[flow_default] Transcription: 02-Secondary Cloth.json

transcriptions/02-Secondary Cloth.json

@@ -0,0 +1,7 @@
+{
+  "audio_file": "02-Secondary Cloth.wav",
+  "text": "Alright, welcome to listen sex in this lesson. We're going to add an additional layer of cloth lower back So it catches the falling sand and will make it breakable as well. So we see more cloth tearing apart So first let's create our grid Make it 1.5 by 1.5 and let's make it point six in the y-axis. And let's remesh our grid like we did with the bag, but this time I want our wireframe to be denser at the middle of the grid. So we have more points and as we go out to the edges, I want fewer and fewer points. So that when the centers are grid apart, we'll have more cloth fragments in the center than out in the edges. So to do this, let's first drop down a remesh node and let's change the way the node computes the edge length from uniform to adaptive. So we get these control attributes and what we want to use is this mesh size attribute called target mesh size. And as always when we have an attribute that we need to manipulate, we always go to, yup, an attribute drop. So let's drop down an attribute drop. And this time instead of using noise we want a controlled gradient from the center going to the edges and there's a lot of different ways to do this. The easiest thing for me in this case is to isolate a point from the center of the grid and compute the distance from that center point to the rest of the points in the grid and that will result in a circular gradient automatically. So let's drop down a last node and let's select a point in the center near the center of our grid. Let's say this one. Delete not selected. So we isolate our points and let's connect our point to the second input for VARP and let's dive inside. So here we have the position attribute of the whole grid and to get the position of the second point we need an input point attribute node and let's connect our file with the op input 2 since our point data is coming from the second input and to compute the distance we'll drop down a distance node and let's connect our positions and to have more control over our gradient we'll use a fit range node and bind export this whole thing as our target mesh size. Let's promote our fit range parameters to the outside and let's create a visualizer for target mesh size and voila. Now we have a gradient that we can use to set our minimum and maximum mesh sizes. So let's say our smallest mesh size is going to be 0.015 and our largest will be 2. And let's widen the area where our small mesh is. So we need to see if this blue area getting larger. So let's say 0.75. And to make the transition between the small and large meshes more smooth than this. We need to take our maximum value up. So let's say 4. So we have a smooth transition between our mesh sizes. So let's see if this worked. Let's get rid of our visual eyes and see our rare frames and let's see our remesh node. And we now have different amounts of points going from our center to our edges. And let's take our gradation instead of 0.25 and let's make a relative density 1 instead of 2 and maybe take iterations down to 1 as well. So I think this is fine for now. So we're ready to convert this into a cloth object now. So first let's drop down an edge fracture node to make it breakable. We can actually copy the one that we already have and just mess around with its parameters. Let's visualize our pieces and I want more pieces than this for this sheet. So let's say our initial pieces will be 150 and let's lie down and mess with our mountain options as well. And we need something to hold our cloth object in place. So let's create four wires that hold it from the angles. So what we're going to do is isolate these four points and we'll scale our grid up, isolate the same points from the bigger grid and connect them together to create a wire. And to do this, we first need to assign an ID attribute to our points so Houdini knows which points to actually connect. So let's drop down an attribute triangle and let's type I at ID so it's an integer attribute equals PT num so now each point will have its own ID attribute and let's isolate our four points so drop down a blast node and select the outer four points. Delete non-selected so we isolate them and let's copy this thing over here. Let's actually copy our group and paste relative references in the x and 1.5 in the z. Let's merge these two nodes together. And to connect them, we will need an add node. Let's go to our polygons tab, connect them by group and add points by attribute. And attribute in this case is the ID attribute. All right, now we have our four wires. And since we don't need our ID anymore, it can actually have right now and we will need more points than this. So let's drop down a re sample node. Let's take down our length to 0.05. So we have more points than this. And I think that's enough this way. A wire will now bend properly. And you can think of this as a way to increase your wires resolution and to have a completely clean setup. Let's fuse our points together. So in case we have points that are on top of each other, they get deleted and don't cause any problems in the solver. So let's drop down a fuse node. And to differentiate our wires from our cloth sheet, let's make a group for them. Let's call it wires and change our group type from preventives to points since we don't have any primitives yet. And now we can actually merge our wires together with a cloth. So let's drop down a villain. Configure cloth. First thing, let's change our group type to points. And in this group right here, let's type sdisk circumflex wires. So now the node understands that we want everything to be treated as a cloth object except for a virus group. And let's set our mass to 0.2 like our bag and get rid of our compression stiffness. Do the same thing with our Vellum weld so we can glue together our fractured edges. Also in the group, let's say point group as best circumflex wires. And let's turn on threshold and let's leave it like this right now and we will actually come back to this again in a second. Let's first drop down another volume constraints node. Set the type to string and we will only need a wires group this time. Let's actually set the mass really high so it doesn't get affected by the cloth getting hit by sand. So let's take this up to 5 and let's actually set our thickness manually. So later when we change this into polywire it doesn't look weird. So let's set uniform our thickness, visualize it and maybe increase this to 0.02. I think We can also use the pin points to make the outer most point of every wire. this case is lower threshold in the center and a high threshold at the edges. So the center will be more easy to tear apart. And since we've already done the same thing with a remesh, I'm going to duplicate our VOP, put it before edge fracture, and let's change our bind export to break threshold scale. So we'll use this attribute as a multiplier for a big threshold. So let's set our minimum value to one. So it multiplies by whatever value we specify in the big threshold scale in the vellum world constraints. And let's take the maximum value all the way up to five. So it really isn't easy to break as the middle parts. Let's visualize this real quick. And let's take a minimum value to 0.6 and the maximum value to let's say 1.5. Okay, let's now go to our vellum weld constraints, scale this by attribute and it will use our break threshold scale that we made and change the type to stretch ratios just as we did with the bag and maybe take this up to 0.2. So it's more easily breakable than a bag, which is now sitting at 0.25. And let's create nulls now. So one for the geometry and one for the constraints. And since we've named this cloth geo, I'm going to be very creative with the naming in this one. I'm going to name this cloth2.geo and cloth2.con. All right. So let's actually input this and put it up now. I'm going to bypass our cloth object, duplicate this, name it cloth2 and change our subpath to cloth2g and cloth2con. And we will need to move our ground plane down as well because if the grains come down and hit our cloth right now, it will hit the ground since they are so close together. So let's take this down to negative one and y axis so it has more room to move freely and let's actually make a flipbook and see what this will get us right now. Alright this is looking good right now I like the way the cloth fragments look and I like this violent rebound that the wires make the bag give us hand. So let's get our bag back in play now so on bypass this and let's actually increase our sub steps to 15 Since the simulation is really starting to get heavy now and I'm going to go into the advanced tab and here under fluids We have a parameter newly introduced in Houdini 19 called special sort integral. I'm going to turn this on so special sorting is a new way of calculating the motion of nibbing points by making sure that the points that are near each other in 3D space are also near each other in memory. And it does this every 20 frames by default. So turning this on means more consistent and accurate results, but it also means faster same time. I actually go over this more in depth in my Houdini Hive video so check this out if you're interested but basically turning this on makes your sims faster and your loose points fewer and actually before I cash this let's go to our grains real quick and I'm going to increase our resolution even more so let's say 0.005 and I guess this is more than enough and we'll call this our final resolution for the grain. So I'm going to cache this out now and maybe we'll do one more adjustment before wrapping up this lesson. Alright, so before we actually can take a look at what we have now, let's focus on our split node for a second. We still have our bag coming from the left output, but now from the right output, we don't only have the grains, but the cloth too and our wires as well. So let's drop down another split node here, choose the stream grains so it comes down from the lift output and let's copy this null over, name it cloth2 and let's actually do another split here and split our group, our wires and invert selection. So we have a cloth coming out of the left output. And from the right output, we have our wires, maybe drop down a polywire node. And if we type here at T scale, this is the attribute controlled by the thickness. So if you remember, we made our thickness 0.02 for the string. So now we have this at 0.02. Let's merge everything together and take a look. Okay, and maybe let's divide this by four. Alright, this makes more sense now. Alright, let's have a different color for our wires, like yellow color or something. And let's actually make a flipbook of this now so we can see what we have. Alright, we're very close now to what I wanted when I first started this project. One more thing I want to adjust is the brick threshold of our bag. Let's see if we can get it to hold on for a bit longer before it breaks apart. So let's go to our weld constraints and let's take this up from 0.25 to 0.4 and before caching this out, I want to talk about something called integration. So let's jump into our dotnet real quick. So right here in the advanced tab, the first option that we get is integration and we have an option called second order and an option called first order. And second order integration is the default. And it's more physically accurate than first order integration when it comes to retaining the object's energy, especially in non-linear motion. And it does a great job of predicting where the object will be in next frame, but it kind of fails when it comes to calculating how much should an object bounce and it could produce artifacts. So technically the first order integration is quote unquote the cheaper option, but practically I often find first order result to be more realistic looking in most cases, especially when we have high constraint iterations like we do now, and it rarely causes any artifacts as well. So Vellum compensates for this integration differences by defaulting to the second order since it's the more mathematically accurate but on detection collision, so if we go to the motion tab, we will find that on collision detection it actually falls back to first order integration. So if Vellum detects any collision that is going to produce an acceleration that equals to or higher than 30, it defaults to the first order integration. And there's actually a great presentation by Jean Lynch about this that I highly recommend you to watch. But giving this information, we have two ways of experimenting with the integration orders here. Number one, we can set this threshold to a lower value. So Vellum doesn't have to wait for a collision that produces a max acceleration of 30 to happen to fall back to first order integration. So let's say we take this down to five, or we can actually send the whole thing entirely in first order integration. So I'm going to try both. But first let's try taking down our max acceleration. Let's catch this out now. And let's see what kind of change does this give us in terms of behavior. Alright, so this is the result of the second order integration with the max acceleration turned down to 5. And while it's not bad by any means, it might be a little bit stable for my taste as in the motion is more stable than I would have liked. I would have liked to see more bounce in the grains more than this anyway. So let's try to cache with the first order integration now and see what we get. But before doing this, there is actually one last parameter that I need to talk about. So let's jump in or dopnet. So in the command tab of a solver in the friction options, we have this parameter that's called static threshold. So if our tangential velocity is less than this threshold right here, it will be completely eliminated through fiction. In other words, raising this up might give us less sliding, and it's actually worth your time to play with this since fiction can be one of those forces that you don't notice too much, but can increase the realism of your work if it's calculated properly. So I'll try to raise this to 2, and let's go to the advanced tab, turn our integration to first order, and I will revert this max acceleration parameter to its default. And let's cache this out now and see if our results are going to be better than the last time. Okay, I'm very happy with this now. The amount of time spent in anticipation before the bag tears apart makes the eventual tearing much more impactful and we now have much more dynamic sand movement as well and having this second layer of cloth tear apart as well makes this whole thing much more stimulating to watch. So this is the end of our time together. Thank you so much for watching and listening to me ramble about Vellum. I hope you've learned at least one new thing from this and please let me know if you have any questions and I'll see you guys in the next tutorial.",
+  "language": "en",
+  "confidence": null,
+  "duration": 1210.14
+}