PDA

View Full Version : New Blog Post: OpenVDB



Julez4001
12-07-2016, 08:48 PM
Because it needs its own thread and not burden in nuggets.....

https://blog.lightwave3d.com/2016/12/openvdb-support/

Snosrap
12-07-2016, 09:35 PM
Now if only the new hypervoxels can make an effect like this. :)

jwiede
12-07-2016, 09:38 PM
I have comments on the rest, but first wanted to make a desperate plea: LW3DG, please pull that second video, ASAP, until you can fix it!

The second video looks horrible (IMO): The "fire" part looks nothing like fire at all (either from an light emission/shadowing standpoint, or from the meshing used). Look at the middle to end of the video, it looks exactly like gray smoke fading into orange smoke, and overall very non-explosion-y. Someone needs to tweak the "fire" emission and shadows, and check whether the "fire" is being imported or generated as wrong type of mesh (perhaps in the source sim itself? regardless, shouldn't be used as is, IMO). The way the initial gray kind of turns orange, and stays orange through the end just makes no sense -- almost as if the color gradient got reversed or something.

I'm really concerned that second video is going to do more damage than benefit as it currently stands, esp. with words like "explosion" so prominent in title and description.

Paul_Boland
12-07-2016, 09:58 PM
So this new OpenVDB allows you to import stuff from other packages that also use it... Perhaps I'm just not knowledgeable enough to get excited by this but...I'm not that excited by this. LOL!! I'm sure others in the know will love it. But why are these external plugs being added to Lightwave instead of having the actual FEATURES added to Lightwave??? Give us these cool smoke tools, and some awesome fluid dynamics, INSIDE LIGHTWAVE, and my jaw will hit the floor. Right now, nice to get a new blog post and the renders do look lovely, but I don't want to be looking to other packages to get these effects, I want them IN Lightwave!

samurai_x
12-07-2016, 10:22 PM
Nice feature. Render time seems a bit long.

jeric_synergy
12-07-2016, 10:26 PM
So this new OpenVDB allows you to import stuff from other packages that also use it... Perhaps I'm just not knowledgeable enough to get excited by this but...I'm not that excited by this. LOL!!
Me neither, Paul: however, this is a pipeline feature that DOES extend LW's functionality in higher-end houses. From the start, Rob P. has shown extreme concern about positioning LW in pipelines, so it maintains its foot/finger/fingernail-hold in VFX houses. I think that was a good idea: if LW plays nice, it gets to stay in the playground.

He also, of course, is more than privy to the new render engine's capabilities. Reading the tea leaves, I'm guessing he believes (with reason) that the new renderer is nifty enough to seduce more VFX customers, especially ones using this OpenVDB feature, and is making it as easy as possible for VFX houses to hold onto their LW seats.

For small and one-animator houses, it's probably no big deal, but for the upper end, mayyybe?

Anyway, I'm busy learning 3d Coat (LW friendly!) and C4D (mographariffic!).

++++++++++
just for ...whatever...



OpenVDB is an Academy Award-winning open-source C++ library comprising a novel hierarchical data structure and a suite of tools for the efficient storage and manipulation of sparse volumetric data discretized on three-dimensional grids.

jeric_synergy
12-07-2016, 10:40 PM
I'm really concerned that second video is going to do more damage than benefit as it currently stands, esp. with words like "explosion" so prominent in title and description.
OR, they could just give it a better name and thus lower expectations!

Jaqen
12-07-2016, 10:49 PM
Actually what is shown on this website is something to get excited about imo - I assume this feature will allow the import of fluids, dust particle sims etc as shown in the lead show-reel.
http://www.openvdb.org/

wingzeta
12-07-2016, 11:52 PM
"Doctor! The blog has awoken from it's coma!"

"Don't pull the life support just yet, nurse."

But seriously, a good new compatibility feature, and some communication is great.

Agree on the "explosion" having color/shading issues, but the simulation is nice.

jwiede
12-08-2016, 12:15 AM
Agree on the "explosion" having color/shading issues, but the simulation is nice.

But that's just it, from the looks of it (because of the color/shading issues, at least in part) the sim kinda looks broken (might be accurate of sim config, not enough info to tell). F.e. smoke doesn't getting hotter hanging around dispersing.

If the purpose of the video is to demonstrate a) "accurate OpenVDB import", and b) show off LW rendering of it, then if the sim looks broken and the shading looks off, I just suspect folks who'd care about it (the target potential customer, after all) aren't going to walk away with a positive impression.

I thought the first video looked fine, and was a reasonable demo of OpenVDB import. As LW3DG hinted previously OpenVDB was coming (circa March, IIRC), and the rendering of the imported data looks like the prior volumetric rendering, it lacks much in the way of "new info", but at least it looks decent for what it is supposed to be showing us.

As for the first sentence, I don't feel the apology sounds genuine, nor has there been any significant indication LW3DG intend to change their behavior. I'm sure some will say I'm being overly harsh, but IMO, given the (yet unchanging) circumstances, their apology just came across as hollow and remorseless. I'm tired of LW3DG apologies about absent communication followed by further extended absences of communication, LW3DG's done that too many times for me to believe they're actually contrite about it this time.

Considering the circumstances, and how long we've all been waiting, the new blog post just felt "empty" to me. YMMV.

m.d.
12-08-2016, 01:03 AM
The second video looks horrible (IMO): The "fire" part looks nothing like fire at all (either from an light emission/shadowing standpoint, or from the meshing used). Look at the middle to end of the video, it looks exactly like gray smoke fading into orange smoke, and overall very non-explosion-y. Someone needs to tweak the "fire" emission and shadows, and check whether the "fire" is being imported or generated as wrong type of mesh (perhaps in the source sim itself? regardless, shouldn't be used as is, IMO). The way the initial gray kind of turns orange, and stays orange through the end just makes no sense -- almost as if the color gradient got reversed or something.


Doesn't look like any emission is actually happening....just different color smoke....
I am sure emission is coming soon (pun not intended) ;D
(Have to add @NAME=HEAT in the group in convert VDB node in houdini to get it into the proper buffer...)

While your at it Newtek...consider adding VDB liquid volume import as well

50one
12-08-2016, 01:04 AM
Poof! goes the voxel.

Ztreem
12-08-2016, 01:12 AM
So this new OpenVDB allows you to import stuff from other packages that also use it... Perhaps I'm just not knowledgeable enough to get excited by this but...I'm not that excited by this. LOL!! I'm sure others in the know will love it. But why are these external plugs being added to Lightwave instead of having the actual FEATURES added to Lightwave??? Give us these cool smoke tools, and some awesome fluid dynamics, INSIDE LIGHTWAVE, and my jaw will hit the floor. Right now, nice to get a new blog post and the renders do look lovely, but I don't want to be looking to other packages to get these effects, I want them IN Lightwave!

You can just buy turbulenceFD to do it inside lightwave but with this feature you're not limited to that plugin and you also get access to the free smoke plugin callef Blender. I'm actually looking forward to bring blender smoke into LW for rendering instead of comping it in in post.

Marander
12-08-2016, 02:48 AM
The second video looks horrible (IMO)

Yeah this can be done way better in TFD within 5 minutes, it looks very fake. OpenVDB support itself is a good step.

hrgiger
12-08-2016, 02:56 AM
As for the first sentence, I don't feel the apology sounds genuine, nor has there been any significant indication LW3DG intend to change their behavior. I'm sure some will say I'm being overly harsh, but IMO, given the (yet unchanging) circumstances, their apology just came across as hollow and remorseless. I'm tired of LW3DG apologies about absent communication followed by further extended absences of communication, LW3DG's done that too many times for me to believe they're actually contrite about it this time.



I do think you're being overly harsh. As much as I agree it is to LW3DG's benefit to communicate regularly, they certainly don't owe anyone an apology for choosing to talk or remain silent. They don't even owe us a final release of LW Next. LW3DG selling us software works because its supposed to be a mutually beneficial exchange. We get new software, they get money and hopefully the cycle continues. We don't pay them for the blog, for them to answer our questions online, for them to communicate to us with anything outside of our purchase agreement or product support agreement. Again, I believe it is in their benefit to do so and that's why I think we see Lino try to answer questions as he can and for Rob to do the same, but none of that is owed to anyone. In the end, they'll put out software or not, you decide at that point if you want the software or not and that's pretty much the extent of the obligation between yourself and LW3DG. Obviously you can take the lack of communication, or the fact that a blog video isn't up to your standards or whatever into consideration on whether you purchase or not, but for me, it always comes down to 'do I want the software or not?'. I can't use a blog post in production.

As far as the explosion video goes, maybe it would be a better tactic to direct questions to Lino directly first about what he intended before suggesting there's something wrong with the sim or shading or whatever.

rustythe1
12-08-2016, 03:16 AM
I do think you're being overly harsh. As much as I agree it is to LW3DG's benefit to communicate regularly, they certainly don't owe anyone an apology for choosing to talk or remain silent. They don't even owe us a final release of LW Next. LW3DG selling us software works because its supposed to be a mutually beneficial exchange. We get new software, they get money and hopefully the cycle continues. We don't pay them for the blog, for them to answer our questions online, for them to communicate to us with anything outside of our purchase agreement or product support agreement. Again, I believe it is in their benefit to do so and that's why I think we see Lino try to answer questions as he can and for Rob to do the same, but none of that is owed to anyone. In the end, they'll put out software or not, you decide at that point if you want the software or not and that's pretty much the extent of the obligation between yourself and LW3DG. Obviously you can take the lack of communication, or the fact that a blog video isn't up to your standards or whatever into consideration on whether you purchase or not, but for me, it always comes down to 'do I want the software or not?'. I can't use a blog post in production.

As far as the explosion video goes, maybe it would be a better tactic to direct questions to Lino directly first about what he intended before suggesting there's something wrong with the sim or shading or whatever.

he did say in the blog it was a test of the actual simulation, so I don't think the rendering had much to do with it, so although it could have been presented differently to kill two birds with one stone (simulation and rendering) I think this enough to see a much bigger picture, it opens up interchange between al sorts of apps, and considering how many people here mention Houdini on the forums its a very good move for them, I think it will go a long way to clawing back any lost market to equal level software like modo, and I am guessing this is a two way street so that straight away I would assume could open up the door to external renders such as Arnold and Vray (all though I don't know the technicalities of any of that but many Renderers are listed on the openVDB site)

Sanchon
12-08-2016, 03:37 AM
I would like to import V-Ray simulation to Lightwave renderer ;) Is it can be done with upcoming Lightwave ?

jwiede
12-08-2016, 03:37 AM
I do think you're being overly harsh. As much as I agree it is to LW3DG's benefit to communicate regularly, they certainly don't owe anyone an apology for choosing to talk or remain silent. They don't even owe us a final release of LW Next.

And you're welcome to that opinion, but it is moot, because they stated the apology. Now the only relevant question is whether the apology was genuine, and even recent past behavior by LW3DG suggests an absence of contrition/remorse, thus not a genuine apology. It's not a huge deal per instance, but LW3DG repeatedly apologizing about behavior they don't intent to stop certainly doesn't improve customer relations or trust.


As far as the explosion video goes, maybe it would be a better tactic to direct questions to Lino directly first about what he intended before suggesting there's something wrong with the sim or shading or whatever.

They are the ones who put a big "explosion" label on the video and description. Regardless of whether that term matches Lino's "artistic intent", the vast majority of viewers are going to expect an explosion, and judge it by how it resembles a real explosion. Pull it down, change the title, fix the sim/render, whatever, there are a bunch of ways they could fix the issues.

S0nny
12-08-2016, 03:44 AM
Nice feature. Render time seems a bit long.

Can't see any render time, where are the infos?

It's a good features for sure, I'd say at this point that the more they open Lw to interchange, the better.

From the render quality stand point, I'm not impressed honestly, I'm not sure why someone would render out an explosion like this in Lw and not in Houdini directly. Maybe it's me, but it looks very similar to what Maya renders out directly in the viewport 2.0.

Note aside, looking at the title, it looks like Lightwave Next is confirmed as official release name.

bazsa73
12-08-2016, 04:33 AM
It can't be slower than Blenders cyclerenderer which produces zillions of tiny little noise artefacts. So I'm generally pleased with this setup, it's the artist's task to make the fire nice looking
not the developer's.

Surrealist.
12-08-2016, 05:02 AM
So this new OpenVDB allows you to import stuff from other packages that also use it... Perhaps I'm just not knowledgeable enough to get excited by this but...I'm not that excited by this. LOL!! I'm sure others in the know will love it. But why are these external plugs being added to Lightwave instead of having the actual FEATURES added to Lightwave??? Give us these cool smoke tools, and some awesome fluid dynamics, INSIDE LIGHTWAVE, and my jaw will hit the floor. Right now, nice to get a new blog post and the renders do look lovely, but I don't want to be looking to other packages to get these effects, I want them IN Lightwave!

Have a look at this:

OpenVD is actually not a plugin. Although one can be written for it.

http://www.openvdb.org/about/

Think of this as sort of Bullet for volumes. It is an open library.

To what extent they have plans to support OpenVDB as a feature in LightWave I don't know. But it would seem that giving it a nod is a big first step.

I am a huge fan of supporting libraries such as Bullet as a solution for FX. It comes with a huge advantage in my opinion. It is a great way to automatically increase your development team without the expense. These larger companies such as Dreamworks and Pixar Dysney for example, have large budgets to work with. And an incentive to spread the technology as far and wide as they can.

Ed Catmull explains it pretty well here:

https://renderman.pixar.com/view/future-of-renderman

The video on "Community"

Bottom line is, I think this is heading toward incorporating it as a feature in LightWave. That is the sense I get. And it is a very economical way to add state of the art features that are then also compatible with other pipelines.

Asticles
12-08-2016, 05:11 AM
Congrats for the new feature!

This is the right way! Lino, can you make another video with openvdb import from blender? This would be good to attract blender users.

Regards!

bazsa73
12-08-2016, 05:16 AM
Congrats for the new feature!

This is the right way! Lino, can you make another video with openvdb import from blender? This would be good to attract blender users.

Regards!
Theoretically it should work, but I also wonder how does it happen in reality. And how shader setup happens. Would be great to see a screenshot of this too.

Surrealist.
12-08-2016, 05:22 AM
Yeah by the way devs, we get it. Just so you know. The creative aspects of the shot are not the point at all. Not even worth commenting on in my opinion. The fact that this feature is added is the point. And thanks for sharing. I can see where this is going. And a smart move.

MichaelT
12-08-2016, 05:36 AM
The render is a test.. it is not really meant to be representative of anything final. The aim is to show that the integration of OpenVDB works.. and they (IMHO) do succeed in that. As for making it look awesome.. that would be the next step. That said, they can probably already set it up to look awesome.. but to say how long that would take? And btw.. the initial (default) rendering of explosions etc.. in Houdini (and TFD too for that matter) does look pretty awful. If anything, I'd say Blender is the easiest to make look good right out of the box, doesn't mean it is anywhere near being best however.

tyrot
12-08-2016, 06:28 AM
i also think .. render times very long ..

50one
12-08-2016, 06:35 AM
I would like to import V-Ray simulation to Lightwave renderer ;) Is it can be done with upcoming Lightwave ?

PhoenixFD? You can't export VDB files from PFD, just Krakatoa.

OlaHaldor
12-08-2016, 07:18 AM
Well, actually, you can.


Btw, there is a beta VDB export in the nightlies - you have to set an absolute path for the output and change the extension from aur to vdb in the edit box.

http://forums.chaosgroup.com/showthread.php?91224-PhoenixFD-3-wishlist&p=717708#post717708


Seeing there's a PFD 3 for Maya beta going on now, I'm quite interested !

lardbros
12-08-2016, 07:21 AM
Wow... didn't take long for the negativity to ensue.
Would be cool if Newtek could just post something cool without it being nitpicked.

I took the examples as not necessarily being photoreal, but show how you can do VFX work and render Houdini (or Blender) VDB sims using LW.

Looks awesome to me!!!

Spinland
12-08-2016, 07:23 AM
Heh. So, if they drop a little bit of in-progress material in response to our outcry for same, that material gets picked apart and treated like a thoroughly edited and seasoned demo.

Then some folks wonder why they don't bother.

"It only takes a few seconds to share an update!"

"Oh, that update better be vetted and polished and produced to the nines!"

M'kay. ;D

- - - Updated - - -


Wow... didn't take long for the negativity to ensue.
Would be cool if Newtek could just post something cool without it being nitpicked.

I took the examples as not necessarily being photoreal, but show how you can do VFX work and render Houdini (or Blender) VDB sims using LW.

Looks awesome to me!!!

This, in spades. Sheesh, people.

MichaelT
12-08-2016, 07:57 AM
I think this is a symptom of the silence. If they keep releasing bits like this, I think it will get much calmer.

S0nny
12-08-2016, 07:57 AM
Ok, I found the post about render times, what are you guys comparing to, to say is slow/fast? For a 1080p without noise or aliasing problems doesn't look so slow.

Spinland
12-08-2016, 08:01 AM
I think this is a symptom of the silence. If they keep releasing bits like this, I think it will get much calmer.

A more-than-fair observation. I, too, shall seek some inner Zen and refrain from yelling at the kids to get off my g-d lawn. :devil:

vncnt
12-08-2016, 08:08 AM
Glad with the new blog post.

Yeah by the way devs, we get it. Just so you know. The creative aspects of the shot are not the point at all. Not even worth commenting on in my opinion. The fact that this feature is added is the point. And thanks for sharing. I can see where this is going. And a smart move.
Very interesting to see what OpenVDB (+Blender) can do for us.

A general FluidFX plugin, for simple smoke and fluid effects, would be very welcome.

ianr
12-08-2016, 08:18 AM
Wadda dis? Him Lino him gonna up ind Smoka?

Sheesh Wadda y'know?:jam::

jboudreau
12-08-2016, 08:38 AM
I have comments on the rest, but first wanted to make a desperate plea: LW3DG, please pull that second video, ASAP, until you can fix it!

The second video looks horrible (IMO): The "fire" part looks nothing like fire at all (either from an light emission/shadowing standpoint, or from the meshing used).

I thought the same thing, this looks nothing like an explosion IMHO, But I don't think they wanted to waste a lot of time polishing it to get the point across that you can now import Open VDB in Lightwave Next.

Looking forward to testing this. Thanks for the update LW3DG

Thanks,
Jason

lardbros
12-08-2016, 08:40 AM
I think this is a symptom of the silence. If they keep releasing bits like this, I think it will get much calmer.

Totally agree... but it just means they'll share less if this is the response. They are people after all, and have feelings.

Spinland
12-08-2016, 08:53 AM
They are people after all, and have feelings.

Indeed. :jam:


https://youtu.be/Svxclyud1iA

MichaelT
12-08-2016, 08:55 AM
Indeed. :jam:


..video removed..

^^^^^^ This!!

SBowie
12-08-2016, 09:00 AM
Indeed. :jam:Totally OT, but I loved Kelly's Heroes .. some excellent stuff in there.

S0nny
12-08-2016, 09:09 AM
Lol, come on guys, if there's someone who doesn't agree with the output of the blog it doesn't mean that the overall response is negative, actually it's always been quite the opposite.
All this fuss about nothing every time someone makes some criticism or concern is just overreacting.

Spinland
12-08-2016, 09:25 AM
Totally OT, but I loved Kelly's Heroes .. some excellent stuff in there.

Heh. The inspiration for posting that clip was actually Oddball's line about "Some of these guys got sensitive feelin's!" but I couldn't find the appropriate clip so I nabbed the next best quote. :D

- - - Updated - - -


if there's someone who doesn't agree with the output of the blog...

Personally? My in-turn reaction was to what I perceived as far more harsh (and, frankly, unwarranted) than simply not agreeing. But that's just me. :D

hrgiger
12-08-2016, 09:25 AM
Note aside, looking at the title, it looks like Lightwave Next is confirmed as official release name.

I kinda doubt that. Not only was it mentioned before that it was an internal name, but I'm guessing it just has more to do with the fact that since they revealed information about the new release, they didn't have a solid release date so instead of naming it 2016 or 2017, they just referred to it as LW next.

Wickedpup
12-08-2016, 10:03 AM
Totally agree... but it just means they'll share less if this is the response. They are people after all, and have feelings.

They arenīt subjected to any more scrutany and nitpicking than any other software developer.....

mummyman
12-08-2016, 10:38 AM
Test.... it's labeled as only a test! Good, bad, ugly... it's something! Progress!

Spinland
12-08-2016, 10:51 AM
Test.... it's labeled as only a test! Good, bad, ugly... it's something! Progress!

I'm reminded of why I no longer let anyone but the director and production manager see my VFX dailies. Even when delivered with a disclaimer detailing the precise thing a clip was intended to demo I'd get crits about aspects completely out of scope. Some folks just feel the need to pick nits for picking's sake. ;D

darkChief
12-08-2016, 11:23 AM
Cooooool, not bad. Good to see Lighwave integrating open standards :)

vncnt
12-08-2016, 11:23 AM
I'm reminded of why I no longer let anyone but the director and production manager see my VFX dailies. Even when delivered with a disclaimer detailing the precise thing a clip was intended to demo I'd get crits about aspects completely out of scope. Some folks just feel the need to pick nits for picking's sake. ;D
I couldn't agree more.

S0nny
12-08-2016, 11:30 AM
Personally? My in-turn reaction was to what I perceived as far more harsh (and, frankly, unwarranted) than simply not agreeing. But that's just me. :D

Probably harsh is a better word, but I still see some reasonable criticism, not personal attack on Lw3dg guys. Anyway, no debate intended, better to get back on topic.


but I'm guessing it just has more to do with the fact that since they revealed information about the new release, they didn't have a solid release date so instead of naming it 2016 or 2017, they just referred to it as LW next.

It could be. Maybe it's just that I like Next more than 2017, it fits better to me.

ianr
12-08-2016, 11:42 AM
Totally OT, but I loved Kelly's Heroes .. some excellent stuff in there.

Now your Talking!

I don't think there's

a duff line in that.

Shear top dallar!

ianr
12-08-2016, 11:46 AM
Couldn't agree more Jeric & nicely put too.

jeric_synergy
12-08-2016, 11:53 AM
The second render is what it is: while it may not be a FIERY explosion (and it certainly is not), it is some kind of explosion: to me, it's more like some sort of detonation under a pile of powder.

The main point is that it was sim'd in Houdini, and rendered in LW. The interchangability is the point, not the convincingness (is that even a word?) of the render.

+++++

Besides Houdini (I got enough headaches), what other applications can generate and export OpenVDB simulations?

and, besides the obvious smoke, fire, and liquids, are there any surprising uses of OpenVDB functionality?

prometheus
12-08-2016, 12:04 PM
Now if only the new hypervoxels can make an effect like this. :)
meh..not really worth to go there, hypervoxels or the new volumetrics is a volumetric engine based on items, not CFD fluids, we already got a fluid engine with turbulenceFD, wether or not we would need any form of native
fluids in lightwave in addition to the turbulenceFD plugin is another story, but I suspect it is better to discuss volumetric voxels and fluids as separate tools, since they are just that.



So this new OpenVDB allows you to import stuff from other packages that also use it..., but I don't want to be looking to other packages to get these effects, I want them IN Lightwave!
Yup...you could use houdini indie for a decent buck I think, or use the free blender and paint in smoke and fire with blenders paint tool, and then use change blenders caching to openvdb, and I think it should be doable to export to lightwave.
Well Yeah, native fluids sure..but at the moment they haven gotten there, so we either get turbulenceFD for a few bucks more, remember that other main apps may cost way more with included fluids.
For a more closer and stable fluid solution, I would also like to see it natively though.


You can just buy turbulenceFD to do it inside lightwave but with this feature you're not limited to that plugin and you also get access to the free smoke plugin callef Blender. I'm actually looking forward to bring blender smoke into LW for rendering instead of comping it in in post.

Yep...I will have to check this, with this machine I am on now, I only have blender 2.76b, and I couldnīt find the parts to change point cache to openvdb in there, so I am not sure if I was guided right or if it isnīt implemented in 2.76b, will have to check a later version on another machine.
It defenitly opens up a lot of possibilities and strengthens the pipeline, at least in a way to not fall behind other software.


I do think you're being overly harsh.

well..I would agree on that it is overly harsch, the fire and smoke sample isnīt perhaps the best showcase, but I expect to see more, or have more options to control the fire shader within lightwave.

Michael

S0nny
12-08-2016, 12:04 PM
I'm reminded of why I no longer let anyone but the director and production manager see my VFX dailies. Even when delivered with a disclaimer detailing the precise thing a clip was intended to demo I'd get crits about aspects completely out of scope. Some folks just feel the need to pick nits for picking's sake. ;D

This is very common, and in all the studio I worked for this is basically a rule when delivering wips to intermediate clients and similar. The exception is if a very trusty relationship with the client is established.

Not a perfect example, but does anybody remember the Star Trek Discovery trailer? Looking at the comments about the CGI it's easy to understand why there's some risk involved in delivering some non-high standard stuff, expecially on youtube. Imagine if that trailer was labeled with Lightwave 3D, as primary software.

ianr
12-08-2016, 12:09 PM
Thank you LINO 4 the TEAM

NOW WE DON'T NEED Houdini Engine4 LW ANYMORE!!!!:beerchug::beerchug::beerchug:

prometheus
12-08-2016, 12:12 PM
+++++

Besides Houdini (I got enough headaches), what other applications can generate and export OpenVDB simulations?

and, besides the obvious smoke, fire, and liquids, are there any surprising uses of OpenVDB functionality?

BLENDER...
http://blender.stackexchange.com/questions/19562/what-voxel-format-is-used-by-blender-to-store-smoke-data

"When baking smoke Blender can save the cache as OpenVDB. Select the smoke domain, go to the physics panel, and go into the smoke cache settings. Change the file format from "Point Cache" to "OpenVDB". You can then choose your compression and data depth."

hrgiger
12-08-2016, 12:15 PM
Besides Houdini (I got enough headaches), what other applications can generate and export OpenVDB simulations?

?

Blender and Realflow are the few I know of other than Houdini.

prometheus
12-08-2016, 12:19 PM
From a staff member at planetside (terragen)...

Question..

Export Clouds to OpenVDB
Ŧ on: January 26, 2016, 01:43:38 PM ŧ
Feature request. Would be very helpful. Thanks.

"
Re: Export Clouds to OpenVDB
Ŧ Reply #1 on: January 26, 2016, 05:31:40 PM ŧ
This is definitely something we are considering for the future. What would your specific use be?

- Oshyan
"

Spinland
12-08-2016, 12:19 PM
The second render is what it is: while it may not be a FIERY explosion (and it certainly is not), it is some kind of explosion: to me, it's more like some sort of detonation under a pile of powder.

My take was akin to how TFD works: you run the actual fluid calcs and get one kind of visual, and that does depend on which channels you have enabled, but after the fact is when you set up the render parameters and the same fluid motion becomes smoke, flames, or what-have-you. LW was showing their ability to import the calcs at all, not what can be done afterwards to tweak the render.

Chrusion
12-08-2016, 12:19 PM
Interesting... a technology with a name acronym that has no definition and no meaning.

From: http://www.openvdb.org/documentation/doxygen/faq.html


What does "VDB" stand for?

Over the years VDB has been interpreted to mean different things, none of which are very descriptive: "Voxel Data Base", "Volumetric Data Blocks", "Volumetric Dynamic B+tree", etc. In early presentations of VDB, we even used a different name, "DB+Grid", which was abandoned to emphasize its distinction from similarly named, but different, existing sparse data structures like DT-Grid or DB-Grid. The simple truth is that "VDB" is just a name. :-)

Well, that was helpful now, wasn't it. hehe I'll go with Volumeric DataBase, K?

Spinland
12-08-2016, 12:22 PM
Probably harsh is a better word, but I still see some reasonable criticism, not personal attack on Lw3dg guys. Anyway, no debate intended, better to get back on topic.

Agreed. My upbringing included internalizing the old saw about attracting more flies with honey than with vinegar and, despite some failures in resolve when I lose my temper, I see no discrepancy between being plain-spoken and to the point and yet still being positive in outlook and respectful of the feelings of others. Different folks have different emotive styles, I get that. Again, that's just me.

m.d.
12-08-2016, 12:31 PM
Blender and Realflow are the few I know of other than Houdini.

can add krakatoa and maya to that list, as well as max using PheonixFD, or FumeFX.....

Basically pretty well every fluid/liquid simulator now exports VDB...with the exception of Turbulence, and support is coming in version 2

50one
12-08-2016, 12:47 PM
Thank you LINO 4 the TEAM

NOW WE DON'T NEED Houdini Engine4 LW ANYMORE!!!!:beerchug::beerchug::beerchug:

You really ave no idea what Houdini engine really is.

tyrot
12-08-2016, 12:52 PM
You really ave no idea what Houdini engine really is.

+1000

mummyman
12-08-2016, 01:01 PM
I'm reminded of why I no longer let anyone but the director and production manager see my VFX dailies. Even when delivered with a disclaimer detailing the precise thing a clip was intended to demo I'd get crits about aspects completely out of scope. Some folks just feel the need to pick nits for picking's sake. ;D

Ha! Yup

bazsa73
12-08-2016, 01:03 PM
I'm reminded of why I no longer let anyone but the director and production manager see my VFX dailies. Even when delivered with a disclaimer detailing the precise thing a clip was intended to demo I'd get crits about aspects completely out of scope. Some folks just feel the need to pick nits for picking's sake. ;D

Even if you explain in detail to seemingly intelligent people they look for reassurance: But it will look better right? I mean in the end these wireframes whatnots, we won't see these in the final right?

Surrealist.
12-08-2016, 01:32 PM
All that is true. But in fairness. That is not what he was on about in my opinion. But that is his fight. I just think this is going in another direction that is not all that fair to what he actually said as I interpreted it. Don't be some damn sensitive to critique!

The test was great, It showed what it needed to. Not everyone agrees on the quality.

Moving on?

cove
12-08-2016, 02:51 PM
The fact that Rob and Lino have decided to show us the new OpenVDB feature
means simply that they have this new feature ready for the "NEXT" release and decided to show us all.
This is in keeping with previous feature releases.
When they have a new feature that will be in "NEXT" to show then they show it.

As for the quality of the 2 video examples.
--------------------------------------------
LightWave Next: OpenVDB Smoke Test.

This was most probably fairly easy to create as it is simply a smoke effect.
--------------------------------------------
LightWave Next: OpenVDB Explosion Test.

Some of you have criticised some aspects of the way the explosion accures/looks eg. orange color continuing to spread over the whole plume of smoke.

Fair enough its not perfect but i asked myself how much time did Lino have
to create this effect.
If his remit was to create a basic example of an explosion to illustrate the new OpenVDB feature then that is exactly what he has done.

Although a lot of us none professionals will not have a use for this important new feature i understand why Rob has posted info at this time.

Hopefully there will be a few more new features ready to post soon.

bazsa73
12-08-2016, 02:59 PM
Do we know how long did it take to render, because someone posted earlier that it took way too long. 10 minutes per frame? Or more?

S0nny
12-08-2016, 03:50 PM
Do we know how long did it take to render, because someone posted earlier that it took way too long. 10 minutes per frame? Or more?

I asked the same, the answer is in the 'nuggets' thread: Lino stated 4 minutes/frame max for the first video and 7 max the second.

jwiede
12-08-2016, 04:43 PM
I asked the same, the answer is in the 'nuggets' thread: Lino stated 4 minutes/frame max for the first video and 7 max the second.

Considering all the things that weren't present (f.e. no emissive enough materials to cause strong self-shadowing), those render times seem pretty long.

tyrot
12-08-2016, 05:03 PM
that is my concern too... JW how you compare those numbers with C4D...

MichaelT
12-08-2016, 05:20 PM
Since we have no idea how the scene is set up, it seems to me that debating if it is too long or not.. is a tad moot.

lino.grandi
12-08-2016, 05:20 PM
Considering all the things that weren't present (f.e. no emissive enough materials to cause strong self-shadowing), those render times seem pretty long.

Render times are very good indeed. How can you judge render times not knowing anything about the OpenVDB size, volumetric settings and lighting? You don't even know how many lights have been used, if they are multisampled lights or not.
You should really try to load a well defined OpenVDB file in another application able to render it, and try to get a clean render. Possibly also using an HDRI for lighting, because that's something I've used for both renders.

To be clear, I've attached a render of the same simulation performed on the same PC (see render time on the image).

Snosrap
12-08-2016, 05:26 PM
Thanks Lino - I'm loving it. Now give it to me!

Spinland
12-08-2016, 05:42 PM
Don't be some damn sensitive to critique!

I am certainly not sensitive to such; but, I personally draw a line between critique and carping. YMMV. :beerchug:

bobakabob
12-08-2016, 06:14 PM
The still image looks brilliant, Lino and the whole point of the blog post is to show the potential of the tech and LW plugging into other apps. The cynics here would be the first to whine if LWNext was Maya + Zbrush + Houdini on steroids.

jasonwestmas
12-08-2016, 06:53 PM
Creating powerful importers and exporters have been a focus for lightwave for a while now. This makes sense. Thanks!

leandropedrouzo
12-08-2016, 09:39 PM
Since I read some folks saying that they don't want importers but the functionality in LW I want to share a personal experience with you guys.
A few weeks ago I got a call from a regular client. They wanted to send me an alembic file for me to try. I asked "what's in the file? Is it a fluid?". (It was a project where a Houdini guy was gonna make some fluid simulations and I was gonna make the rest) They said "yes". "I'ts not gonna work" I said "not with changing topology". They couldn't understand it. "But it worked here in Maya and 3DMax". I explained that we had the option of exporting obj sequence, but it felt like I was stuck 10 years ago.
IMHO importers are a MUST today, specially when you have to work with others.
If your software of choice gets in the way instead of help you, most likely it's the program fault. This feels like a much needed step in the right direction. The 9 months gap in communication is a different matter.

m.d.
12-08-2016, 09:57 PM
Since I read some folks saying that they don't want importers but the functionality in LW I want to share a personal experience with you guys.
A few weeks ago I got a call from a regular client. They wanted to send me an alembic file for me to try. I asked "what's in the file? Is it a fluid?". (It was a project where a Houdini guy was gonna make some fluid simulations and I was gonna make the rest) They said "yes". "I'ts not gonna work" I said "not with changing topology". They couldn't understand it. "But it worked here in Maya and 3DMax". I explained that we had the option of exporting obj sequence, but it felt like I was stuck 10 years ago.
IMHO importers are a MUST today, specially when you have to work with others.
If your software of choice gets in the way instead of help you, most likely it's the program fault. This feels like a much needed step in the right direction. The 9 months gap in communication is a different matter.

Had experiences like that.....

From my point of view, liquid sim imports are more important and in demand.

Smoke/VDB....Octane users have access to that right now. Great that lightwave has it native....
Alembic fluids/VDB fluids.....no go. Not with any plugin outside realflow, and despite the philosophy of strong interchange.....one of the few apps that cant natively (perhaps the only one?)

I am sure they are all on the way, and this is a welcome step

Surrealist.
12-08-2016, 11:32 PM
I am certainly not sensitive to such; but, I personally draw a line between critique and carping. YMMV. :beerchug:

Yeah I agree. Just seemed the responses were getting more indirect rather than direct. And more of an unfair characterization rather than a point by point rebuttal.

Characterizations come off as sensitive to me (and personal). Point by point rebuttal nor matter how harsh, I think is a better - and much more fair approach.

Just my opinion. Carry on.

ncr100
12-09-2016, 01:29 AM
Render times are very good indeed. How can you judge render times not knowing anything about the OpenVDB size, volumetric settings and lighting? You don't even know how many lights have been used, if they are multisampled lights or not.
You should really try to load a well defined OpenVDB file in another application able to render it, and try to get a clean render. Possibly also using an HDRI for lighting, because that's something I've used for both renders.

To be clear, I've attached a render of the same simulation performed on the same PC (see render time on the image).

What constitutes the substance of the smoke - voxels? Is it what the volumetric renderer does? An OpenVDB is just a file that the new renderer is capable of rendering?

Surrealist.
12-09-2016, 01:48 AM
http://www.museth.org/Ken/OpenVDB.html

I would think to use VDB data that they would have had to implement the VDB data structure in LightWave. So that i effect they are actually rendering VDB data. That is the whole point of having and using the data structure. Its unique way of storing and using the data. Which implies tools must be developed to access and manipulate as well as render this data.

Ztreem
12-09-2016, 02:32 AM
that is my concern too... JW how you compare those numbers with C4D...

As far as I know C4D don't import OpenVDB nativly...

bazsa73
12-09-2016, 02:53 AM
In Blender that quality what Linow has shown to us in his latest post takes 7 or more minutes in HD, rather more and there is a big chance for noise unless you push AA and other render values to the sky. Look at these tests: https://vimeo.com/171076042

lino.grandi
12-09-2016, 04:05 AM
In Blender that quality what Linow has shown to us in his latest post takes 7 or more minutes in HD, rather more and there is a big chance for noise unless you push AA and other render values to the sky. Look at these tests: https://vimeo.com/171076042

Amazing tests!

About render time, I can't stress enough how settings can be important, as well as the file used. ;)

Surrealist.
12-09-2016, 04:15 AM
Thanks for the sneak peek on the interface. (whether final or not) And also great to see this tech coming to LW. Also.... nice render for so fast. :) In general, smoke looking real good in my opinion.

Spinland
12-09-2016, 04:28 AM
Just my opinion. Carry on.

Fully agreed. :thumbsup:

Every4thPixel
12-09-2016, 04:46 AM
I like this! Keep up the good work LW3DG! The colors of the second movie are a bit weird but I know this is just a test. I hope shading VDB files will be easy in the final release.

rustythe1
12-09-2016, 04:51 AM
Thanks for the sneak peek on the interface. (whether final or not) And also great to see this tech coming to LW. Also.... nice render for so fast. :) In general, smoke looking real good in my opinion.

yep, blows HV out of the water, I see fun times ahead, but can I stand the wait, nooooooooooooooooooooooooo.

Ztreem
12-09-2016, 05:29 AM
Amazing tests!

About render time, I can't stress enough how settings can be important, as well as the file used. ;)

Looks great! Thanks for sharing.

bazsa73
12-09-2016, 06:01 AM
Amazing tests!

About render time, I can't stress enough how settings can be important, as well as the file used. ;)

Great to hear it!
I share some of the scenes. Not organized, just rar-ed the last ones, maybe it helps your job.
http://www.visionsupreme.net/blenderexplo/

lino.grandi
12-09-2016, 07:57 AM
Great to hear it!
I share some of the scenes. Not organized, just rar-ed the last ones, maybe it helps your job.
http://www.visionsupreme.net/blenderexplo/

For sure it does! Thank you so much!

MAUROCOR
12-09-2016, 10:27 AM
Great to hear it!
I share some of the scenes. Not organized, just rar-ed the last ones, maybe it helps your job.
http://www.visionsupreme.net/blenderexplo/

Congratulation, Peter! ^ THIS is to add some positive action to the theme!

m.d.
12-09-2016, 11:10 AM
Amazing tests!

About render time, I can't stress enough how settings can be important, as well as the file used. ;)

Ya that's a rather small step size.....lots of details

prometheus
12-09-2016, 02:22 PM
In Blender that quality what Linow has shown to us in his latest post takes 7 or more minutes in HD, rather more and there is a big chance for noise unless you push AA and other render values to the sky. Look at these tests: https://vimeo.com/171076042

Yes..I have seen those before, great tests Peter, did you use manual gradient node coloring and tweaked that manually? or did you use the blackbody node in blender?
Edited...forgot you posted some scene samples, I may take a look at those.

And the renders by Lino with openVDB seem pretty fast too, exciting.

Michael

wyattharris
12-09-2016, 03:46 PM
Drat, two days old. Where have I been?
Glad to see the update fellas. Interesting tech. Was definitely a good read and update.
Hopefully it gets followed by a few more... ;)

Matt
12-09-2016, 04:03 PM
the shading looks off

Just needed a better gradient setup is all. No need to be all doom and gloom!

135276

hrgiger
12-09-2016, 04:05 PM
Very nice Matt. Thanks for the render. Would love to see it animated.

Matt
12-09-2016, 04:07 PM
Very nice Matt. Thanks for the render. Would love to see it animated.

Yeah I agree, just don't have time to tie my machine up at the moment.

Matt
12-09-2016, 05:02 PM
Very nice Matt. Thanks for the render. Would love to see it animated.

Okay, here's a small one, different VDB file out of Houdini. Only had 42 frames generated.

135278

135277

bobakabob
12-09-2016, 05:19 PM
Way beyond hypervoxels. The new volumetrics are looking great.
This might sound like a silly question... Does the glow illuminate a scene with radiosity applied?

hrgiger
12-09-2016, 05:25 PM
Again, thank you Matt, that looks lovely.

- - - Updated - - -


Way beyond hypervoxels. The new volumetrics are looking great.
This might sound like a silly question... Does the glow illuminate a scene with radiosity applied?

I'm sure it does, it uses emission which is PBR version of luminosity.

Ztreem
12-09-2016, 05:43 PM
Just needed a better gradient setup is all. No need to be all doom and gloom!

135276

Nice! Thanks for sharing...

Matt
12-09-2016, 05:49 PM
Way beyond hypervoxels. The new volumetrics are looking great.
This might sound like a silly question... Does the glow illuminate a scene with radiosity applied?

Yes it does.

Surrealist.
12-09-2016, 06:00 PM
Very nice but you had me at OpenVDB and Houdini. :)

Seriously. Thanks for the share!

Paul_Boland
12-09-2016, 09:07 PM
Just want to say, sorry if my post at the start of this thread came across badly. I was just trying to say, I'm more interested in what I can do with Lightwave itself, and not want I can do with other software packages and Lightwave. I see a lot of excited people here about this so thumbs up!!

Farhad_azer
12-10-2016, 03:37 AM
I am exactly like you Paul_Boland and plugins and alike do not interest me.

There are lots of dark corners in LW that i have not been able to find and experiment with.

Marander
12-10-2016, 04:02 AM
Okay, here's a small one, different VDB file out of Houdini. Only had 42 frames generated.

Looking good, the ability to render openVDB this way (and as seen in Lino's first smoke video) is indeed a big step ahead for LW, congrats!

Maybe you could render some frames more and add it to the already posted videos.

Surrealist.
12-10-2016, 05:33 AM
Just want to say, sorry if my post at the start of this thread came across badly. I was just trying to say, I'm more interested in what I can do with Lightwave itself, and not want I can do with other software packages and Lightwave. I see a lot of excited people here about this so thumbs up!!

Yes. But don't forget that with this Open VDB in place, it means it is implemented to some degree in LightWave. Which means a new set of volumetric tools and fluid tools that are already proven state of the art can now be developed in LightWave. That is my understanding of the significance of this beyond a mere plugin to import the data. To play and render the data, that means the system exists to some level within LightWave. I don't think it is converting it so something else. That defeats the purpose- I think. Though I could be wrong. But the interface does seem to show it is reading an openVDB file.

I think the data system needs to be there to play and take advantage of this. The next logical step is developing editing tools for the data so that it can all originate in LW. That is how I read into it anyway. I can not say I know what their plans are. But they are in the process of revamping the underlying data structure of LightWave anyway, so this makes sense.They are replacing the mesh system. And this would allow them to replace the old volumetric system - eventually I suppose.

inkpen3d
12-10-2016, 05:42 AM
In the paper that Richard Culver kindly provided a link to in the nuggets thread (see http://www.museth.org/Ken/OpenVDB_files/Museth_TOG13.pdf) the caption to Fig 1 states "The final animated sparse volumes ... are rendered using a proprietary renderer that exploits VDB’s hierarchical tree structure."

Which begs the question: Has the new LW render engine been fully optimised to efficiently handle openVDB data structures?

Regards,
Peter

Julez4001
12-10-2016, 07:09 AM
Yeah I agree, just don't have time to tie my machine up at the moment.


I have 5 Dual Z280 XEON CPU @ 3.5 GHZ sitting idle at work doing nothing, sign me up for Beta use and I will render out all you ever want.

matter of fact, let me take over the blog post page and I have this forum head spinning around you never have to worry about the griping.
:hey:

Surrealist.
12-10-2016, 07:11 AM
QUOTE=inkpen3d;1491930]In the paper that Richard Culver kindly provided a link to in the nuggets thread (see http://www.museth.org/Ken/OpenVDB_files/Museth_TOG13.pdf) the caption to Fig 1 states "The final animated sparse volumes ... are rendered using a proprietary renderer that exploits VDB’s hierarchical tree structure."

Which begs the question: Has the new LW render engine been fully optimised to efficiently handle openVDB data structures?

Regards,
Peter[/QUOTE]



In that case it wold have to. If it is reading the file. That is the whole point I would think. Or you would not be calling it OpenVDB. Right? So when they say they now support OpenVDB, it means the LightWave renderer has been optimized to read/use the data. The other option would be convert it to a mesh or some other data in Lightwave. This is all above my pay grade.... lol Just trying to look at it logically.

prometheus
12-10-2016, 07:13 AM
Just needed a better gradient setup is all. No need to be all doom and gloom!

135276

good detail there, great..and we get the point of adjusting shaders I reckon.
What is to be seen is how illumination methods can be employed, rayleigh values etc.

Spinland
12-10-2016, 07:18 AM
Just needed a better gradient setup is all. No need to be all doom and gloom!

Woo hoo! Some of us never doubted, Matt. We understand the difference between the simulation and the shader. ;)

Julez4001
12-10-2016, 07:57 AM
Thanks Lino for interface sneak peak.

See, they are sharing now after months of "nuggets".... just exude positive energy and see where it goes.

Julez4001
12-10-2016, 08:02 AM
If the render times are super fast and produce superb quality, LW may be able to re-capture a lot of the TV VFX that have went to Maya shops (Flash, Supergirl,League and Netflix Marvel shows)

The new STAR TREK series would definitely benefit with this development.

inkpen3d
12-10-2016, 08:35 AM
In that case it wold have to. If it is reading the file. That is the whole point I would think. Or you would not be calling it OpenVDB. Right? So when they say they now support OpenVDB, it means the LightWave renderer has been optimized to read/use the data. The other option would be convert it to a mesh or some other data in Lightwave. This is all above my pay grade.... lol Just trying to look at it logically.

No, I have to respectfully disagree - just because your software can read a given data set and use it, doesn't automatically mean that your code is optimised to do so in the most efficient way. So you could make bold claims to support openVDB, but in reality you are not doing so in the most effective way. For instance, your code might be navigating and manipulating the data set in a completely naive and inefficient manner when compared to code written by someone who has a deep understanding of the openVDB data structures and who can write optimised code to efficiently handle this data. The end result being that, in comparison, your code exhibits lengthy render times.

Regards,
Peter

MichaelT
12-10-2016, 09:33 AM
The renderer have nothing to do with the data from OpenVDB. You collect the data you need, then compile it into a form that is usable by the renderer. How efficient that process is, is of course a different matter. But is in any case, greatly affected by the amount of data you have to process from the database. It doesn't stop you from ignoring parts of that data however, to speed things up. Some of the data can also be predicted.

bazsa73
12-10-2016, 11:36 AM
amazingly cool these flame renders are

Surrealist.
12-10-2016, 11:46 AM
I was not intending to start up a technical debate none of us can win.

That is impossible and not even worth arguing at this stage which such absoluteness. We will all know when they release more data and we test render times.

In the mean time this is the information we have to work from.


ex·ploit
verb
3rd person present: exploits
ikˈsploit/

1.
make full use of and derive benefit from (a resource).
"500 companies sprang up to exploit this new technology"
synonyms: utilize, harness, use, make use of, turn/put to good use, make the most of, capitalize on, benefit from; informalcash in on
"we should exploit this new technology"



http://www.openvdb.org/


OpenVDB is an Academy Award-winning open-source C++ library comprising a novel hierarchical data structure and a suite of tools for the efficient storage and manipulation of sparse volumetric data discretized on three-dimensional grids. It is developed and maintained by DreamWorks Animation for use in volumetric applications typically encountered in feature film production.

So LightWave render engine is also exploiting this library along with other competing render options such as Renderman, VRay and Arnold and including DCC/FX apps Like Modo, Realflow and Houdini.

I don't think there is much else to go on at this point.

MichaelT
12-10-2016, 05:19 PM
Fine... but unless they've designed their rendering engine in a very unusual way. Any renderer will have its own data system, from which it draws its images. OpenVDB will (in general terms) only be another source from which the final data is assembled, before being pushed to the renderer. Designing it differently would mean you lock in the code.. making it very difficult to change to another system later. I know, I did that myself.. but only when coding demos on Amiga :)

Surrealist.
12-11-2016, 02:15 AM
Yes of course. But in this case I think the app has to recognize the data as well. In Maya for example you can render vdb in both Arnold and Renderman. In Maya you can further modify the data with nodes as far as I understand, and visualize that apart from rendering in OGL.

https://www.youtube.com/watch?v=R8Y9VqU_YTI

And here is a video showcsing the plugin in more detail.

https://vimeo.com/73294767

Clearly this is reading and using the VDB data. And that is the entire point of having the data structure. To say it is only a saving and loading application is oversimplifying it greatly.


So my point is to use Open VDB data it first has to be opened and visualized into the host app. And there the data can be manipulated. Then it is sent to the render engine which must also be able to read (or convert) that data. So that means in the case of 3P rendering, there are two separate implementations. But both the app's implementation and the render implementation have to be present. Then I suppose shading networks as well which would have to be another use/conversion of the data.


Again this is what it says about Open VDB.


open-source C++ library comprising a novel hierarchical data structure and a suite of tools for the efficient storage and manipulation of sparse volumetric data discretized on three-dimensional grids.

How far it goes in being able to manipluate thaty data I suppose is going to change from app to app. The Maya plugin looks as if it is accessing the various attributes of the file system.

In Houdini it looks fairly extensively incorporated into the creation tools as well as shading and rendering:

http://www.openvdb.org/download/openvdb_houdini_2015.pdf

How far they will take this in LightWave, we don't really know yet. That is the part I think I going out on a limb about. The rest of it, as far as I understand is fairly accurate. If a LW developer wants to correct me. Please do.

MichaelT
12-11-2016, 11:59 AM
Hmm, I think we look at 'renderer' differently. Visualizing data in the IDE is one thing, but the renderer (as in the part doing the final render) is usually its own thing though. I think we talked beside each other in this case :).

hrgiger
12-11-2016, 01:25 PM
We could just wait until lw next and see for ourselves. Just a thought.

jeric_synergy
12-11-2016, 01:46 PM
Sooooooo, if we're not importing a mesh (apparently) it's really just a way to import particles, correct?

Nothing wrong with that, no? And then the new renderer will do some HV-esque magic to convert those particles to visible/renderable volumes?

What with all the verbiage being expended here, I'm suspecting it's more than just that.

hrgiger
12-11-2016, 03:55 PM
OpenVDB is not importing just particles, its importing volumes (Hypervoxels can also be volumes)

You can read on it here: http://www.openvdb.org/about/

m.d.
12-11-2016, 04:31 PM
VDB holds particles, volumes and meshes...either converting or internally holding all as a volume structure.

The implementation in the host software can either mesh it into polygons for render or openGL (liquids for example)... or output as native volume straight into the render.

Right now all we are seeing from LW is volumes.

MichaelT
12-11-2016, 06:35 PM
I am fairly certain they can use most (if not all) of what OpenVDB contains.

m.d.
12-11-2016, 11:31 PM
I am fairly certain they can use most (if not all) of what OpenVDB contains.

Can is one thing....right now with the examples and interface shown, it appears like only volumes

Under the VDB object properties that Lino showed you see, emission, scattering and absorption values.... the 3 values you need to describe typical volumetric shading.....that along with step size, which is a volume only related value and has nothing to do with a mesh....along with the absence of any meshing parameters at all, leads one to assume this is a volume only interface.
And the title "VDB Object"....not "VDB Volume" or "VDB mesh" leads you to assume it is the only VDB interface.

There could be things buried in the node editor but as of right now there is no evidence to show any use of openVDB meshing

But....I'll happily be wrong:)

jeric_synergy
12-12-2016, 12:43 AM
OpenVDB is not importing just particles, its importing volumes (Hypervoxels can also be volumes)

You can read on it here: http://www.openvdb.org/about/
Can someone translate "non-manifold geometry" for me?

EDIT: AND "divergence-free advection". <--Is that last word even real? :eek:

Surrealist.
12-12-2016, 12:56 AM
There could be things buried in the node editor but as of right now there is no evidence to show any use of openVDB meshing

But....I'll happily be wrong:)



Yeah I would say no too. But did you have a look at that Maya plugin? It seems to have access to the VdB data itself in the interface. I am not sure what you are referring to when you say volume data. I think the definition of that changes when talking about vdb. I mean,say compared to a Maya volume or a LightWave hypervoxel for that matter. It looks to me like an entirely different way to access volume data.

Here is the link again.

https://vimeo.com/73294767

I encourage anyone who is interested in this to check it out. Visually it helps you understand what kind of data seems to be present within the vdb file when accessed and viewed within the 3D app.

There are some similarities though. In Maya a volume can have resolution. But it seems completely different here.


Will be interesting to see how it is implemented in LightWave.

MichaelT
12-12-2016, 01:45 AM
Can someone translate "non-manifold geometry" for me?

EDIT: AND "divergence-free advection". <--Is that last word even real? :eek:

It means geometry that cannot be unfolded. Meaning the geometry cannot be laid out flat, and still be connected by an edge.
For instance, a 'T' shape with a depth of 1 unit :)

As for advection.. if you have a liquid cooler for your CPU, the heated water is cooled in a fin structure (typically aluminium) outside the
case. That transfer of heat from the liquid to the fins and out is called advection. As for divergence free, it means the volume remains the
same. As in it doesn't shrink or grow in volume.

bazsa73
12-12-2016, 02:01 AM
Can someone translate "non-manifold geometry" for me?

EDIT: AND "divergence-free advection". <--Is that last word even real? :eek:

manifold geometry example
http://help.autodesk.com/cloudhelp/2015/ENU/MayaLT/images/GUID-7368F973-3DE2-4763-BAE2-0250359B7BCF.png

jeric_synergy
12-12-2016, 03:00 AM
As for advection.. if you have a liquid cooler for your CPU, the heated water is cooled in a fin structure (typically aluminium) outside the
case. That transfer of heat from the liquid to the fins and out is called advection.
Huh. I would have called that "conduction". Does it denote >1 material or something?

m.d.
12-12-2016, 03:05 AM
It means geometry that cannot be unfolded. Meaning the geometry cannot be laid out flat, and still be connected by an edge.
For instance, a 'T' shape with a depth of 1 unit :)

As for advection.. if you have a liquid cooler for your CPU, the heated water is cooled in a fin structure (typically aluminium) outside the
case. That transfer of heat from the liquid to the fins and out is called advection. As for divergence free, it means the volume remains the
same. As in it doesn't shrink or grow in volume.

Advection in this case is data flow between voxels (level sets)
Density, curl, velocity, particles, points.... dozens of possible values to 'advect'....including temperature

Turbulence FD famously promised LW users 'particle advection' but only delivered to C4d

MichaelT
12-12-2016, 05:14 AM
Advection in this case is data flow between voxels (level sets)
Density, curl, velocity, particles, points.... dozens of possible values to 'advect'....including temperature

Turbulence FD famously promised LW users 'particle advection' but only delivered to C4d

I'm not entirely sure they actually can have advection in the current version of LW. At least not in TFD itself. Maybe if they create another plugin that can manipulate LW particles/objects by using the TFD cache perhaps? But even then I am guessing. LW particles just isn't designed for ad hoc interaction when containers meet (or objects entering the container) a real shame actually. Something I hope they fix.

Thomas Helzle
12-12-2016, 05:36 AM
Interesting.
I'm using Houdini mostly these days and this looks like a welcome addition to Lightwaves arsenal.

The way you can work and model with VDB in Houdini is quite fascinating even outside of classical volume stuff:
http://www.entagma.com/houdini-boolean-volume-denting/#more-73

Now somebody has to write a fully transparent Houdini to Lightwave connection... :-)

Cheers,

Tom

mummyman
12-12-2016, 08:31 AM
Interesting.
I'm using Houdini mostly these days and this looks like a welcome addition to Lightwaves arsenal.

The way you can work and model with VDB in Houdini is quite fascinating even outside of classical volume stuff:
http://www.entagma.com/houdini-boolean-volume-denting/#more-73

Now somebody has to write a fully transparent Houdini to Lightwave connection... :-)

Cheers,

Tom

Wow... this looks very cool. (tedious for node idiots like me)

vbk!!!
12-12-2016, 11:24 AM
Thanks Newtek to add this features to Lightwave ! It's a really good news.
And the render time of the PBR looks pretty decent to me. I don't think Houdini/Mantra is able to render that fast the test Lino did.

Now my question is :
There is a plan to get motion blur volumetric ? In other terms, will Lightwave be able to use velocity field and/or "geometric motion blur" ?

Every4thPixel
12-12-2016, 11:31 AM
Okay, here's a small one, different VDB file out of Houdini. Only had 42 frames generated.

135278

135277

Look at those shadows! I love them! I can't wait to play around with the new LW. It would actually come in handy right now. I need to make a photo realistic 3D sponge... Now I'm here..., does anyone have any tips?

jwiede
12-13-2016, 05:21 PM
Amazing tests!

About render time, I can't stress enough how settings can be important, as well as the file used. ;)

Lino, in the UI shown, there are OpenVDB sections for emission, scattering, and absorption. Is that just for prototyping, or is that how OpenVDB surfacing will be handled, at least in the coming release? Also, is the node editor shown for masking/modifying the OpenVDB data itself, or is that part of the surfacing options as well?

m.d.
12-13-2016, 06:34 PM
good questions....

step size is an irrelevant parameter if it isn't a volume however...so it looks like that specific menu is volume only and not mesh, hopefully there is another object type listed as VDB mesh in that dropdown :)

vbk!!!
12-14-2016, 07:18 AM
AFAIK, VDB mesh doesn't exist in itself. But VDB library provides tools to manipulate sparse volume and convert it to mesh.
Now, regarding the VDB panel, I can't see any place to load/map velocity fields ... :(

creacon
12-14-2016, 08:10 AM
OpenVDB can take a bunch of particles, convert it to a levelset, apply filters on that levelset (or join, subtract different levelsets) and then convert the resulting levelset to a mesh.
But that's not the part that Lino is showing, he's showing how LW can access the data in an OpenVDB grid and render that data. Probably raymarching through the OpenVDB data.

creacon



AFAIK, VDB mesh doesn't exist in itself. But VDB library provides tools to manipulate sparse volume and convert it to mesh.
Now, regarding the VDB panel, I can't see any place to load/map velocity fields ... :(

m.d.
12-14-2016, 09:41 AM
AFAIK, VDB mesh doesn't exist in itself. But VDB library provides tools to manipulate sparse volume and convert it to mesh.
Now, regarding the VDB panel, I can't see any place to load/map velocity fields ... :(

Yes it's all volumes, the point is the entirety of this panel is devoted to volumetric (raymarching) VDB rendering

To do VDB mesh rendering we should have a lot more mesh related controls, as it's very likely lightwave will still have to convert VDB volumes to polygon for mesh renders.

creacon would know more of the nitty gritty.... as so far as we know he (and possibly dark chief)is they only one to implement any kind of vdb meshing in lightwave

Velocity fields are already contained in the VDB level sets, and LW should be reading that for MB.... not sure why you would want to override the existing ones unless you forgot to cache them. Hopefully in the nodal panel we could manipulate the various fields.....but that's asking a lot, and personally I doubt I would ever need to manipulate the velocity field....just need a read node to drive (future) particle interactions. Of course there would be interesting things you could do with full nodal control of all the fields, but that is more Houdini's wheelhouse.

pinkmouse
12-14-2016, 09:51 AM
If you don't ask, you don't get! ;)

m.d.
12-14-2016, 10:05 AM
Step 1: implement an undo system

Step 2: dethrone Houdini

:)

darkChief
12-14-2016, 12:20 PM
creacon would know more of the nitty gritty.... as so far as we know he (and possibly dark chief)is they only one to implement any kind of vdb meshing in lightwave



Not "possibly", most definitely. Will post a Lightwave integration preview video later today, showing the interface and some of the meshing previewing in the viewport, on my thread in the third party sub forum.

m.d.
12-14-2016, 12:37 PM
Not "possibly", most definitely. Will post a Lightwave integration preview video later today, showing the interface and some of the meshing previewing in the viewporf, on my thread in the third party sub forum.

Excellent news :)

jwiede
12-14-2016, 01:38 PM
Velocity fields are already contained in the VDB level sets, and LW should be reading that for MB

Another good question (presuming the UI shown is "customer-intended", ofc), though I'd understand if MB had to wait for a later release as well. Also would be interesting to know if we'd have any abilities to adjust the scattering algorithm used (f.e. depending on scale of particles, IOR)?

One thing that I don't see (in UI shown) is much in the way of control over duration/timing, though I suppose that could be handled in the node editor, or perhaps the file selector has some onboard ability to control OpenVDB frames used? IMO, customer ability to adjust OpenVDB timing beyond just a simple 1:1 relationship, and some ability to select/adjust start and end (versus just using all present), are both essential. In that regard (IMO), OpenVDB is no different than any other imported animation sequence. Tweaks like non-linear timing interpolation, etc. would also be nice, but I don't see them as fundamentally essential.

Ofc, if the UI shown isn't "customer-intended" then none of this matters much.

vbk!!!
12-14-2016, 02:53 PM
Yes it's all volumes, the point is the entirety of this panel is devoted to volumetric (raymarching) VDB rendering
Velocity fields are already contained in the VDB level sets, and LW should be reading that for MB.... not sure why you would want to override the existing ones unless you forgot to cache them. Hopefully in the nodal panel we could manipulate the various fields.....but that's asking a lot, and personally I doubt I would ever need to manipulate the velocity field....just need a read node to drive (future) particle interactions. Of course there would be interesting things you could do with full nodal control of all the fields, but that is more Houdini's wheelhouse.

I was thinking the way Maya works. Last time I use Maya you had to declare the velocity field in a similar panel to allows you to use geometric motion blur.
Now , as you notice, it could nice to add "vdb nodes" to the nodal to reshape the fields.

prometheus
12-14-2016, 02:54 PM
I'm not entirely sure they actually can have advection in the current version of LW. At least not in TFD itself. Maybe if they create another plugin that can manipulate LW particles/objects by using the TFD cache perhaps? But even then I am guessing. LW particles just isn't designed for ad hoc interaction when containers meet (or objects entering the container) a real shame actually. Something I hope they fix.

I was one of those who frequently asked and wondered about when particle advection would come for the turbulenceFD plugin for lightwave, and it was One of some factors that ruled out my decision to buy turbulenceFD.
We did got information that it was planned for a later release, very much later Jascha told me it wasnīt possible with the current lw sdk.

So we can only guess and speculate if the lighwave next version has changed the SDK to open up for turbulence in order to make it possible to advect lw particles in the next release, unless someone from the Lw team or jascha himself says anything that is.

Michael

m.d.
12-14-2016, 04:13 PM
I was thinking the way Maya works. Last time I use Maya you had to declare the velocity field in a similar panel to allows you to use geometric motion blur.
Now , as you notice, it could nice to add "vdb nodes" to the nodal to reshape the fields.

Interesting....hadn't used VDB in Maya
I suppose an easy way to do MB in a volumetric render would be to advectively vector blur the density of the volume based on velocity of the camera and individual voxels....

That would be an interesting hack to achieve motion blur in a volume without having to sub sample time

You could do this right now in Houdini, with some clever nodes and LW camera data.... you could import 'pre motion blurred' volume fields, course that loses the flexibility of a camera change

vbk!!!
12-14-2016, 04:42 PM
Interesting....hadn't used VDB in Maya
I suppose an easy way to do MB in a volumetric render would be to advectively vector blur the density of the volume based on velocity of the camera and individual voxels....


At the time I used vdb from Houdini in Maya, it was necessary to remap density and velocity values to get something correct.
Declaring and modifying a vel field could be good to tweak motion blur value per primitive.

I don't know about to compute the vector blur from density advection since you can't get the voxel neighbouring relationship only with density.

By tha way,I never wondered how Houdini can mix geometric blur and camera motion blur. I always had the feeling that, when you check on the geometric blur option of the object, it overides the camera motion blur. But I never properly tested it ...

m.d.
12-14-2016, 05:05 PM
At the time I used vdb from Houdini in Maya, it was necessary to remap density and velocity values to get something correct.
Declaring and modifying a vel field could be good to tweak motion blur value per primitive.

I don't know about to compute the vector blur from density advection since you can't get the voxel neighbouring relationship only with density.

By tha way,I never wondered how Houdini can mix geometric blur and camera motion blur. I always had the feeling that, when you check on the geometric blur option of the object, it overides the camera motion blur. But I never properly tested it ...


I would compute vector blur with velocity fields inside the VDB (from a simulation)....but apply it to advect the density.... this would only hold true for non moving camera....at that point you would have to add the cameras velocity to the velocity field

If this is similar then geo blur would have to exactly match camera blur, as a bright spot shining through a volume would have to be motion blurred to the same degree as the volume it is going through....so an override would make sense

darkChief
12-14-2016, 05:26 PM
Excellent news :)

As promised...


https://www.youtube.com/watch?v=v29o9XDt9YM&t=2s

m.d.
12-14-2016, 06:32 PM
As promised...


https://www.youtube.com/watch?v=v29o9XDt9YM&t=2s

Awesome...awesome....

You have some really good features there :)

MichaelT
12-14-2016, 07:26 PM
I was one of those who frequently asked and wondered about when particle advection would come for the turbulenceFD plugin for lightwave, and it was One of some factors that ruled out my decision to buy turbulenceFD.
We did got information that it was planned for a later release, very much later Jascha told me it wasnīt possible with the current lw sdk.

So we can only guess and speculate if the lighwave next version has changed the SDK to open up for turbulence in order to make it possible to advect lw particles in the next release, unless someone from the Lw team or jascha himself says anything that is.

Michael

Yeah,, suspected as much. And I too have decided against TFD for this very reason. Not that it matters for me now anyway, when I have Houdini. But it would be much nicer to be able to blend containers. Life would be so much easier :)
I still haven't decided on C4D though, but it is kind of a low priority at the moment. I have been too busy with private matters this fall. Which have hurt many things I would like to have been doing instead.

jwiede
12-14-2016, 07:59 PM
As promised...

I realize that's a prelim. UI/UX, but I noticed absence of progress indicators and processing abort capabilities in a couple spots where needed. Presumably you'll add both progress indicators and processing abort capabilities before release?

DrStrik9
12-14-2016, 08:17 PM
Now if only the new hypervoxels can make an effect like this. :)

Indeed. If so, awesome, If not ...............

Ernest
12-15-2016, 12:58 AM
I realize that's a prelim. UI/UX, but I noticed absence of progress indicators and processing abort capabilities in a couple spots where needed. Presumably you'll add both progress indicators and processing abort capabilities before release?

I hope he'll add them after the initial release. Way too many fluid solutions have never gotten released because because the developer wanted to make it just a little bit better, first.

darkChief
12-15-2016, 03:16 AM
I realize that's a prelim. UI/UX, but I noticed absence of progress indicators and processing abort capabilities in a couple spots where needed. Presumably you'll add both progress indicators and processing abort capabilities before release?

Will see what happens. Multi threading code (mainly parallel code) kills the Lw monitor progress bar. So depends on finding a work around before the first version is released.

vbk!!!
12-15-2016, 03:45 AM
I would compute vector blur with velocity fields inside the VDB (from a simulation)....but apply it to advect the density.... this would only hold true for non moving camera....

ok ok ! That's the gridless advection trick . Silly me : I totally forgot that. In Houdini , it can be really slow ... but nice ... but slow.

@darkchief :
Really nice test, thanks to share. Do you think it could be possible to mesh the fluid at render time ? Or generate a cache on disk ?
And my last question : Will be a non-gui render option ?

darkChief
12-15-2016, 04:52 AM
Particle caching is already supported. The preview is just for look development. A mesh sequence for relevant frames will be generated. And Deep Rising will have its own mesh sequence plugin. Working on this at the moment. Non gui as in screamer net? If so not sure yet.

vbk!!!
12-15-2016, 06:29 AM
I was thinking to a nongui simulation ala realflow or ala houdini via command line that allows you to make your cache without the need to open your soft and wait for OGL refresh etc ...
Reaflow simulations are faster with non gui
Another way could be to generate the cache without to have to watch it frame by frame in the viewport of the software. Houdini has cache nodes for that.
Do you think theses things could improve the performance of your solver ?

darkChief
12-15-2016, 06:37 AM
Yes, of course they would improve the simulation. Not by much though. A commandline version could be available in the future.

Opengl doesn’t update while the simulation is running.

prometheus
12-15-2016, 01:21 PM
Yeah,, suspected as much. And I too have decided against TFD for this very reason. Not that it matters for me now anyway, when I have Houdini. But it would be much nicer to be able to blend containers. Life would be so much easier :)
I still haven't decided on C4D though, but it is kind of a low priority at the moment. I have been too busy with private matters this fall. Which have hurt many things I would like to have been doing instead.

I hope You get up to your game plan soon.
Personally I have a new temporary job that just got extended to the end of may 2017, then I do not know if they gonna keep me, nor if I would enjoy working with it for a long time etc, itīs just administration and some minor marketing stuff, but it only takes 12 minutes by bus..and 5 minutes of walking.

So that is good that the money flows in, which I may later invest in lightwave or something else...meanwhile I am awaiting and watching how the next lightwave develops, and lightwave 11.6.3 and 2015 will be my tool for hobby personal projects only I suspect, then itīs just a matter of time time and time, and also having time to learn more blender....investing in cinema4d with itīs pricetag and learning that(though I think it may be easy to learn) would just complicate it with more time to learn, and extremly expensive for someone not working in the 3d industri for the moment.

The last time I used Lightwave in work was for doing gym machine renders at the end of 2012.

I would also need to really focus and simply cut out a lot of what I would want to do, I donīt think itīs possible to do all stuff of projects I have stored in my mind, reckon that is the case for the most uf us, the problem is to sort out what should be focused on and not, or invested in or not.
So hang in there in this darkness and slippery roads we got right now up here in the north.

CaptainMarlowe
12-15-2016, 01:21 PM
Well, it's really interesting. I'm most impatient to see where you will go with your plug-in.