PDA

View Full Version : TurbulenceFD and Lightwave 2018/2015



prometheus
01-19-2018, 05:32 PM
Hi folks!
a little discussion again around turbulenceFD perhaps, itīs been a while since I tried it even in 2015.
I was tinkering with the smoke and fire in blender, some things are great there, but in the end you want to edit the smoke smoothly and have fast renders and fast simulations, after messing with smoke and fire in blender cycles and tweaking the shaders, I couldnīt stand the slowness of cycles for the smoke, maybe I can improve on cycles preview speed and lower some settings..but I went to install the turbulenceFD 1433 version ..released 2018-01-15

simulation fast as usual I think, if not faster..and at a scale that blender would have a hard time to cope with I think.
Anyway..some quirks As always with setting turbulenceFD up, such as no default density value in the force channel, no density active in the container settings, and when I simulate, the objects disappears, only showcasing the smoke sim, so if I have a boat on fire..the boat dissappear while simulating, then back when finnished, can not recall if this was an issue before as well.

another issue, now in 2018, the openGL viewport with smoke density set to smoke shader, it just gives a flat grey or whatever color you choose for the smoke, in 2015 it could represent the smoke more accurately, in fact..initially the smoke became all black, since the color of the smoke had a gradient on the color smoke channel, and it was set to previousl layer, I changed it a fluid channel gradient, not much better, then I clicked held shift down and removed the texture, and that at leas got me a bright smoke, but still no real depth as in 2015 for the openGL preveiw, if I use shader instead of smoke, that yields the standard blue green display..but smoke preview seems wonky here, and when I try to go in to the color smoke channel again and try a fluid channel gradient again, there is no option to choose the channel gradient..itīs Gone, maybe there is a certain container sim channel that needs to be recalculated for it to access that fluid channel?

otherwise, renders decently, If you lower some settings, for testing..lowering smoke illumination quality can help vpr refresh better( that is if you choose to use illuminate smoke in the settings panel)
to be aware of..vpr needs volumetric shadows on as well.
I tried the multiple scattering, but it has to calculate a bit before it finnish any vpr update, quite slow I think, and any scene or object movement made, it automaticly re-calculates..so a bit of a slow workflow here.

More to be aware of, tweaking smoke density maps is proned to often crash, it was the case for me as well with older turbulenceFD version, all through the product line for years it seems unfortunately.
I have had, and now have some more questions regarding some of the above questions, unfortunately the forum at jawset is quite silent, and jascha I havenīt heard from for ages..so that doesnīt bring much light to questions unfortunately.

"Kat" Kelly Myers also brings attention to these current flaws, as well as some guidance on how to install turbulenceFD correctly....

My fixes would be..

fix default setting for easier start and get going with turbulence, as I mentioned about density channels not set nor active.
fix all the crashes with tweaking density maps ( may have something to do with VPR coding and not being able to deal with these kind of volumetrics properly)
fix the smoke shading in opengl, which seem to now have popped up..
fix the lack of both fire and smoke in openGL, (blender has it..houdini has it)
fix a possible bug? with fluid channel in smoke sometimes getting lost.
fix a faster multiscattering solution ..if possible.



get a better workflow or get new kind of forces, that affects..any bullet body, particles as well as the actual fire and smoke fluids, instead of having to set any object as force.
Implement particle Advection, again..and again :)

then I am off to the shop :) cause I like the simulation speed and how to tweak the smoke + fire shader an controlling noise etc.



https://www.youtube.com/watch?v=ihBUKP8gCL0

Mr_Q
01-19-2018, 05:36 PM
It's been made clear from the dev that development of TFD 1.0 is at the end of its cycle. Only support for bugs really are to be expected. (hence 1433) He's moved on to his next project and it is quite a beast. Nice to see he did spend time getting MSCAT VPR enabled. That falls in to a grey zone of bug or feature.

jwiede
01-19-2018, 06:00 PM
That OGL smoke shading issue sounds like a real downer for TFD + LW2018. It actually sounds like TFD1.x is still best hosted in LW2015.3 (at least among LW hosts) for a number of reasons.

It's unfortunate, for all the work Newtek/LW3DG did with OGL on LW2018, they didn't see fit to really better enable third-party devs' access to viewport OGL as well, and as a result LW2018 is coming out behind LW2015 w.r.t. third-party OGL stuff. It wouldn't have been that difficult to better document and open up third-party access to their OGL contexts, and had they done it in a decently-flexible manner, they'd actually be ahead of competitors packages' viewport integrations in some ways.

Anyway, I can't wait to see all the improvements in TFD 2.x, but Jascha needs to hurry up because (at least on C4D), competitors like X-Particles are now a lot more "serious" competition than they were in the past.

prometheus
01-19-2018, 06:11 PM
It's been made clear from the dev that development of TFD 1.0 is at the end of its cycle. Only support for bugs really are to be expected. (hence 1433) He's moved on to his next project and it is quite a beast. Nice to see he did spend time getting MSCAT VPR enabled. That falls in to a grey zone of bug or feature.

what..wasnīt aware of that, so you mean there will be no more future turbulenceFD version? and no replacement? or is this next project something new in fluids that will be available for Lightwave?
Or do you mean only the end of a certain version cycle of it?

- - - Updated - - -


That OGL smoke shading issue sounds like a real downer for TFD + LW2018. It actually sounds like TFD1.x is still best hosted in LW2015.3 (at least among LW hosts) for a number of reasons.

It's unfortunate, for all the work Newtek/LW3DG did with OGL on LW2018, they didn't see fit to really better enable third-party devs' access to OGL as well, and as a result LW2018 comes out behind LW2015 in many ways. It wouldn't have been that difficult to better document and open up third-party access to their OGL contexts, and had they done it in a decently-flexible way, they'd actually be ahead of competitors packages' viewport integrations in some ways.

Anyway, I can't wait to see all the improvements in TFD 2.x, but Jascha needs to hurry up because (at least on C4D), competitors like X-Particles are now a lot more "serious" competition than they were in the past.

Aha..so there was just a cycle end to it all?

jwiede
01-19-2018, 06:31 PM
Aha..so there was just a cycle end to it all?

As far as I know, Jascha is just finally ending 1.x support and moving on to 2.x. Considering just how long 1.x lasted, I imagine he's desperately in need of upgrade revenue -- I'm shocked he didn't end 1.x a long time ago, frankly.

Greenlaw
01-19-2018, 06:49 PM
I thought I read a while back that the next Turbulence is going to be a standalone tool. Is that accurate or just a rumor?

Thanks in advance for any reliable answer.

MichaelT
01-19-2018, 11:28 PM
Oh I sincerely hope that is not the case.. he'll loose customers because of it. Because then you can just as well stick with Houdini or Blender, which can generate smoke etc.. for free. That it integrates is a major selling point.

prometheus
01-20-2018, 09:09 AM
Oh I sincerely hope that is not the case.. he'll loose customers because of it. Because then you can just as well stick with Houdini or Blender, which can generate smoke etc.. for free. That it integrates is a major selling point.

A bit worrying yes, and I agree with you..
Jascha has been very silent on his own forums, canīt blame him if heīs really busy etc, but it would be good to have him a tiny bit active in their and answer a bit of questions, or if he can, make a clear statement of wether or not turbulenceFD2 will be standalone only or working within cinema4d and lightwave..or both?

MichaelT
01-20-2018, 09:21 AM
With that I'm guessing that I might have to get the 1.x version before it is too late :/

prometheus
01-20-2018, 09:26 AM
With that I'm guessing that I might have to get the 1.x version before it is too late :/

Why...do you think he would close the shop of the lightwave plugins after that?and as you said yourself, why not go for houdini or blender?

Greenlaw
01-20-2018, 09:44 AM
Oh I sincerely hope that is not the case...That it integrates is a major selling point.

Yes, I agree. I was an early adopter of TFD because it was a LightWave plugin, and because of Jascha's excellent support. It's been an awesome tool and I've used it in dozens of films and commercials over the years.

If TFD is indeed going standalone, I may upgrade if the workflow is still easy and it plays well with LightWave. But of course, if it goes standalone, I may as well consider other programs too (thanks to OpenVBD support.)

BTW, not trying to stir trouble or anything...going standalone was just something I saw mentioned in another thread and it may only be a rumor. I was only asking if anybody has official info from Jascha about this.

If TFD continues as a LightWave plugin in the future, I'm definitely on-board with it.

thomascheng
01-20-2018, 10:07 AM
I'm pretty excited about TFD2, from the last presentation he did on the technology a long time ago. If he can pull it off, I wouldn't mind it being a standalone if it is easy and intuitive.

MichaelT
01-20-2018, 10:45 AM
Why...do you think he would close the shop of the lightwave plugins after that?and as you said yourself, why not go for houdini or blender?

Oh no.. that's not what I meant. Usually when a program changes major version, the previous one goes away (Just look at the LW store) What I meant is that if it goes standalone, then the only option to have it integrated is to get it now.
If I choose to do this.. then I'll get it for C4D, and not LW. Because of the lack of support for advection in the LW version.

MichaelT
01-20-2018, 10:56 AM
Just checked out this video:

http://on-demand.gputechconf.com/gtc/2015/video/S5611.html

He says right in the beginning that TD2 will intergrate into LW & C4D (and a few more)

prometheus
01-20-2018, 11:09 AM
Oh no.. that's not what I meant. Usually when a program changes major version, the previous one goes away (Just look at the LW store) What I meant is that if it goes standalone, then the only option to have it integrated is to get it now.
If I choose to do this.. then I'll get it for C4D, and not LW. Because of the lack of support for advection in the LW version.

Same for me..almost, I think the c4d version seems better intergrated, the problem is that I am not keen on investing in cinema4D, and not motivated unless I need it for some work and not as now as Mainly hobbyist, itīs not only particle advection, itīs also about previewing smoke ..which doesnīt seem to work that well with VPR, and cinema4d has itīs own, then I wonder if turbulenceFD in cinema4d can use weigth maps for emission.



Just checked out this video:

http://on-demand.gputechconf.com/gtc/2015/video/S5611.html

He says right in the beginning that TD2 will intergrate into LW & C4D (and a few more)

Great research, I will look at that, currently both friday and this saturday, and problably sunday..I have set aside to try both blender smoke..openVDB exports and messing with the new turbulenceFD, may thread about blender openVDB just might be interesting as well..
And I have yet to install houdini apprentice again and try that workflow...puh.

Michael, You havenīt considered 3d max and fumefx? Not sure if you have had the time to check out Mc kays fire and explosion, destruction effects series?

MichaelT
01-20-2018, 11:15 AM
I did way back.. but when I got C4D I never renewed the subscription. I could no longer motivate paying for something that is nearly fully covered among the tools I already have.

Marander
01-20-2018, 11:15 AM
irrelevant - removed

MichaelT
01-20-2018, 11:24 AM
Watched through the clip now.. and TFD2 looks very interesting. As I understood it, it will work like this. You have a server that runs independently, but it can run on the same machine. Then the client is the part that will integrate into LW for instance, and then just do communication with the server for what it needs. This way the program is free to divide tasks in the system more efficiently, and only need to focus on handling requests, and deliver results to the client. This means the performance will be much better, and if you need more performance, you only need to get more parts for your computer. Lots of good things in that video, I suggest watching it.

Marander
01-20-2018, 11:33 AM
irrelevant - removed

Greenlaw
01-20-2018, 01:30 PM
Just checked out this video:

http://on-demand.gputechconf.com/gtc/2015/video/S5611.html

He says right in the beginning that TD2 will intergrate into LW & C4D (and a few more)

Thanks for the info! That's fantastic news.

Greenlaw
01-20-2018, 01:38 PM
As I understood it, it will work like this...

Oh! So it sounds like it's sort of both: TFD2 runs as its own (server) program that you can interact with in other 3D (client) program probably by plugin? Interesting.

MichaelT
01-20-2018, 01:42 PM
Yes, that's how I got it. Which means you have the server run on a GPU farm for instance, and the client communicates with that (since he did mention it can run remotely) which would increase performance quite considerably.
Barring he changed his mind for some reason :) But I wouldn't bet on that one.

AnimeJoex
01-20-2018, 01:43 PM
So it sounds like it's sort of both: TFD2 runs as its own (server) program that you can interact with in other 3D (client) program probably by plugin? Interesting.

I've only played around with the demo for a little bit but isn't that how RealFlow works?

prometheus
01-28-2018, 08:34 AM
been fiddling a bit more with turbulenceFD in Lightwave 2018.

Previously Kelly Myers reported issues when tweaking maps, and crashes, and I knew I had that issue with all previousl beta versions of turbulenceFD, strangely though..I havenīt had one single crash now with 2018 and tweaking the maps..while running VPR active, so that is kind of weird, and I tweaked opacity channels as well..where it often used to crash before, I do not like though that tweaking the maps is extremely slow, can hardly move the control points in the maps due to lag, or vpr interfering with it.


https://www.youtube.com/watch?v=ihBUKP8gCL0


To note, turbulenceFD requires the old legacy volumetrics, so you will have to check that in the volumetric tab before anything can be seen in VPR or rendering, this also poses a workflow issues of dealing with two kind of volumetric environment that doesnīt fuse together so to speak, you can not render explosions, fire or particle fluids with smoke within clouds made from the new volumetric system..in any other way than to render out separate passes, which could be problematic when you maight need smoke go down below clouds and similar, casting shadows etc.
This means the volumetric scattering of light will not be doable for turbulenceFD either, so no casting of volumetric rays if you plan to save out cloud cache etc, or just light streaks behind a nuke or something, so that means post trickery.

TurbulenceFD comes with defaults I still to this day do not get why certain things are shut off from getting started from scratch as easy as possible, meaning, the density isnīt activated in the fluid container, nor is the temp or density channels in the emitter force, having one of them is crucial to get anything playing ...and for those newbies, you would be confronted with not getting started as easy as you could have.

regarding VPR speed, you can always lower the illumination quality drasticly while tweaking, only to raise it more for renderings, it helps a lot in VPR update speed, and....keep your VPR windows to 1/4 full screen, going above that and it will be very slow.

other issues I have that simply make the smoke not appear, it is the smoke color channel, it has a fluid channel active, but it will prevent anything from be seen unless that is removed..so that is weird.

Love the quality of the renderings though and the smoke and fire shader.
Crashes?...yes I had that...it showed up during higher res cached files at the end of the simulation with lotīs of crashes in such case, so some kind of memory issue there, but crashing it should do.

sadkkf
01-29-2018, 07:05 AM
So while LightWave isn't necessarily GPU aware, TFD 2.0 will be? This makes me wonder how well it will work with one GPU.

01-29-2018, 07:22 AM
Works a treat with a Titan. Can't imagine simming without it.
Rober

sadkkf
01-29-2018, 07:33 AM
Works a treat with a Titan. Can't imagine simming without it.
Rober

I'm sure it does, but not all of us can or want to spend $1000 US on a video card. :)

jwiede
01-29-2018, 09:22 AM
So while LightWave isn't necessarily GPU aware, TFD 2.0 will be? This makes me wonder how well it will work with one GPU.

Huh? TFD 1.x is already "GPU-aware", why shouldn't TFD 2.x be?

You can use a GPU to accelerate TFD (and should), but it isn't _required_. My 980Ti makes quick work of many tasks in TFD.

01-29-2018, 09:30 AM
My point was, as Jwiede has stated, that TFD has had gpu functionality for some time. It wasn't an advert for what you ought to have.
Robert

Greenlaw
01-29-2018, 09:35 AM
JWiede beat me to it but, yeah, it's always been capable of GPU processing and not a requirement.

I think GPU is currently used for calculation only though, not rendering. Maybe that part will change in 2.0? In any case, GPU use will probably still be optional. (That's just a guess...I don't really know.)

prometheus
01-29-2018, 11:02 AM
two sides of this coin..
GPU as Render aware
GPU as simulation aware

GPU has been in there for a long time in Lightwave when simulating, and my gtx 1080 beats my intel i7-6700 3.40 ghz dual core, canīt imagine simulating with the CPU, at a certain amount of voxels memory...GPU will of course switch over to cpu since it can not handle that amount of data, same with Blender Fire and smoke simulations, except you can choose to render with CPU or GPU.

GPU rendering directly in lightwave, not possible.
GPU rendering aware with third party renders together with lightwave...yes, that platform is there with octane, our Lightwave Guru has worked a bit with that.

not lightwave guru..but some other info..

https://www.youtube.com/watch?v=AL7itR1BxM4

sadkkf
02-04-2018, 09:23 AM
Finally installed TFD with the latest build, but it does nothing but crash.

I open Layout, press the Containers button, or any other, and instantly it crashes.

Am I missing something?

JamesCurtis
02-04-2018, 09:31 AM
Did you make sure that you put the dependency files in the LW bin folder?

prometheus
02-04-2018, 10:05 AM
Make sure to check Kelly Myers intro, where he goes through where to install turbulenceFD, itīs important, I watched it for precautionary actions and I had no crash.
Kelly Also warns for tweaking maps while having vpr on..constant crashes for him it seems, and that was valid for the old versions for me, but what is weird now..is that I do not have those crashes so far in 2018 and the 1433 turbulenceFD build.

there are issues of the smoke color gradient making the shading not showing up properly in open GL on my installement.
and I wonder how the future will be with the new volumetrics, since turbulenceFD is build around the old legacy, and will not work with new volumetrics of clouds and the likes, so we got old stuff and new stuff not able to work together properly.


https://www.youtube.com/watch?v=ihBUKP8gCL0

OlaHaldor
02-05-2018, 03:25 AM
not lightwave guru..but some other info..

https://www.youtube.com/watch?v=AL7itR1BxM4


True, I'm no guru. But that was a pretty fast setup to see if I could render decent smoke with Octane. :)
And for mid to far shots, I think that works good enough. I wouldn't use the render for close ups, unless it's out of focus perhaps..

Since you can't use any displacement on the TFD sim at rendertime with Octane (like you can with the native LW renderer) you have to sim at a higher res, or make due with what you got.

Fire and explosions on the other hand is a bit tricky with Octane. Mapping a ramp on the sim to make it look realistic is hard imho.

sadkkf
02-05-2018, 04:54 AM
Make sure to check Kelly Myers intro, where he goes through where to install turbulenceFD, itīs important, I watched it for precautionary actions and I had no crash.
Kelly Also warns for tweaking maps while having vpr on..constant crashes for him it seems, and that was valid for the old versions for me, but what is weird now..is that I do not have those crashes so far in 2018 and the 1433 turbulenceFD build.

there are issues of the smoke color gradient making the shading not showing up properly in open GL on my installement.
and I wonder how the future will be with the new volumetrics, since turbulenceFD is build around the old legacy, and will not work with new volumetrics of clouds and the likes, so we got old stuff and new stuff not able to work together properly.


https://www.youtube.com/watch?v=ihBUKP8gCL0

Sigh...Thank you. I wish the instructions were updated to include this info.

Spinland
04-27-2018, 07:58 AM
To note, turbulenceFD requires the old legacy volumetrics, so you will have to check that in the volumetric tab before anything can be seen in VPR or rendering

Ah! Thanks for that tidbit!

Started a new gig recently and decided to give 2018 a go since there's some ramping up time built into the mix. They want me to blow up a building as part of the job and TFD wasn't playing nice. This fixed it.

Thanks again! :jam:

Mark

prometheus
04-27-2018, 09:48 AM
Ah! Thanks for that tidbit!

Started a new gig recently and decided to give 2018 a go since there's some ramping up time built into the mix. They want me to blow up a building as part of the job and TFD wasn't playing nice. This fixed it.

Thanks again! :jam:

Mark

Personally I do not want to wrestle with this two part volumetric system, and itīs one of many many factors that weighed in when I decided to not upgrade.
TurbulenceFD is incompatible with the new volumetrics ..so clouds and something related with turbulenceFD like aeroplane smoke, fire, tornados, or dynamic cloud parts in concuntion is not possilbe...not without switching between these two
volumetric system and the comp in post, which may have drawbacks.

Glad that the heads up on this solved your problem though.
Hoping for Turbulence or whatever comes to work with the new system in the future.

Spinland
04-27-2018, 10:03 AM
I hear ya. As I see it, nothing is perfect so I'm good with "good enough to get the job done." Right now that's all I need out of TFD, I get some quality time learning more about 2018, and the client loves the work.

Win-win! :beerchug:

Thomas Leitner
04-28-2018, 04:33 AM
...TurbulenceFD is incompatible with the new volumetrics ..so clouds and something related with turbulenceFD like aeroplane smoke, fire, tornados, or dynamic cloud parts in concuntion is not possilbe...not without switching between these two
volumetric system and the comp in post, which may have drawbacks....

Until the update TurbulenceFD for the new volumetrics, you can use OpenVDB to render TurbulenceFD simulations in LW new volumetric system.

ciao
Thomas

prometheus
04-28-2018, 06:01 AM
Until the update TurbulenceFD for the new volumetrics, you can use OpenVDB to render TurbulenceFD simulations in LW new volumetric system.

ciao
Thomas

I know..thanks for making that clear, which I forgot to mention.

SBowie
04-28-2018, 07:21 AM
They want me to blow up a building as part of the job... Now there's a line guaranteed to set off alarm bells at the NSA. Any big black SUVs parked across from your house yet? ;)

Spinland
05-02-2018, 03:13 AM
Now there's a line guaranteed to set off alarm bells at the NSA. Any big black SUVs parked across from your house yet? ;)

:help:

djlithium
05-02-2018, 09:33 PM
Hi Kats. New video out here today. http://www.liberty3d.com/2018/05/turbulencefd-advanced-concepts-and-projects-engine-fire/

Regarding some of the issues we have been having with TFD in 2018.0.x
I haven't tested just yet but many of the bugs I was running into with TFD and LW 2018.0.x - x.0.3 are supposed to be fixed. These largely have to do with the calculation slow downs and some of the other funk but I have yet to test but I'm doing that next (like today) and then proceeding with 2018 and TFD vids. I will keep you posted regarding the findings in terms of performance much like I did in the video from a couple of months back so please stay tuned to www.liberty3d.com for updates there. Jascha has kinda gone dark, which is somewhat concerning - but hopefully he is just really deep into making awesome and not pulling a CanTarcan.

thomascheng
01-25-2019, 01:33 PM
Anyone try TFD with 2019? Does it work well?

Markc
01-25-2019, 02:22 PM
It works, only added a container to make sure it didn’t crash.
Will give it a proper go though....:)

jwiede
01-25-2019, 03:37 PM
Jascha has kinda gone dark, which is somewhat concerning - but hopefully he is just really deep into making awesome and not pulling a CanTarcan.

The lack of updates (esp. to the Mac LW version of TFD), combined with the lack of interaction by Jascha w.r.t. LW TFD in forums, etc. definitely raises questions. He's still actively updating the C4D versions of TFD for both Win & Mac, occasionally even interacting with TFD C4D users in the TFD forums. He's clearly working on TFD (at least for C4D), but in some ways that makes the lack of LW TFD activity even more concerning.

Marander
01-25-2019, 04:59 PM
The lack of updates (esp. to the Mac LW version of TFD), combined with the lack of interaction by Jascha w.r.t. LW TFD in forums, etc. definitely raises questions. He's still actively updating the C4D versions of TFD for both Win & Mac, occasionally even interacting with TFD C4D users in the TFD forums. He's clearly working on TFD (at least for C4D), but in some ways that makes the lack of LW TFD activity even more concerning.

Actually the last TFD updates were pretty recent, LW 2018-12-11 (v1439) and for C4D 2019-23-01 (v1440).

I have licenses for both.

In C4D R17-R20 it works perfectly and stable on any machine.

In LW it only works only on one machine that is not used for 3D without GPU. On all other machines it crashes immediately Layout with clicking on any TFD button (any TFD or LW version).

I'm using the latest NVidia drivers and TDF is installed outside the LW folder. Using NVidia cards up to GTX1080.

Maybe somebody has an idea. My support incidents at Jawset were never answered / resolved unfortunately. I was never able to use TFD in LightWave. I guess it's a conflict with a dll or another process.

It works in LW2019 (again, only in the one installation without GPU).

next_n00b
01-26-2019, 03:19 AM
My TFD v1439 is working fine on win7, K6000. TFD is installed in %lightwave_root%\Plugins\TFD... and version prior to that was working fine with gtx1070Ti. And this old K6000 Quadro works slightly faster than gtx 1070Ti in most (not all) cases.

EDIT: forgot to mention LW version, LW2018.07

Marander
01-26-2019, 06:04 AM
My TFD v1439 is working fine on win7, K6000. TFD is installed in %lightwave_root%\Plugins\TFD... and version prior to that was working fine with gtx1070Ti. And this old K6000 Quadro works slightly faster than gtx 1070Ti in most (not all) cases.

EDIT: forgot to mention LW version, LW2018.07

Thanks next_n00b. I guess it's a conflict with another software installed (that I have on all machines).

I have tons of graphic applications and plugins installed, they all work fine and stable, as well as other software (like VMware Workstation or IDEs like Visual Studio and Eclipse).

Really weird because that happens only in TFD for LW. Maybe one day I'll figure out what the conflict is.

next_n00b
01-26-2019, 07:19 AM
I think you should try moving/"fresh install" of TFD inside LW install folder. I do not know if TFD worked for you before installed this way? Anyway I feel your pain. I sleep better if I know that my workstation is back in a software order in just 15 min with Norton Ghost.

Marander
01-26-2019, 08:22 AM
I think you should try moving/"fresh install" of TFD inside LW install folder. I do not know if TFD worked for you before installed this way? Anyway I feel your pain. I sleep better if I know that my workstation is back in a software order in just 15 min with Norton Ghost.

Thanks! Ahh Ghost, I used that too many years ago and it was great, don't know that it still works with Win10. But I find the current Symantec software too bloated.

I tried various combinations of the TFD install, inside or outside the LW or NewTek Folder, copying the dll's, fresh install of LW without other plugins etc.

One thing I know is that the old version 9xx worked, I had a demo that didn't crash. Any version 1xxx doesn't work on various generations of GeForce. Luckily TFD works within C4D without issues. Hopefully I figure it out one day and keep trying with every new version of LW and TFD.

next_n00b
01-26-2019, 09:20 AM
Last time I was checking there was no Ghost for win10.

In that case, I think you should start looking for your answers on a fresh windows install with LW as the only software installed. I assume that you have already exhausted all read/write permissions possibilities.

Marander
01-26-2019, 11:06 AM
Last time I was checking there was no Ghost for win10.

In that case, I think you should start looking for your answers on a fresh windows install with LW as the only software installed. I assume that you have already exhausted all read/write permissions possibilities.

Yes, tried different permission settings, UAC, run as admin etc.

When I get my next new machine I will install LW and TFD very first and observe at which point / after what software install it gets broken. My current workstations and laptops are quite new and they work perfectly otherwise, so that it will have to wait some time. Thanks anyway!

Lito
01-26-2019, 12:40 PM
Working here in 2018 and 2019 in Win 10 w/ Nvidia 1080Ti Driver 416.81. Have you checked the windows event viewer? It might give you a clue as to what is causing the crash. Start LW press TFD button let it crash then open up computer management->event viewer and look under application and system to see if it left you a clue there.

Markc
04-17-2019, 11:52 AM
The lack of updates (esp. to the Mac LW version of TFD), combined with the lack of interaction by Jascha w.r.t. LW TFD in forums, etc. definitely raises questions. He's still actively updating the C4D versions of TFD for both Win & Mac, occasionally even interacting with TFD C4D users in the TFD forums. He's clearly working on TFD (at least for C4D), but in some ways that makes the lack of LW TFD activity even more concerning.

Just looked on the Jawset forums, noticed Jascha STILL hasn't replied to your thread from Feb 2018?

prometheus
04-17-2019, 01:09 PM
Just looked on the Jawset forums, noticed Jascha STILL hasn't replied to your thread from Feb 2018?

and that is why I dont bother purchasing..partly, I have given up on posting any questions on that forum..and it certainly isnīt promisng for the future.
it wasnīt one thread..it was several threads unanswered.

and apart from not getting answers from my questions, there are hardly none many months, itīs too uncertain and I have a feeling or kind of a fear of it slowly being killed off for lightwave and cinema4d is where he may put all energy and effort on.

Even though I purchased Lightwave 2019, my experience with turbulenceFD in 2018 isnīt good enough in terms of multiple scattering, and that it can not be used simultaneously with the new PBR volumetrics, the mutliple scattering is frustratingly slow and with constant calculation time update for each tweak, so much easier and faster in blender.

and I need weight paint emission and full fire and smoke opengl presentation, that is in blender.
Blenders forces works on all bullet bodies as well as fluids, that isnīt possible with turbulenceFD, those forces are made within the fluid system and doesnīt interact with bullet bodies

There are things I prefer in turbulenceFD though, such as I think it calculates faster, it is easier to set various noise procedurals to your liking compared to blenders, and it is harder to get good quality stepless volumetrics in blender.

Markc
04-17-2019, 02:09 PM
I had the same thoughts on the LW version longevity.......are TFD’s days numbered.

wingzeta
04-17-2019, 02:29 PM
I had the same thoughts on the LW version longevity.......are TFD’s days numbered.

At some point there was a video of the next version of TFD, being worked on. It was a stand alone application that worked outside the "host application". The idea being, it could calculate much bigger and better, separate from your scene overhead, if I am recalling correctly. Anyhow, it seemed the other part of the idea was, he would develop one application, and then it would be a simpler matter to make a I/O plugin for each 3D app. Like GoZ, but obviously it would be able to interact with your scene geometry. So it may not be so dark for LW, if all it takes is a plugin to interface with the same TFD app that will also work for C4D and others. If anyone remembers the video I'm talking about, feel free to post a link.

Markc
04-17-2019, 03:58 PM
It still doesn’t explain why there has been no reply to forum posts for over a year though.
Specifically when paid customers are asking questions.

gdkeast
04-17-2019, 04:19 PM
This is helpful to know. So is it safe to say that TFD isn't really being supported for LW any longer...buy at your own risk?

wingzeta
04-17-2019, 05:02 PM
The LW community gets the silent treatment once again. I agree that is bad. My suspicion is that Jawset, which I think is just one guy, intended the new version to be released by now, but it has taken longer to finish, and because he is one guy, he's done a bad job of communication, while having nose to grindstone, trying to get the new version out. I imagine he was planning to end support for the old version of TFD when the new one came out. He did update it to work with 2018+, but it doesn't take advantage of 2018+ new capabilities, because that would be covered by the new version. I would only get the current version now, if you need it for a job. It does work.

Qexit
04-17-2019, 05:35 PM
Hm, maybe this will leave an opening for Deep FX Studio when it is finished ?

http://www.deepfxworld.com/deep_fx_studio.php

prometheus
04-18-2019, 12:52 AM
At some point there was a video of the next version of TFD, being worked on. It was a stand alone application that worked outside the "host application". The idea being, it could calculate much bigger and better, separate from your scene overhead, if I am recalling correctly. Anyhow, it seemed the other part of the idea was, he would develop one application, and then it would be a simpler matter to make a I/O plugin for each 3D app. Like GoZ, but obviously it would be able to interact with your scene geometry. So it may not be so dark for LW, if all it takes is a plugin to interface with the same TFD app that will also work for C4D and others. If anyone remembers the video I'm talking about, feel free to post a link.

To me that workflow feels like a longshot, but Jascha should know better than me about these things I reckon, questions are how bullet fractured pieces in the scene would be treated in the standalone version in order to calculate interaction, and with the native host version..you would get an instant render from letīs say a cloud simulation in scene context, you would probably loose out on those aspects, thus I do not believe in that approach of a standalone version, and I may start to completely switch to blender for this kind of stuff or houdini.
If he provides a standalone version for when it is needed with only standalone and then a host tfd version, that may work..but probably too much work for Jascha to do.

darkChief
04-18-2019, 05:33 AM
Hm, maybe this will leave an opening for Deep FX Studio when it is finished ?

http://www.deepfxworld.com/deep_fx_studio.php

Still haven't worked out a gas solver yet, looking for something versatile and fast. Still haven't decided if it will be classic Eulerian or something like Blenders (very multi threadable). Both have disadvantages and advantages. There are many ways to simulate liquids, but not so many for gas strangely.

I encourage everyone to try Deep FX Studio, looking for as much feedback as possible. Also offering a free license to anyone who owns Deep Rising FX already. Offer is limited till next year. Just get in touch with support.

Oldcode
04-18-2019, 07:07 AM
Still haven't worked out a gas solver yet, looking for something versatile and fast. Still haven't decided if it will be classic Eulerian or something like Blenders (very multi threadable). Both have disadvantages and advantages. There are many ways to simulate liquids, but not so many for gas strangely.

I encourage everyone to try Deep FX Studio, looking for as much feedback as possible. Also offering a free license to anyone who owns Deep Rising FX already. Offer is limited till next year. Just get in touch with support.

Hey Dark Chief,

You're fluid software looks amazing. I've been reluctant to buy only because I already have Real Flow, but if you can crack the flames, gas and smoke thing, I'll definitely be buying it! :D

I've wanted to get TurbulenceFD for a while, but it was just too expensive for one effect that can be done by other tools in Lightwave, although not as well. Even with the group buy in, it was still a lot and after what I'm hearing about lack of support, I'm kind of glad I held off.

Good luck with your research.

raymondtrace
04-18-2019, 08:48 AM
It still doesn’t explain why there has been no reply to forum posts for over a year though.
Specifically when paid customers are asking questions.

There is a major difference between "paid customers" and "paying customers", especially in a one-person operation.

Markc
04-18-2019, 11:19 AM
Still haven't worked out a gas solver yet, looking for something versatile and fast. Still haven't decided if it will be classic Eulerian or something like Blenders (very multi threadable). Both have disadvantages and advantages. There are many ways to simulate liquids, but not so many for gas strangely.

I encourage everyone to try Deep FX Studio, looking for as much feedback as possible. Also offering a free license to anyone who owns Deep Rising FX already. Offer is limited till next year. Just get in touch with support.

Any progress on mac development?

darkChief
04-19-2019, 02:27 AM
Any progress on mac development?

I've been looking at this for a while. This is probably not going to happen anytime soon. Sorry.

Markc
04-19-2019, 03:08 AM
Understood......:grumpy:
Keep up the good work all the same.....:thumbsup:

wingzeta
04-19-2019, 01:26 PM
To me that workflow feels like a longshot, but Jascha should know better than me about these things I reckon, questions are how bullet fractured pieces in the scene would be treated in the standalone version in order to calculate interaction, and with the native host version..you would get an instant render from letīs say a cloud simulation in scene context, you would probably loose out on those aspects, thus I do not believe in that approach of a standalone version, and I may start to completely switch to blender for this kind of stuff or houdini.
If he provides a standalone version for when it is needed with only standalone and then a host tfd version, that may work..but probably too much work for Jascha to do.

My impression from the video presentation, which I haven't been able to find unfortunately, was that it would still work within your scene as it always did, but that the heavy lifting of calculations would be handled in the outside app. The outside app may have also been for getting a much faster preview to get the look tweaked faster. So the workflow looked pretty good. It didn't seem like capabilities were being lost, just made faster. Recalculating after tweaking is one of the big slowdowns in simulations, so it seemed like he was solving that.

Again, this is just from memory, and I could be getting it mixed up, but the simulations he was showing were also taking advantage of some advancements where rather than having the volume have to be contained in a cube, that could sometimes be bigger than it needed to be, or too small and clip, it was able to carve up the area of the simulation into only the spaces needed, thus being more efficient in calculations. Hopefully it will be available for LW if he gets it done.

prometheus
04-20-2019, 06:01 AM
My impression from the video presentation, which I haven't been able to find unfortunately, was that it would still work within your scene as it always did, but that the heavy lifting of calculations would be handled in the outside app. The outside app may have also been for getting a much faster preview to get the look tweaked faster. So the workflow looked pretty good. It didn't seem like capabilities were being lost, just made faster. Recalculating after tweaking is one of the big slowdowns in simulations, so it seemed like he was solving that.

Again, this is just from memory, and I could be getting it mixed up, but the simulations he was showing were also taking advantage of some advancements where rather than having the volume have to be contained in a cube, that could sometimes be bigger than it needed to be, or too small and clip, it was able to carve up the area of the simulation into only the spaces needed, thus being more efficient in calculations. Hopefully it will be available for LW if he gets it done.

If that is the case, then it would be far more interesting and feasable, will have to see..to little info on his site on his development and no answers..so it leads to uncertainty and speculations anyway.
I can just hope it doesnīt take so long time that I go on with something els, only to discover that he then would release something awesome.

As for calculating areas near the actual fluid simulation instead of bounding boxes, that is what houdini has implemented well I think.

The foremost issue for me with TFD currently, is that it can not use the new volumetrics..no PBR volumetrics, and the slow multiscattering.
After that comes stability issues when tweaking curves where it crashes, and is also slow, and then feature lacks like missing weight paint and emission and openGL fire and smoke representation at the same time.

If those where in place I may have purhcased it, also missing particle advection.