Hi Gérald, I would like to request a feature for VWD that would make it possible to use it for tube, or rope-shaped meshes. After watching your chain tutorial I tried to do something similar to a long tube, so it would look like a suspended cable or something. But of course that cannot work now because rigidity for the cross-section of the object cannot be set to be near infinite, and at the same time having length-wise rigidity low enough so the object can bend (without collapsing). For soft clothing the possibility for splitting those parameters might not be very useful, but I think for things like trees, jewelry or other bending objects/accessories it would be. Thanks in advance for considering this option. Erik

Hello Erik,

In fact, I think you can do it with VWD in its current version. Could you show me some examples you would like to simulate.

These examples could be a very good test for the tutorial showing the different rigidification methods.

I don't tell you that you will be able to simulate everything in VWD, but I am sure all its simulation possibilities have not be shown.

I am very interested by extending the capabilities of the program. More the users will have ideas, more the program will improve.

It is always very interesting to discover ( and then repel :-) ) the limits.

Have a great day.

Gérald

Hi Gérald! Here's a screenshot of the type of objects I mean.

I first tried to bend a 'Climbing Rope' by 3-D-C (for sale here on Renderosity), but no matter what VWD settings I tried it kept collapsing into a floppy mess. What I would like to see in VWD is a material setting that allows cross-section rigidity of an object to be (very) high, and high flexibility along the length of the same object. Just like a rope or cable does not substantially change its circular cross-section while bending and coiling. This would for instance allow animating flexing trees and grass with the advanced wind emulation you are developing for VWD. And so many other 'bendy' things : ) I hope I made myself more clear to you, and maybe there already is a way to do this in VWD, I'd sure like to learn how to do it then. Thanks! Erik

Hi Gérald,

As a variation on the question above, I have also this one: Could you create the possibility to make a dynamic object flexible, but not compressible? Like a water-filled balloon can be deformed without changing its volume? Here's an example:

A person's bottom sitting on a hard surface will flatten, but the muscle and fat tissue cannot be compressed, so the 'seating area' of that person will get wider. Likewise a tight garment will 'squeeze' a human body, but that body will then bulge around the edges of that piece of clothing. Would such a deformation be possible to do in VWD? Erik

Hello Erik

as I made a scene with some pipes simulated in VWD a long time ago I tried again. Here is a quick result:

An Idea for VWD: I fixed both ends of the pipe it could be interesting if we could move one end (select the fixed vertices of one end and move them were we want in the scene) or if the "move vertices on fly" give the habitlity to select a group vertices and move them as the rest of vertices are simulated.

Erik, Have you tried Vertices Neighborhood Rigidify for the part of the rope you want to hang, and Fixed Vertices for the ends you want to remain in place? I just tried it and with a Vertices Neighborhood Distance greater than 1.0 (I used 1.4), where my average Spring Length (from the Utilities tab) was about 1.4, and the rope itself draped fine and didn't collapse on itself, and the ends stayed fixed.

Hi ebergerly, I tried quite a few combinations, a Neighbourhood Distance of 2 (I think) gave results that came closer to what I'd like to see. But still the tube folds or kinks in places, and resistance to self-collision is more or less 'symbolical'. I see the same when draping clothes when cloth turns inside-out, I've spent a LOT of time repairing stuff in ZBrush after using VWD. All down to user error I'm sure, but a bit frustrating at times. Here's an image of a knot I tried to make (just for the fun of it)

@erik-nl : Hello Erik. About the rope, do you use the 'Climbing Rope' by 3-D-C in the image you sent us? I can buy it or make a mesh close to this one. I will make some tests but the rigidity by neighborhood is perhaps not large enough to rigidify correctly the rope. This depends on the mesh structure.

For the deformation of the character, Yes, it is possible but you have to define the character as cloth actor and the chair as collision actor. You use a rigidification by extension with a high value, 5 for example and a low softness 0.01. So the character mesh modification will be extended on a large size. This will give the feeling there is a volume deformation. Perhaps, it would be interesting to add a small inflate function. This needs some tests. The chair must have a correct mesh size (max 1.0). Otherwise, you can use the subdivision function. On the character, you can set the vertices that don't have to move as fixed vertices and then, you remove all the dynamics in the simulation (Inertia, Gravity etc...). Then you make a short static simulation and you export the character to the host. I can do a short tutorial showing this.

In the current version, it is possible to use the self-collision on a character imported as cloth actor. This allows to correct the pose for a cross-leg sitting pose or correct the armpits. This will certainly be the next tutorial I will do . The next version will allow you to correct a full character animation using the self-collision. This will be very helpful if the simulation is not doable starting on the final pose.

@Smaker1 : I am not sure to understand, Stef. Do you want to say that some vertices of a "cloth" would be animated by the host and other vertices would be animated by the dynamics?

Hi Gérald,

I used the low-rez version of the Climbing Rope by 3-D-C, but I'll try again with the hi-rez version to see if the higher density helps prevent the shape from collapsing. Of course any cilinder primitive would be suitable, but this one can be pre-bent, and for tests everyone can use identical geometry to make results comparable. If successful I'll then try if I can get a tree model to bend in the wind in VWD, would be great to see it work!

The method you describe for deforming a human figure is the one I used for the picture of the seated woman I posted. Unfortunately not much of the effect is visible, because the contact areas of the legs and buttocks flipped from convex to concave, in other words, became dents instead of flat areas, without pushing the sides of the legs outward to make them wider. Perhaps adding internal pressure can help to improve the result.

For this, and many other effects you need a very gradual increase of the displacement to prevent a sudden step in the geometry. But as it is now there is no way to check the desired distribution of the displacement before pressing the simulate button. Would it be possible to add coloured 'heatmaps' for selected vertex areas to help visualize the area of influence corresponding to chosen settings? Similar to the one you created for the resulting vertex stress levels after a simulation?

I think in general such a visualisation of parameters could help dramatically shorten the 'trial and error' process we have to go trough each time.

Erik

ps: If you'd like to see some of the results of my VWD learning-efforts you can find them here erik-nl at deviant art

Hi Erik,

I made some tests with a rope and I have the same issues as you. It is necessary to move the dynamic vertex very slowly and we are not sure to have a good simulation.

The dynamic deformation has been written for modify a cloth. For a rope the direct displacement of a vertex is very violent. It would be better to displace a part of a new spring. By making some tests, the self-collision works correctly when the simulation runs alone but it don't work well with the dynamic deformation, because this direct displacement.

I am confident for the simulation of a tree. I certainly will try it soon. For me it is necessary to use a very low gravity because its action has no utility.

What are the settings you used for the sitting pose. For me, the best choice is a rigidification by extension with a high extension (5 or more) and a softness set to 0.01 or less. But you are right. VWD is not able to do a volumical simulation. This kind of simulation is not easy to do. I tried to do so, many years ago (with another name) by using a volumical meshing. There is a test here : https://www.youtube.com/watch?v=GAd4SH_OuWM The simulation needs a volume (a mesh fully closed) and a tetrahedral meshing. All these functions are in VWD but the constraint are very high (a volume). The tetrahedral meshing is not really a problem. To make a perfect volumical deformation, it will be necessary to know the volume of the area you want to modify and, during the simulation, always try to keep this volume. I thought to a solution that I can explain you if you are interested, but this is not so easy to write. You are right, it is possible to make a first test using the inflate function.

The vertices coloration for a selection would be a good improvement. I promise to think to this.

I looked at the images you posted to Deviant Art. They are really wonderful. I look forward to see your hair results. If you are interested by hair simulations ... of course. Are you sure that these are just learning efforts, master Erik...

Have a great day.

 Gérald

Thank you for considering my suggestions Gérald, I really appreciate that you have an open mind for things like that. I've only just begun to explore the possibilities of VWD, I feel I need to get a solid grip on the basics before venturing into more hairy stuff ; ) But I will eventually, no question about it. Of course I should have made notes of my settings for the sitting figure, but I didn't, sorry about that. But I'll try yours to see if they can improve the result. Thanks also for having a look at my images at Deviant Art. If you like them enough, feel free to use them as reference for your wonderful program!

Erik

Hello

yes , for rope it's mandatory to move the vertices very slowly. As I do still render, I often export and then use my saved object for new simulation if it goes wrong. I also saw (trying the "knot exercice" !) that extreme folds made self penetration.

@Gerald: about my question, I made a new test with a strong rigidification of both end of the rope and then moving the ends. It works. I have better results than in my old projects.

@Erick: didn't though about tree simulation that an intersting idea that I must try. Wonderful gallery and I agree with you: plenty of old clothes found a new life with VWD. Do you use the DS4/Octane bridge? What's you feedback?

In combination with VWD that climbing rope by 3-D-C doesn't have to be only a rope, there's so much more you can do with it!

Here's an example, made by scaling and pre-posing it in DAZ Studio, then draping it in VWD: After draping it I used ZBrush for some corrections and 3DCoat for better UV's, but VWD was an essential step in the process of creating it.

@Smaker: what can I say about Octane for DAZ Studio, you've seen my results : )

Converting materials by hand for Octane can be very time-consuming, but is usually well worth the effort.

Erik

I'll try to not hijack a VWD topic too much ! Does the brigde manage DS4 instances? about converting Iray Materials? can we save a complete Octane scene we can render later without opening DS4? can we export abc ? Today I don't have the bridge and do it manually with obj or abc exports but I miss instances and sometimes obj export forget textures (depending of the ds4 material configuration).

@Smaker1: Octane for DAZ Studio is not really a bridge, it's more a plugged-in render application with an embedded version of Octane in it.

But you can export a complete render file of a scene to be rendered in Octane Standalone if you want to, and for certain Octane features (volumetrics, object motion blur) you'd even have to.

Grouped instances in DAZ Studio used to be problematic for Octane for DAZ Studio, but that issue has been solved (as can be seen in my DA collection). However, it still is finicky about the origin of parented stuff. So using DAZ Studio's bone editor to move those origins to 0,0,0 is a necessary step, even with scenes made by the likes of Stonemason.

By the way, all my DA images with VWD experiments that show more than one figure are done with instanced doubles and triples. Just like the seated woman and chair I posted here.

@erik-nl : Thank you Erik for your proposal but they are your renders and, for me, I cannot use them even with your agreement. Great image.

I don't know Octane render. I saw great renders that use it. Erik and Stef, could you explain me why Octane is your prefered render engine. Its speed? its quality? other advantages?

erik-nl posted at 9:00AM Wed, 14 June 2017 - #4307400

@Smaker1: Octane for DAZ Studio is not really a bridge, it's more a plugged-in render application with an embedded version of Octane in it.

But you can export a complete render file of a scene to be rendered in Octane Standalone if you want to, and for certain Octane features (volumetrics, object motion blur) you'd even have to.

Grouped instances in DAZ Studio used to be problematic for Octane for DAZ Studio, but that issue has been solved (as can be seen in my DA collection). However, it still is finicky about the origin of parented stuff. So using DAZ Studio's bone editor to move those origins to 0,0,0 is a necessary step, even with scenes made by the likes of Stonemason.

By the way, all my DA images with VWD experiments that show more than one figure are done with instanced doubles and triples. Just like the seated woman and chair I posted here.

Thanks Erik, I was thinking about a bridge because you need to have the standalone license + the bridge license. So I understand that the plug contain Octane, it's only a matter of license . I'll try the demo to see if it fits my needs

@Gerald: I discovered Octane with the beta test of Octane V0.9 (a long time ago!). My firsts interests was unbiaised render, qualtiy and interactive rendering.

Materials are more simple even if I'm far from mastering all the subtilities. My main lack: if someone know books, tutorials on this topic: I'm interested!

About rapidity, Octane changed my life !! Not only for the final render but also working renders to make the scene : lighting, framing,.... I made (at least for me) a big jump in term of quality because it's very easy to see quick results and improve it . Final render doesn't take more than one night for very very big resolutions (average of 60 Mb, far from what I post) . I remenber waiting for weeks with Bryce or Vue!

With Octane V3 I would say that I have a complete solution filling my needs : graphic card memory is less a trouble. My graphic cards are: a GTX 680 with 4gb dedicated to Octane (I made a scene with 23 Genesis 3) and a GT780-2GB for display

We have subsurface scattering, instances, open vdb ,...

I didn't try Iray !

Same here. Long before Iray became available I was using the open-source unbiased Luxrender through Paolo's Reality plugin, because what came out of biased renderers like 3Delight was definitely NOT what I wanted. Downside was render times had to be expressed in days, rather than hours (CPU only).

So the VERY reasonable introduction pricing of the v0.9 (beta) Octane Standalone + DAZ Studio plugin bundle was irresistible for me, and I haven't used any other renderer since.

Never touched Iray, but I might, when (and if) its features expand enough to come close to what I have now with Octane Standalone and Octane for DAZ Studio.

Erik