Somebody just attach it to a gradient or node and we are in business.

loki74
07-15-2007, 10:00 PM
HV Deformer may as well not exist for all the practical use it is. I've only seen it used effectively for a still image. I could not get even a simple splash of liquid landing flat on a floor, let alone pouring over an object. Anyone have a good example of HVdeform being used for anything convincing?

I get tired of people pointing out little tidbit examples that work only in innocuous tests but not in practical reality. Its a waste of time to explore unless you can show how it works in a convincing example, and not just a couple of balls doing something.

A proper Distance Between Particle parameter would work wonders, particularly in the Blend of voxels. And the data is already there. Somebody just attach it to a gradient or node and we are in business.

Well, if the results you got making fire with LW's native tools (that was you, right?) is any indication, you'll probably be the first to get it right with HVs.

I'd rather see a level set based isosourface. Eulerian is WAY better for liquids (fluids in general I think)... at least in laminar cases, that is.

I nearly shat myself when i saw this thread... until I scrolled down and saw RedBull's post, "not CFD" ;_;

Anti-Distinctly
07-17-2007, 04:33 PM
The level set isosurface seems to be the trickiest part of getting a decent fluid sim. Wouldn't that have to be a seperate object per frame of animation too?

HowardM
07-18-2007, 04:38 AM
anyone got more vids of LiquidPack?
:)

lino.grandi
07-18-2007, 02:43 PM
I made some other videos...this plugin really ROCKS!

http://www.youtube.com/watch?v=Irc2gxNfwU8

http://www.youtube.com/watch?v=BfUn2vP0ucA

One more to came!

JamesCurtis
07-18-2007, 09:43 PM
Could anyone help me with the install of the demo plugin steps since the install is in japanese. Please list steps!!

Dodgy
07-19-2007, 03:17 AM
Just keep clicking the default selected button, that's all I did :)

dballesg
07-19-2007, 11:42 AM
Hi,

I downloaded the demo and what impressed me most was the speed of calculation. For the demo scenes looks like the collisions are better handled by this plugin than the original ones programmed by the same guy for lightwave???.

Anyhow, as other pointed here, a particle based fluid system needs of an isosurface generator.

I think Blender fluids solved that quite well! :) Except their algorithm takes a lot of memory, more than a SPH system I think.

If someone can confirm if in Layout would be possible to generate and SAVE dynamic geometry (AntiDistincly is asking as well on other thread) it would be possible to port Blender fluids right?

Best regards,
David

mav3rick
07-19-2007, 11:57 AM
i think we are very limited to liquid pack and it support since it is targeted to japanees market.... i dont see why dstorm decide not to release english version?
is japanees LW market that bigger than rest of world or ?

loki74
07-19-2007, 06:35 PM
The level set isosurface seems to be the trickiest part of getting a decent fluid sim. Wouldn't that have to be a seperate object per frame of animation too?

Well, if you wanted to save/export the simulation, yes, you would have to have a mesh object for every frame, as well as morph data if you wanted motion blur. The process would be very much like that in RealFlow.

However, as far as mesh generation in general, that's not the hugest problem. I mean, HV's or Level Set, a mesh must be generated, right? Although I believe it is possible to raytrace a level set directly, (trace the ray thru the field until phi=0, then take the gradient to get the surface normal) I imagine it would be noisy or "pop"...

Anyway, mesh generation really isn't the hardest part of making a fluid simulation--there's a lot of documentation on marching cubes method, I imagine LW already does it someplace or other... The hard part is solving the NSEs themselves.

dstorm
07-20-2007, 05:05 AM
Hello, this is D-Storm.

Thank you for many requests for "Liquid Pack" ! :)

In Japan, we released Liquid Pack today. And, we decided to release English version in near future thanks to your requests. However, I am sorry that I can not tell you exact release schedule because we do not have enough time to prepare it while working on other many tasks. So, please enjoy Liquid Pack trial version for a while.

The limitation of Liquid Pack trial version is as following:

-Expiration date: at the end of August 2007
-Could not save motion data.
-GridSize in NaminamiFX is no more than 100.
-Particle number in FluidParticle is no more than 5000.
-Could not render via network.

Thank you.

moussepipi2000
07-20-2007, 05:56 AM
liquid pack is very fast and great but i cant find any parameter about mesh adherence, for example a scene with emitter and a simple ball under this emitter. if the liquid fall on the ball, it have to splash of course but it have to follow the shape of the ball until the lower hemisphere of this ball too but it doesnt... anyone found this parameter or it simply not yet able to do this kind of this (very important for liquid like JAM, or high viscosity liquid)

thank

RedBull
07-20-2007, 08:16 AM
Hello, this is D-Storm.

Thank you for many requests for "Liquid Pack" ! :)
In Japan, we released Liquid Pack today. And, we decided to release English version in near future thanks to your requests.
Thank you.

Thankyou Great to hear about English version...
Any chance for a 64bit version?

Lightwolf
07-20-2007, 08:41 AM
Any chance for a 64bit version?
I suppose that depends on the availability of a 64bit PhysX SDK, last time I checked (this week) there is no 64bit version available (but planned, with no release date). An OSX version _might_ come some time in the future.

Cheers,
Mike - being nosy again.

nemac4
07-20-2007, 09:31 AM
Well, if you wanted to save/export the simulation, yes, you would have to have a mesh object for every frame, as well as morph data if you wanted motion blur. The process would be very much like that in RealFlow.

However, as far as mesh generation in general, that's not the hugest problem. I mean, HV's or Level Set, a mesh must be generated, right? Although I believe it is possible to raytrace a level set directly, (trace the ray thru the field until phi=0, then take the gradient to get the surface normal) I imagine it would be noisy or "pop"...

Anyway, mesh generation really isn't the hardest part of making a fluid simulation--there's a lot of documentation on marching cubes method, I imagine LW already does it someplace or other... The hard part is solving the NSEs themselves.

I wonder if there could be a way to convert the particles into metaball groups.

RedBull
07-20-2007, 07:08 PM
I suppose that depends on the availability of a 64bit PhysX SDK, last time I checked (this week) there is no 64bit version available (but planned, with no release date). An OSX version _might_ come some time in the future.

Cheers,
Mike - being nosy again.

Cheers, Yeah i know the OSX is absent, but i also checked the other day, and thought i read 64bit support, although 64bit support doesn't always mean native 64bit libraries i guess.

However XSI also uses PhysX, and there are 64bit plugins for Maya and Max.
So not really sure...(I'm trying not to buy any more 32bit plugins.)

And the hardware PPU's also have 64bit drivers....?

Lightwolf
07-21-2007, 02:58 AM
Cheers, Yeah i know the OSX is absent, but i also checked the other day, and thought i read 64bit support, although 64bit support doesn't always mean native 64bit libraries i guess.

Officialy you get driver for 32bit apps on a 64bit OS... but no 64bit libraries.


However XSI also uses PhysX, and there are 64bit plugins for Maya and Max.
So not really sure...(I'm trying not to buy any more 32bit plugins.)

These are ports by the developers, which have also paid for a license which includes source code access.
I suppose for that to be worthwhile for Ino to consider, he'd need to 10x the price... ;)

And the hardware PPU's also have 64bit drivers....?
Apparently yes.

However, I wonder how much sense they still make in the age of quad cores and GPU bases calculations...

Cheers,
Mike

Cheers,
Mike

archijam
07-21-2007, 03:03 AM
Cheers.
Mike.

Is anyone aware of the suitability of this or any other plugin for sumulating realistic river flow (and perphaps overflow, flooding etc) ?

Would not need to be realistic close up, but from a birds eye view ...

Cheers,
James

JamesCurtis
07-21-2007, 09:12 PM
Has anyone reading these threads done any LW example scenes they could share? I've no clue how to use the demo.

nemac4
07-22-2007, 12:58 AM
Here is a test render.

prospector
07-22-2007, 01:04 AM
looks like LW normally does

What special thing am I looking for?

nemac4
07-22-2007, 01:06 AM
It seems better to me. I think it would work well for stuff like sand.

Mr Rid
07-22-2007, 01:43 AM
looks like LW normally does

What special thing am I looking for?

I still can not find this liquid pack thing. The previous links, and any online search only lead to pages in Japanese, and I can find no info about it on the English version of DStorm's site(?).

Mmm, yeah the pouring liquid behavior Ive seen so far is no different from what I can get out of LW, but it sounds like it calculates much faster. LW is not really practical for liquids because of how slow the self-interaction is.

This brief sim of 20,000 self-colliding particles took three hours to calculate in LW9, which is not nearly enough to really do something serious. http://www.box.net/shared/static/ytqofua2te.mov When I tried 30,000 points, the calculation froze on frame one forever so I gave up the thought of liquid in PFX.

I still think liquid is best left to RealFlow. There is a free, full featured demo available. A 'Register Now' nag pops up every so many minutes and kills any currently rendering flow, but use autosave to restart.

cagey5
07-22-2007, 01:54 AM
The link for the lliquidpack demo is the last link on that first page.

http://www.dstorm.co.jp/products/plugins/liquidPack/LiquidPack92JDemo.zip

Trevor
07-22-2007, 02:13 AM
This brief sim of 20,000 self-colliding particles took three hours to calculate in LW9, which is not nearly enough to really do something serious. http://www.box.net/shared/static/ytqofua2te.mov When I tried 30,000 points, the calculation froze on frame one forever so I gave up the thought of liquid in PFX.

.

Yikes, 3 hours.!
This little test I did did with dstorms pfx took only 3 mins, for a 230 frame simulation, You have to admit , thats pretty impressive, on a single 3 gig processor.
By my calculations, 5000 points per emitter, I cloned it 17 times, thats 85,00 points.
http://www.ogc-clan.com/images/screens/fluidfxp.avi
Divx codec 3 megs.

lino.grandi
07-22-2007, 02:41 AM
Really, really cool!

Trevor
07-22-2007, 02:56 AM
Oops, I meant 85,000 points, not 8500.
Im sure there used to be an edit post tool, :stumped:

Lewis
07-22-2007, 03:23 AM
Yikes, 3 hours.!
This little test I did did with dstorms pfx took only 3 mins, for a 230 frame simulation, You have to admit , thats pretty impressive, on a single 3 gig processor.
By my calculations, 5000 points per emitter, I cloned it 17 times, thats 85,00 points.
http://www.ogc-clan.com/images/screens/fluidfxp.avi
Divx codec 3 megs.

Yes Trevor this really look very cool. can you render it wiht surfaces so that we see is it gouing to look nice then :)?

dballesg
07-22-2007, 03:25 AM
Has anyone reading these threads done any LW example scenes they could share? I've no clue how to use the demo.

Hi, The demo include several scenes to try.

Its use is very similar to Particle FX.

Create a NUll, and add PX_FluidDemo on the Custom Object plugins section.
Open the plugin option and change parameters there! :)

BTW I saw and option called Use Grand Plane. My test shows that create an invisible "ground" collision plane, so here we go again with a misspelled option on one of Mr Ino plugins, looks like it is his trademark! :) So that option means really "Use Ground Plane" ;)

The collisions works the same but the plugin is PX_Block.

Here it is a test make in two minutes, so do not expect anything scpectacular! :)

Mr Rid,

The speed of calcuation and puring effect are way different of the usual Particle FX, this particles behave much more warter, but as you pointed, without another plugin able to generate a mesh around them instead of using HV the results "looks like" normal Particle FX.

Realflow it is another beast that has been developed on the last 10 years, there is no comparison possible.

For example I HATE the 3D Max interface of Realflow.

I remember I had a conversation with Victor on their office when they shoiw me Realflow 1.0. And I told them that it was sad that they choose a 3D Max style interface instead of one similar to LW.

I said that because they developed the Realflow plugins for Lightwave on first place! :) Maya, Max, and XSI support came later! :)

Best regards,
David

Trevor
07-22-2007, 04:50 AM
Yes Trevor this really look very cool. can you render it wiht surfaces so that we see is it gouing to look nice then :)?

Well, the only solution within LW is still hypervoxels, cant get away from that blobby look no matter how many particles you use.
Someone over at spinquad suggested using flowtracer instead, I did try that, but the renders had awful artifacts, in not sure if its because its in demo mode or not.
Render times on that plugin were really long too, so I gave up on that solution.

The ability to save out a generated mesh sequence, like realflow would be awsome, but we can only hope and pray on that I guess.

On another note my only fault with the dstorm plugin is that on the mesh collision there are no settings to tweak, you just plug it onto the mesh you want your particles to collide with and thats it.

No matter how much I tweaked the particle settings, I could not get it to flow on a flat surface, like water, not fast enough.
I kept ending up with oil like fluid.

nemac4
07-22-2007, 05:22 AM
Here is another test render.

Gregg "T.Rex"
07-22-2007, 06:00 AM
LiquidFX seems like an decent effort for some fluid particle behavior and the Naminami part of the plugin is even cooler for ocean ripples and waves, with object interactions. But still, we are stuck in a program that can't render realistic fluid simulations. Real Flow's meshes is the only vital option we have right now, since Next Limit's Flow Tracer plugin that could render out Real Flow's or LightWave's particles realistically, broke at 9.0, afaik...

Anyway, i'll stick with Real Flow for now; LiquidFX is cool and all, but imho very limited for true and various types of liquids that keep their volume throughout the simulation. For comparison, here's some Real Flow work, for a not-aired-yet tvc job i'm currently working on. First one is a LightWave render of a 500K polygon Real Flow mesh and the second is a Real Flow calculation of almost 1 million particles in a container. Calculations took about one hour for the first sim and 12 hours for the second...

1st: LightWave 500K mesh render (http://www.sub-pixel.com/RFLW1.mov)
2nd: Real Flow 1M particles simulation (http://www.sub-pixel.com/RFSim.mov)

Though, the price DStorm is asking for the LiquidFX pack is extremely attractive and almost a give away, for an alternative option to umber expensive Real Flow, even if the options are lesser or not as advanced...
A welcome addition to our tool set, nevertheless...

Cheers,
T.Rex

Lewis
07-22-2007, 09:51 AM
No matter how much I tweaked the particle settings, I could not get it to flow on a flat surface, like water, not fast enough.
I kept ending up with oil like fluid.

Hmm so basically with this fluid simulator you still can't really render WATER :(. Maybe it would be nice to get some renders from Mr. Ino himself to show us is it possible or not ?

Castius
07-22-2007, 11:58 AM
I was surprised to see so few setting liquidFX had. Then i realized that is how it has to be if you want real control though nodes. To change the way the water behaves you do this though the node editor. What you do to speed up fluids is detect the collision and modifier there behavior. You can make then bounce more or increase there speed. Combining a liquid simulator and a node based editor is simply amazing. The amount of control you can have over this is simply insane.

Ernest
07-22-2007, 12:44 PM
Has anyone tested this with and without a PhysX card? Is the speed radically different? Does the simulation look any different from the CPU version?

nemac4
07-22-2007, 01:53 PM
Here is yet another test using item info from the ball to drive the external force in nodes. No physX card here.

Dodgy
07-22-2007, 03:22 PM
That looks really nice :)

Faster hypervoxels, kill kill!

Trevor
07-22-2007, 04:14 PM
What you do to speed up fluids is detect the collision and modifier there behavior. You can make then bounce more or increase there speed. Combining a liquid simulator and a node based editor is simply amazing. The amount of control you can have over this is simply insane.

Thanks Castius, I dio not know that.

What I did try is the demo in the control panel for phsx, even in software mode
its very impressive, fun too, shoot the boxes in real time and watch them interact.
Seems to me, this could blow away LW's hard fx, if implemented as hard fx as well as its current fluid sim.

duke
07-22-2007, 08:07 PM
Is this multi-threaded?

Gregg "T.Rex"
07-23-2007, 12:33 PM
Is this multi-threaded?
No, it is not...
But it's fast, even with one core...

T.Rex

archijam
07-24-2007, 02:54 AM
Speaking of PhysX, has everyone seen this realtime demo (http://www.ageia.com/physx/tech_demos.html)? (at bottom).

Anyone familiar with the limitations of this, tech, i.e. at what scale it slows down?

Looks quite similar in effect to the liquid pack, tho the way it also moves the objects in its path is ... sweet.

j.

mav3rick
07-24-2007, 10:19 AM
http://www.youtube.com/watch?v=fXkSxph75_I&mode=related&search=


dstorm how about update on cloth fx....... and hard fx with ageia?
from this examples i have to say ageia got pretty nice examples

Wickster
07-24-2007, 12:24 PM
http://www.youtube.com/watch?v=fXkSxph75_I&mode=related&search=


dstorm how about update on cloth fx....... and hard fx with ageia?
from this examples i have to say ageia got pretty nice examples
Maybe LiquidPack is the first stage. :)

Here's hoping.

archijam
07-24-2007, 01:17 PM
By the way, I meant to say good work on LiquidPack :)

Now, since we've all gone and requested distance between particles, we can expect even more from the next version ;)

j.

Gregg "T.Rex"
07-24-2007, 01:34 PM
Well....
Let's put costs aside for a while...

Although, i most certainly will buy Liquid Pack and put it in good use, i don't understand why we have to settle with this solution for liquid sims inside LightWave. I mean, is this the LW's future for liquid and other dynamic sims, we all wanted for so many years? Feature-wise comparison with Maya's Nucleus (http://area.autodesk.com/custom/?id=4547) or Real Flow (http://www.nextlimit.com/realflow/), Liquid Pack or any similar upcoming dynamic solver, looks like once again a somewhat "limited" solution. Maybe, it's LightWave itself that make any solution look limited, or maybe i'm wrong as all i've seen is the demo of Liquid Pack, but comparing apples to apples, Nucleus (http://area.autodesk.com/custom/?id=4547) and Real Flow (http://www.nextlimit.com/realflow/) look way more advanced...

For once, i would like to see NewTek step up to the challenge and either contact big brains to code next-gen dynamic solutions, like Nucleus, or buy rights to use next-gen dynamic engines, like Ageia, Havoc etc. If more money are required for this task, then do it as an extra add-on Dynamics pack for LightWave and charge whatever you see fit. I'm SO tired of waiting for decent solutions from 3rd party developers like Ino or Worley. I want NewTek to deliver the goods internally, one way or another and in any way possible, regardless of costs or expenses...

And for once and for all, i would like to see LightWave's road map. What plans NewTek has for LightWave? Where does NewTek want to take LightWave? And i don't mean corporate blah-blah, but a simple and clear road map without dates or anything. Just what NewTek plans to do with LightWave. I don't care if they go there in ten years or never accomplish them, but what they have in mind right now, like today...

I'm with LightWave since 1993 and as a devoted user, i ask for some inspiration from NewTek's end to keep things rolling. If we expect and ask too much from NewTek, then i'd also like to know...

Thanks,
T.Rex

Dodgy
07-24-2007, 02:10 PM
I don't know. The fluid stuff in Nucleus in the pics looks like a fluid/HV combo, which is what we have. I've used the soft body stuff in maya, and if you get into cm scales, the values become very twitchy and despite what anyone says the collision system for cloth is exactly the same point poly collision we have in cloth fx, and suffers from the same problems. I've actually found it easier to get decent cloth and other soft bodies effects in LW than in maya's spring based system, as clothFX has more values you can play with. HardFX is where LW falls down a bit, with it's inability to do proper self collision, and lack of constraints. If those could be added in, it would be a lot stronger.

Also, if nodes could be used to set values, or even clothFX and other FX noded within deformations, LW's dynamics would be very good indeed. XSI's system is similar to clothFX, with a lot of the same settings, but they can link them to any driving value, not just a weight map. This would really be a boost to LW's dynamics.

mav3rick
07-24-2007, 03:11 PM
well l after i saw fluids(not talk about namini) all i can tell it can be put in good use.... i dont die about MATH precise fluids and i dont see reson to go that far with some dynamic..... it has to look convincing and not math precise....
but than again i am user that do tv commercials and i can do most of things without using any HARD BRAIN coded stuff..... for now i would like hv3 update or dynamite update so we can get better looking renders of fluids.....

Gregg "T.Rex"
07-24-2007, 03:44 PM
The fluid stuff in Nucleus in the pics looks like a fluid/HV combo, which is what we have.

Yeah....
Maya's fluids rendering sucks as much as HVs for LightWave. But, we don't compare liquid rendering engines, rather liquid dynamic solver engines and Nucleus fluids can be combined easily with cloth, soft and rigid dynamics, all together with tons of options to set up right in front of you, instead of hidden behind a node layout, with hard to decipher connections...

Again, my question remains; why NewTek does/say nothing for such tasks and instead "wait" for a 3rd party developer to come up with a solution, by his own initiative?

Dodgy
07-24-2007, 04:51 PM
Nucleus and Real Flow look way more advanced


But, we don't compare liquid rendering engines, rather liquid dynamic solver engines and Nucleus fluids can be combined easily with cloth, soft and rigid dynamics, all together with tons of options to set up right in front of you, instead of hidden behind a node layout, with hard to decipher connections

Have you actually used Nucleus, or just seen the demos they've chosen to show you? I'm confused as to your experience. I'm going on what I've seen, and Nucleus to me as they've stated is a particle based system. And believe it or not, HardFX and Clothfx and ParticleFX are all particle based systems. This means you can have hard objects sitting in cloth objects in LW, which until nucleus you couldn't in maya. ParticleFX particles can deform and rest on ClothFX objects, and ClothFX objects can rest over Hardfx objects. The one outsider is SoftFX, which is kind of integrated, but not as much, and you can do most of SoftFX effects Using ClothFX, so there's overlap, but if you want to use SoftFX for a specific type of effect you can do that too.

Maya is entirely node based, so nucleus will be the same, so I fail to see how making LW's dynamics nodal need have any harder to decipher connections than a maya system. Maya has script interfaces to its dynamics, (I assume, since most things in maya are scriptable) which is why LW would be well served to add scriptability to the whole package, since that is maya's strength. I've already written a plugin to autosew ClothFX objects, that's a step towards a nucleus type usability, but it does have the advantage of just being written all together, tossing out the old code, so as to more neatly fit together. It's very nicely interactive too, which allows you to test settings on the fly.

For example, here's how they set up a cloth flag tearing.http://www.studiodaily.com/studiomonthly/tutorials/trainup/7903.html
In LW you can do similar, except you used to have to setup the tear by hand (unless you use my plugin). Maya has a nice tear by stress function, but by nodalising Clothfx, LW would have this too.


I would definitely say that NT should have kept Ino on board to keep this new fluid system integrated with the current system, perhaps making it a subset of particles, as well as expanding on it.

Gregg "T.Rex"
07-24-2007, 05:02 PM
Well....
We've tested Maya 8.5 with Nucleus for this current job we do, but eventually we ended up with Real Flow meshes. Nucleus is much more than a mere particle base system and integrates extremely well in Maya, in areas LW fall quite short, like morphs and soft/hard bodies...

I agree, that NewTek should have kept Ino in chains, though.....

Dodgy
07-24-2007, 05:10 PM
Agreed, in and of themselves though, the dynamics are not that far apart, LW's are behind, but not as far as many claim. Nodalling the dynamics, so you can mix morphs and dynamics and other deformations more easily would put LW on an even or better level as nucleus. For example I use DPont's mdd nodes to mix my dynamics with regular deformations to a limited degree, but being able to put a ClothFX node in there instead would be so much more powerful.

zapper1998
07-24-2007, 05:11 PM
is there any way of getting the docs like the read me and the Help File translated to english???

for LiquidPack???

I installed it and well what is what??
its all in japanese

hmmmm

zapper1998
07-24-2007, 05:20 PM
Hello

This is D-Storm. Thank you for testing Liquid Pack trial version!

This plugin's author is INO-san who was at NewTek. His contract with NewTek expired last year and had been developing Liquid Pack by himself since at the start of this year. Thus, NewTek has no plans to build Liquid Pack in LightWave next version for now.

Currently we D-Storm has no plans to release English version because of the language problem. However, if we will receive many requests for English version, we might change our plan.

If we will release Liquid Pack English version, we will need help from you to correct documents in our poor English.

Thank you. :)

Need help translating
I will Help For free
Translating it...

Michael

Stooch
07-24-2007, 08:12 PM
Agreed, in and of themselves though, the dynamics are not that far apart, LW's are behind, but not as far as many claim. Nodalling the dynamics, so you can mix morphs and dynamics and other deformations more easily would put LW on an even or better level as nucleus. For example I use DPont's mdd nodes to mix my dynamics with regular deformations to a limited degree, but being able to put a ClothFX node in there instead would be so much more powerful.

lws dynamics are obsolete in comparison to maya.

voriax
07-24-2007, 10:01 PM
I still think it sucks that Blender, a free program, has a built-in fluid simulation that looks way better than anything LW particle dynamics can do, and so far it still outdoes this new fluid sim software for a realistic look/movement. It may take a lot of time and HD space, but in the end it's FREE.

Imagine a blender/realflow-style fluid sim built into LW that interacts with LW's particle/hard/soft/clothFX.. mmmmm. Now I'm dreaming.

Gregg "T.Rex"
07-24-2007, 11:15 PM
I guess, that's what we always do with LightWave; imagine "what if"...

RedBull
07-25-2007, 04:46 PM
Imagine a blender/realflow-style fluid sim built into LW that interacts with LW's particle/hard/soft/clothFX.. mmmmm. Now I'm dreaming.

While i like the idea, Blender will continue to outpace commercial 3D application Development, in the same way that Linux will outpace Microsoft in OS development. Opensource is not without it's advantages.

What Newtek need to do is increase the integration with 3rd party tools and other applications such as Blender. As it's only going to increase it's abilities over LW's.

As Blender is a free open source program, it's basically a free plugin for LW, the only problem is the exchange of data between the two programs.
I would prefer LW had Collada support, so i could use Blenders new particles, cloth and fluids and export the results to LW....

At the moment i need SurfReplacer from NextLimit just to surface (no nodes) my imported sequences.

Also Blenders fluids have been built for years it was first part of fluiid solver program called "El'Beem" which was transported to Blender under Googles SOC project.

Also the creater of El'Beem is now working with a research group for Ageia
in realtime fluids simulations.... http://graphics.ethz.ch/~thuereyn/

Here is some info on Blenders upcoming Particles and......
http://wiki.blender.org/index.php/BlenderDev/Particles_Rewrite

Cloth:
http://wxtools.com/

I also understand that XSI is working on a new FX system, and that Nucleus from Autodesk will be in full swing in the next round of Autodesk products..

So it seems CFD, PFX and Cloth are popular at the moment.

Gregg "T.Rex"
07-25-2007, 05:54 PM
So it seems CFD, PFX and Cloth are popular at the moment.

Newtek, doesn't think so as it seems....

duke
07-25-2007, 06:24 PM
I'm sure Newtek would like to do everything, but there's only so much they can concentrate on per-release, especially given Lightwave's price-point and thus the dev team size/capacity (I could be wrong).

RedBull
07-26-2007, 03:36 AM
I'm sure Newtek would like to do everything, but there's only so much they can concentrate on per-release, especially given Lightwave's price-point and thus the dev team size/capacity (I could be wrong).

I agree, Blenders CFD are (correct me if i'm wrong) the only LBM solver in a 3D application. XSI/3DMax/C4D/Modo/LW don't include any type of CFD...

Maya has a NS Solver, but personally i prefer Blender or Realflow..
And let's face it for NT to make an equiv to Maya/Blenders or Realflows established CFD, it would likely be not on par with the others.

Developers who have spent 15 years researching CFD at a University level, tend to be not cheap to hire.

So it's Obviously not something that much richer companies are willing to spend a lot of time researching as it's a quite complicated and expensive to develop.

I'm sure Newtek (and I'm sure Jarno and others) would be better off introducing something like Implicit or ISOSurfaces into LW, as this would be the kind of thing that would allow CFD and plenty of other things to be implemented into LW.

It's just lucky that LW has an SDK, as Dynamite, Liquid Pack, Syflex, ODEfl, and other plugins are the only thing that keeps LW comparative with the others in terms of dynamics systems.

PixelFarmer
07-27-2007, 11:58 PM
For some reason, all of Nemac's .mov's attached to this topic show up as gobbledy code. I have the latest QT installed and don't have that problem on other sites. Anyone else running into this?
M

cagey5
07-28-2007, 02:25 AM
What you need to do is a right click and save. When the save screen comes up you will notice it has an htm end instead of a .mov end. Change the htm to mov in the filename and it should play just fine.

Sensei
07-28-2007, 05:39 AM
For some reason, all of Nemac's .mov's attached to this topic show up as gobbledy code. I have the latest QT installed and don't have that problem on other sites. Anyone else running into this?
M

When this mess stops downloading just click File > Save Page As.. and it'll be saved on disk as regular original file that you can watch in off-line player..

PixelFarmer
07-28-2007, 10:00 AM
Thanks guys. Both of those solutions work, as long as the target file is saved as ".mov".

Nemac:
ItemInfo2.mov is impressive.

phillydee
08-01-2007, 04:46 PM
totally off topic... but couldn't resist :lol:



BTW I saw and option called Use Grand Plane. My test shows that create an invisible "ground" collision plane, so here we go again with a misspelled option on one of Mr Ino plugins, looks like it is his trademark! :) So that option means really "Use Ground Plane" ;)


Ha, it's funny they didn't spell it like "GLOUND"... LOL... grew up in Japan and they always mix up R's for L's, like the band GLAY... (http://en.wikipedia.org/wiki/Glay_(band))

Philbert
08-12-2007, 02:15 PM
For some reason, all of Nemac's .mov's attached to this topic show up as gobbledy code. I have the latest QT installed and don't have that problem on other sites. Anyone else running into this?
M

What I did was right-click and "Copy Link Location", then open Quicktime and paste it into File > Open URL. Then I could watch while it was downloading.


Ha, it's funny they didn't spell it like "GLOUND"... LOL... grew up in Japan and they always mix up R's for L's, like the band GLAY...

That's funny, I know the band, but it never occurred to me that they purposely misspelled their name.

creacon
11-21-2007, 09:04 AM
Hi,

I noticed that the demo size has changed and that the demo isn't working in 9.3.

Will there be an update soon and will we be able to create a surface (mesh) from the particles in the future?


creacon


Hello, this is D-Storm.

Thank you for many requests for "Liquid Pack" ! :)

In Japan, we released Liquid Pack today. And, we decided to release English version in near future thanks to your requests. However, I am sorry that I can not tell you exact release schedule because we do not have enough time to prepare it while working on other many tasks. So, please enjoy Liquid Pack trial version for a while.

The limitation of Liquid Pack trial version is as following:

-Expiration date: at the end of August 2007
-Could not save motion data.
-GridSize in NaminamiFX is no more than 100.
-Particle number in FluidParticle is no more than 5000.
-Could not render via network.

Thank you.

beverins
11-21-2007, 10:27 AM
Hello

This is D-Storm. Thank you for testing Liquid Pack trial version!

If we will release Liquid Pack English version, we will need help from you to correct documents in our poor English.

Thank you. :)

If you can provide a document in some form of English, with some side explanations of what the feature does, I might be able to correct the English into something proper.

Alternatively, you might try shopping it around to Wordware Publishing (I am not speaking for them, I don't know if they are willing or not) - since they also sell some plugins through ancillary deals, and they are home to a bunch of authors, they might be willing to help.

lino.grandi
11-21-2007, 12:12 PM
I downloaded the latest demo and it works...wind forces added! I think Liquid Pack is a must have for Lightwave owners.

ericsmith
11-21-2007, 05:11 PM
Is it actually available to purchase in an english version? The last I heard, it isn't ready yet.

Eric

lino.grandi
11-22-2007, 01:08 AM
Even if in beta version, you can purchase Liquid Pack from DStorm.

mccabejc
11-22-2007, 03:31 PM
DStorm has been around for how many years? And they have always had a problem finding time to get some college kid who understands English to spend a few minutes translating. Unbelieveable. I thought English was taught in all Japanese schools.

I just don't get it....

Philbert
11-22-2007, 03:44 PM
I thought English was taught in all Japanese schools.

It is, that doesn't mean everyone fully gets it or remembers what they do get a couple years later. It's tough to remember a language that you don't use every day. Not to mention English is a much harder language to learn than Japanese. I once made friends with this girl who moved here from France for a year. After 6 or 7 months she went back home for Christmas and told me later she was horrified to see how much French she had forgotten.

mccabejc
11-22-2007, 03:45 PM
Anyway, I downloaded the demo, all docs are in Japanese, and it comes with a whole bunch of demo scene files. I loaded the plugins, but without documentation on how stuff works it's not worth the time to muddle thru it.

BeeVee
11-23-2007, 02:05 AM
There may be problems with buying the Japanese version anyway. I have a user who did and is having trouble getting it going, after the demo version working fine.

B

Cageman
11-24-2007, 03:09 PM
Yikes, 3 hours.!
This little test I did did with dstorms pfx took only 3 mins, for a 230 frame simulation, You have to admit , thats pretty impressive, on a single 3 gig processor.
By my calculations, 5000 points per emitter, I cloned it 17 times, thats 85,00 points.
http://www.ogc-clan.com/images/screens/fluidfxp.avi
Divx codec 3 megs.

Hooly mooly... looks really promising... 3 minutes you say... I'm impressed and depressed at the same time... just shows how slow LWs dynamics are.

papou
11-24-2007, 04:01 PM
There may be problems with buying the Japanese version anyway. I have a user who did and is having trouble getting it going, after the demo version working fine.

B

yep, that is me.
I'm curious to know if i'm alone.

Mr Rid
11-24-2007, 05:48 PM
Hooly mooly... looks really promising... 3 minutes you say... I'm impressed and depressed at the same time... just shows how slow LWs dynamics are.


"Slow"... PFX might as well be "stationary" in comparison. Check post #60.

Stooch
11-24-2007, 08:53 PM
It is, that doesn't mean everyone fully gets it or remembers what they do get a couple years later. It's tough to remember a language that you don't use every day. Not to mention English is a much harder language to learn than Japanese. I once made friends with this girl who moved here from France for a year. After 6 or 7 months she went back home for Christmas and told me later she was horrified to see how much French she had forgotten.

hmm i always though that japanese was ranked as one of the most difficult languages to learn next to russian and some african tribes, while english being one of the easiest.

I learned to speak english fluently in about a year from 0 when i came to the US and still speak russian fluently after almost 17 years. hmmm. anyway there is no excuse for no documentation. not for a product that is being sold to english speakers.

Verlon
11-25-2007, 03:59 AM
Every langage thinks their's is the hardest. I think it has to do with where you start. If you start from german, spanish is painful (from what I hear). If you start from english, russian can be very difficult (from personal experience).

Learning conversational english in college is one thing, being able to translate a technical document in a specialized field is another. I had to take a foriegn laguange in high school here in the US. I can still count, ask where the bathroom is, and tell you what color my pencil is in french. I CAN'T explain how to use weight-maps, bone deformations, and hyper-voxels, and I doubt my french teacher could either. My french was never that good, and her Lightwave is probably a bit weak. So it is really THREE skills they need (Japanese, English, and Lightwave) and the free time to do it.

The plugin looks very interesting. ETA on the english version? Consider this an additional request.



Hah! I edited this post!!!!

Philbert
11-25-2007, 07:01 AM
As a native English speaker I personally can't say how hard it is, I've just been told that others say it's one of the hardest. From my personal learning I found German to be fairly easy, many of the words are similar or the same as there English equivalent. "Das haus ist gross und weis" Japanese was fairly easy to learn the spoken part, especially since there are no plurals, it's writing that's difficult. "Sono inu wa chiisai desu. (in Romaji)" French is ridiculous, I don't even know how the French learn it. :lol: "Je suis le grande fromage."

Elmar Moelzer
11-25-2007, 07:17 AM
I am a German native speaker and I think that english is one of the easiest languages to learn.
The grammer in English is simplistic and the vocabulary has both german and latin influences which should give both roman and german language natives a slight advantate. Also there are almost no exceptions to rules. In German e.g. Zunge (tongue) is female (which does not make sense at all).
French was OK, but I honestly lost most of it again in the many years since High School (I never quite liked it very much either).
I kinda liked Latin, even though it is very complicated and you will find having a hard time actually speaking it (it is a dead language after all). Even the romans did not really speak the latin you read in a Ovid, or Cicero, but a simplified more casual version of the languate that later became todays italian.
Still, I dont regrett learning it, ever.
CU
Elmar

Verlon
11-25-2007, 07:24 AM
no exceptions to rules in english? You ARE joking right?

As my english teature used to say "In english, the first rule is that there is an exception to every rule except this one."

duke
02-05-2008, 04:59 PM
Apparently the demo has expired. Nuts.

RedBull
02-05-2008, 06:40 PM
Until Nvidia support CUDA under Vista, LiquidPack doesn't work with accelerated GPU under Vista. :(

Japanese, French, English and German, Spanish and Italian are all basically easy to learn, speak and read....Only Chinese (mandarin) and Russian are hard to get the old head around IMO.

It would be nice to see NT promote or at least ask Dstorm to get an English page up, but than again NT don't even offer a Demo themselves, so i shouldn't be surprised.

loki74
02-06-2008, 12:57 PM
Japanese, French, English and German, Spanish and Italian are all basically easy to learn, speak and read....Only Chinese (mandarin) and Russian are hard to get the old head around IMO.


EH? I think Mandarin is WAY easier than Japanese... or even English for that matter... I mean yea speaking is a bit tougher, but Mandarin grammar is like 10x simpler than English! And if you know enough Kanji to be fluent in Japanese, then it should be easy to learn enough Han Zi to be fluent in Mandarin.

Elmar Moelzer
02-06-2008, 02:04 PM
Sorry, Verlon, but if you think that english has to many exceptions, you havent tried to learn german yet...
CU
Elmar

RedBull
02-06-2008, 03:23 PM
EH? I think Mandarin is WAY easier than Japanese...or even English for that matter... I mean yea speaking is a bit tougher, but Mandarin grammar is like 10x simpler than English! And if you know enough Kanji to be fluent in Japanese, then it should be easy to learn enough Han Zi to be fluent in Mandarin.

Well to put it this way, in our education system Japanese, German, Spanish and French are taught at a high-school level, where Chinese never will be, it's considered too hard for high school age students... Hirigana, Katakana and Kanji are all taught, and were all dead easy i can still speak and understand a fair bit of Japanese. I guess it's all subjective, but I'm talking from our education POV, which doesn't intend to add it, despite our business relations and close ties and distance. Mandarin or pinyin is also spoken in other hubs like Singapore.

I've worked with Chinese people for years, and it was way harder to pickup, i understand a bit, but can only speak a few phrases. Technically their is more incentive for us to learn Chinese, because it's spoken by over a billion people, where Japanese is obviously less than 200million? But each to their own. If you think it's easier feel free to learn it... :) My problem with Chinese, is so many simple phrases can mean so many different things, the context and if a certain slant or accent, or inflection etc, is added a certain time, will make one of twenty different meanings....

Personally French,and Japanese were the easiest to learn with German and Spanish slightly more difficult, but still quite easy... English is native.... (Well Australian is.....;) I think this is now well OT: Sorry!

Cageman
02-06-2008, 03:39 PM
Ok... I'll crack the OT-bubble now.. ;)

Does anyone use this plugin?

RedBull
02-06-2008, 04:40 PM
Ok... I'll crack the OT-bubble now.. ;)

Does anyone use this plugin?

Yes i bought into it at around page 3..... But because my main workstation is Vista x64, there is no 64bit compile, and Vista doesn't support CUDA, I didn't get the PPU because of the x64 concerns..... Well Nvidia are slow and stupid. (And now they own Ageia) (There are posts on the CUDA forums, saying it's in the next release, over 12months and two releases later, nothing!)So i can only use the standard version under LW32. Grrrr!

It's still really good for what it is, however i gave up on HV3 a long time back, it's just too limited in 2008 and that means most of what i render in LiquidPack, (much like the pics on Dstorms website) really look like crap..
But the solving is really good and it's fast and fairly accurate. Blender/Houdini gets used more for fluids these days, because of HV lack of updates.

Generally i use it as a Particle Replacement engine, and on occasions use NaminamiFX for stuff i don't need real CFD.

loki74
02-06-2008, 05:51 PM
Well to put it this way, in our education system Japanese, German, Spanish and French are taught at a high-school level, where Chinese never will be, it's considered too hard for high school age students... Hirigana, Katakana and Kanji are all taught, and were all dead easy i can still speak and understand a fair bit of Japanese. I guess it's all subjective, but I'm talking from our education POV, which doesn't intend to add it, despite our business relations and close ties and distance. Mandarin or pinyin is also spoken in other hubs like Singapore.

Ah, so you're also accounting for the availability of education. I think the idea that Chinese is too hard for HS level is ridiculous. I mean, they offer classes like AP History and AP Calculus... Chinese is most definitely easier than these (again, subjective, but still...) And, as far as I know, PinYin isn't a language, it's just an approximation of Mandarin phonetics using the Roman alphabet. I don't really get what you mean by "pinyin is also spoken..."


My problem with Chinese, is so many simple phrases can mean so many different things, the context and if a certain slant or accent, or inflection etc, is added a certain time, will make one of twenty different meanings....

If it's a different slant/accent/inflection (or rather, "tone"), then it isn't the same word with a different meaning--it's an entirely different word/phrase. There are four tones--flat, rising, falling then rising, and falling. And, even native Chinese speakers won't get 100% accurate differentiation between the second (rising) and third (falling then rising) tones. As you said, it's very context dependent, and that, IMHO makes it easier.

And IMHO, this difficulty is well compensated for by the fact that in Mandarin, tenses, conjugation, subject/verb agreement do not exist. It also has the same word order as English.


If you think it's easier feel free to learn it... :)

Indeed... I'm about to start my homework. ^_^ Do be honest, I find my university Mandarin class to be immensely easier than my HS Spanish class... O_O



....sorry to bring back the OT bubble!! >_< or rather, "dui bu qi!"

Titus
02-06-2008, 10:24 PM
Sorry, Verlon, but if you think that english has to many exceptions, you havent tried to learn german yet...

Or French.

On the other side, if you learn any of the latin languages, you may understand the rest very easy ;).

I hope nVidia adds support for PPU in their graphic cards very soon.

Verlon
02-06-2008, 10:48 PM
Have taken German (admittedly a very little a long time ago), French, and dabbled in Cambodian and Russian.

While Russian grammar seems deliberately designed to KILL native english speakers, there are enough exceptions in the english language to make me wonder what the blazes those people who made it up were thinking.

beverins
02-07-2008, 11:00 AM
So far, I'm using it for about a day and I like it a lot.

It has its limitations, and when compared to RealFlow there's no contest that RF is the winner for overall control, flexibility and stability.

However, LiquidPack has some advantages over RF in terms that it works natively inside LW and seems to be faster than RF. It's not as accurate as RF, but its good enough for most purposes. The one area where an update would be good is copying RF's ability to pause the frame advance and just "fill" a volume with fluid. As it stands now, seems to be that you have to resimulate the whole bloody thing each time you change a setting. So if you have a glass of water pouring out, you have to run a negative frame sequence of the glass getting filled up to get to the part where it starts pouring out. Or am I missing something?

When it comes to rendering, I think that HV's aren't really up to the task of simulating the appearance of liquids. I found the nflowtracer.p plugin off nextlimit's site to do a far, far better job. Of course, I'm far from being a master of controlling HV rendering... and of course it also depends what kind of fluid you are doing. For something like a white-water flood HV's are perfect. I'm currently doing more clear-water type things (particles fill up something and then flow around) and flowtracer does a hugely better job.. though being old and unsupported it likes to crash.

Being someone who likes to have one simulation do it all, I was trying the MAZE tutorial and I got something that looks like some sort of watery gel but not water. The particles for the fountain are too big but if I reduce the size of the droplets in nflowtracer to make the fountain look better, the water that collects and fills up the MAZE looks gel-like.

So much learning left to do. Anybody with some hints or tips? I saw that the particle size in liquidpack will affect how fast something fills up, and hence affects the resolution of the fluid. So that is something to look at.

But, overall, its a very worthy plugin for LW. Without the pack-in ageia card its about $300 compared to RF's $1000, so its well worth it, I think.

And on a side note to Newtek...
Since nextlimit doesnt really support flowtracer.p anymore... could you PLEASE consider buying it from Nextlimit and improving it for 9.3? For example, it hates Final Gather rendering mode. I honestly don't think they'd charge you much, and while the argument is that HV's can do it, I think this plugin makes things EASY and STRAIGHTFORWARD and has BETTER surface clumping ability.

Gregg "T.Rex"
02-07-2008, 11:14 AM
And on a side note to Newtek...
Since nextlimit doesnt really support flowtracer.p anymore... could you PLEASE consider buying it from Nextlimit and improving it for 9.3? For example, it hates Final Gather rendering mode. I honestly don't think they'd charge you much, and while the argument is that HV's can do it, I think this plugin makes things EASY and STRAIGHTFORWARD and has BETTER surface clumping ability.

I really doubt, but i second that...

Unless, NewTek comes up with HV 4.0, that do liquids as they should be... :)

Elmar Moelzer
02-07-2008, 01:11 PM
Well, I had latin in High School (not that much left almost 15 years later though). I still could not understand spanish or italian. I do have to admit though, that I could guess some meanings, every now and then.
French, while being a lating language, still was comparably hard for me to learn. Had a much easier time with english, which is IMHO a bit of a mishmash of old german with latin influences (and some french as well).
The grammar is easy enough in english, IMHO and the conjugations and declinations... well there is not much of that ;)
In Latin in contrary... I just say the english rose( or roses in plural) would look like that in latin:
singular: rosa, rosae, rosae, rosam, rosa
plural: rosae, rosarum, rosis, rosas, rosis...
I skipped the vocative even on that one...
And there are multiple types of declinations. The rose is the a- declination.
But you also have the us and the i, the is and the um declinations (if I remember this correctly... its been a long time).
E.g. philosophus, oppidum, mons, pater, etc
They all are different too. E.g. philosophus: philosophi, philosophi, philosiphum, philosiphi, philosophi, philisophorum, philosophis, philosophos, philosophis...
Oh did I mention that these were the easy ones?
Now the thing is that a manus (hand), can be female, while a bridge is masculine and so on. This is simillarily confusing in german actually...
In english all this is much more logical. Things are neutral females are female and males are masculine. No things that are masculine or something. The only exception that I can think of on top of my head are ships (which are female).
CU
Elmar

HowardM
02-07-2008, 01:22 PM
elmar cant you guys rework your voxel thingy to do better hvs? ;)

Elmar Moelzer
02-07-2008, 01:40 PM
We do have some plans and even some experimental builds here that do a lot of what HVs do (and more). Those would be quite usable for fluids as well, actually...
One reason why we are focusing more on the medical and scientific market right now is Dynamite. It seems to be aiming directly for the HV- replacement- segment of the market (for us, this would be more of a side- product). Unfortunately, it has been somewhat slow in its development recently.
CU
Elmar

-EsHrA-
02-18-2008, 07:42 AM
slow or dead? ..

id say start coding on sum cool alternatives for release! :)


mlon

mav3rick
02-18-2008, 08:20 AM
We do have some plans and even some experimental builds here that do a lot of what HVs do (and more). Those would be quite usable for fluids as well, actually...
One reason why we are focusing more on the medical and scientific market right now is Dynamite. It seems to be aiming directly for the HV- replacement- segment of the market (for us, this would be more of a side- product). Unfortunately, it has been somewhat slow in its development recently.
CU
Elmar

i would not call it development more like discontinue support

Stooch
02-18-2008, 01:09 PM
We do have some plans and even some experimental builds here that do a lot of what HVs do (and more). Those would be quite usable for fluids as well, actually...
One reason why we are focusing more on the medical and scientific market right now is Dynamite. It seems to be aiming directly for the HV- replacement- segment of the market (for us, this would be more of a side- product). Unfortunately, it has been somewhat slow in its development recently.
CU
Elmar

please by all means replace dynomite! i dont think its much of a competitor anymore.

EastCoastCreati
10-30-2013, 09:08 AM
Bump...I did email dstorm to see if there is any further development regarding Liquid Pack