Add transcription for: frames/MDA_TheSymbioteCreature_DownloadPirate.com_SYMBIOTE_CH007_INTIAL_VELLUM_SETUP_1080p_mp4_frames.zip

transcriptions/frames/MDA_TheSymbioteCreature_DownloadPirate.com_SYMBIOTE_CH007_INTIAL_VELLUM_SETUP_1080p_mp4_frames_transcription.json

@@ -0,0 +1,8 @@
+{
+  "text": " So in this in the previous lesson we stopped at the master transform, I'm gonna call it master transform, and I want to just clarify some things. Yesterday I mean the previous lesson we I told you that we had to create these two attributes the density and stretch stiffness, but if you notice the max density mult and min density mult are inverted so instead of having the min density multiplier here and the max density multiplier right here so they should have been swapped but I did it on purpose because I want to have more density where if you visualize the ramp here I wanted to have more density where it's black and less density where it's white so in order to for these arms to pull the body they have to be strong enough to pull it so I wanted to compensate the weight, like the weight of the body itself by lowering its density. Because otherwise if we leave it as is, the weight of the body will be so high that the arms will not be able to pull it forward. So it's kind of a fake, it's a solution that kind of fakes the dynamics of this creature. Basically the white area is going to be lighter than the black area. So this is why we have have a magnitensity multiplier here, so the zero values of this ramp are going to be mapped to this magnitensity value, and the values that are higher will be mapped to this one. The same applies to the stretch stiffness, so I want the stretch stiffness to be higher when there's black areas and the opposite for the white areas. So we're going to deactivate this visualizer and set properties, creature. We have the master transform now, we have the ground set up in the previous lesson. In order to switch between the viewports as I do here, I press space and then press 1, 2, 3, 4. This switches the views. It's really a handy shortcut for navigating the scene. In this lesson, we're gonna be tackling one of the most interesting parts of this simulation, which is the vellum solver itself. I usually prefer to use the dot version of the vellum, basically reconstruct the solver from scratch with all the nodes, with all the microsolvers in it. But we're just gonna, just for the sake of convenience, we're gonna use the subversion of the vellum solver. And we're gonna end up in the same kind of result the same result. It's going to be easier to dig into the solver and add some microsolvers and additional nodes there. So let's start by adding a null which is going to be called outrest. What I call outrest is usually the static version of our creature and you see the difference in performance here because of the attribute blur. So it's fastest to drag the timeline when you have the attribute blur disabled. It adds kind of a strain to the performance of our timeline. So keep in mind that you may bake this in advance. Usually this is what I would do, but I don't know if I'm going to change the timing later on or not. So let's keep it like this. Maybe we'll cache it before simulating. We'll see. So the alt rest, no. It's pointing to the master transform the geometry with all the attributes, blinking attributes active, the live attribute transferred to the tetrahedral mesh. So yeah, well we're gonna check just in case we're gonna check that the geometry is tetrahedral. We have like a third mouse like you click with the scroll wheel on the null you'll see that it says polygons and tetrahedrons. So we have a polygons on the surface and the tetrahedrons inside of it. If we clip our geometry here, let disable the... hide the grid that we had and let's clip it. Yeah, so we have to clip it in the positive direction. So as you can see earlier I was testing this geometry. You see that the attributes are transferred properly inside the volume. The trihidrons they get their attributes also, which is how it's supposed to be working. So yeah, we have a volumetric mesh. So let's create the first valent constraints. Let's say valent cloth first. The valent cloth constraint is going to be applied to only one single group. And the group is going to be the surface triangles that we had created earlier, with the tetrahedralize that conform. As you see, we added the surface triangles. So we're applying these constraints only to the surface triangles primitive group. What I'm going to change here is the stiffness of the stretch and the bend stiffness are going to be really minimal. So yeah, the cloth constraints. I'm going to set the mass to unchanged and the same for the thickness. I'm not going to change these, but the stiffness is going to be really really low. Something like this. really low and I'd want to the skin to really stretch and bend in all its glory and the band stiffness is going to be really minimal. We're gonna also I don't like leaving these output groups by default this is why I always name my VELM constraints with a CSTR prefix and you'll see why because we'll have a bunch of constraints that we have to use inside the VELM solver and you'll be able to recognize easier the names of these constraints. So one is going to be called constraint clove stretch and constraint clove bend. We're not going to change anything else here. And we're going to duplicate this valem constraint node. And if you want to connect, you see that all valem constraint nodes, they have three inputs, three outputs. So the easiest way to connect these is press J on the keyboard and it's going to turn a cursor on in a kind of a needle and you just cross both nodes like this and it's going to connect both nodes with all the inputs into the outputs. So instead of the development cloth here we're gonna change to tetrahedral stretch and here we're gonna click preserve volume and use the non-linear A-Rap. The non-linear A-Rap is a constraint from the previous versions of Houdini, when the tetrahedral constraints were introduced in Vellum, and A-Rap means as rigid as possible. So this type of constraint will try to really preserve the volume of your tetrahedrons. And if you have enough sub steps in your solver, it's gonna try to preserve the volume of these constraints. So let's see what we do with our mass. mass. So for the mass we're gonna choose 200. This is a proven kind of value that I used before and in the thickness we're gonna use 0.32. I calculate uniform. So yeah the mass here is gonna be calculated depending on the size of the tetrahedron and the size of the edge between the points inside the tetrahedrons. So the mass is gonna be varying depending on that. And we're gonna scale this mass by an attribute that we created earlier, which is density, the one that we created right here in the set properties of the creature. So you see the F density, this is where it's going to go. So this 200 value is going to multiply by the density attribute that created earlier, and the thickness is going to be uniform for all the points. So if you zoom in and visualize the thickness, you'll see the radius, the collision radius of these points. we're gonna hide this and for as for the stiffness we're gonna just the stretch stiffness we're gonna set it to 0 0 0 1 and scale it by attribute which is going to be the stretch stiffness the default for the damping ratio we're gonna set to 0 0 15. The damping ratio is a parameter that will try to stop any unwanted energy inside the tetrahedral constraints so any jittering and any unwanted movement or like really fast movement, it's going to try to calm it down. It's like an internal friction force. And this is a value that we'll play around with a lot. I mean, I play around with it a lot because it helps with achieving really stable simulations. And we're going to rename this constraint with CSTR, that's stretch. So our CSTR, the tetrahedrons will be, tetrahedral stretch constraints will be saved in this primitive group. Also, we apply this to surface triangles. We have to delete this. And we're gonna apply it to the GRP tests because we are applying this to the internal tetrahedrons. Now as you can see, if you dig into, go into inside the geometry, you can see that like you have kind of a web of primitives that are constraining all those tetrahedrons together and making one single soft body. Yeah, we're done with this tetrahedral constraint. And we're also going to create, let's say, a Vellum constraints, press J, drag the cursor like this, and go here and create a tetrahedral fiber. And the tetrahedral fiber constraints is a special one that's really underrated, one of the most underrated constraints in Vellum and in Houdini in general, because this is what creates constructions expansions of muscles and it's used a lot in the new muscle system inside Vellum. If you click, if you just search the muscle for the muscle nodes you can see that you have muscle properties, certified, solver, Vellum. These are all related to the new muscle system for the Vellum solver and it's quite a new thing and it's not, it's still kind of in beta but it's really promising and I think in the future I'm gonna be using it really extensively. But yeah, maybe we'll explore it in another tutorial, but this is a really complex one. So let's get back to our dead fiber constraints. And that fiber constraints will be applied to the legs primitive group. And you can see that the legs are really only the legs are constrained here. And the thing is we are not going to apply this to the primitive group, but to the point group. The one where the tips are not included. So if you remember we had, if we scroll like middle mouse click onto our vellum cloth, vellum constraints, you can see that we have a GRP legs which is a point group and a GRP legs primitive group. We're gonna use the one that is in the point section, so which doesn't contain any tips. We want to compress only this part of the leg. Along the vector that we created earlier, if you remember we created the strider arm. This is the vector that is going to be used as a compression vector for the tetrafiber constraints. The tetrafiber constraints are really kind of a dumb type of constraints, it just compresses these muscles in this specific direction. It uses, by default, it uses the material w attribute, which is a vector attribute, and it's going to define the direction of compression. Also, we're going to set the stiffness to zero because we don't want to to have any initial stiffness or initial compression and we're going to set this to 1. Also for the damping ratio we don't want vibrations or jittering so I'm going to increase it to 0, 0, 3 and I'm going to create an output group which is going to be called CSTR, that's fibers. This is it, this is done and now the core of our simulation, the Veyron solder. Press J, drag to the vellum constraints, and let's rename this vellum fibers. vellum bits. vellumcloth. This is done. So now, I'm going to activate the vellum solver. It's going to think for a second. I'm going to hide the visualizer for the arm vector. I'm just cut this connection. In order to cut the connection, you press Y and just drag your cursor above this line, on top of this line, and it's gonna just cut it. I'm going to connect the ground directly here. So we have the ground visualized in our SIM. And yeah, now for the interesting part, because this is where everything is gonna get interesting. We're gonna dig into the solver. When you double click onto the development solver, you'll find yourself in a dopnet. Basically, any SOP version of any solver contains inside a top version. So it's not a surprise because you can go back, like if you want to go back to the origin where you were before, you have to go right here. But if you double click and go back to the dopnet, you'll see that it's a really simple, I mean, it's a regular dopnet with external static objects to the ground lane. Everything is linked together. You have the forces, you have everything. But yeah, we're gonna end up here. So the force output is for connecting any microsolvers and the source same for connecting microsolvers, but they're related more to the sourcing. So if you wanna emit any vellum objects, emit a bunch of spheres, a bunch of objects from the same point or any geometry, this is where you go. So we're gonna activate the force and we're gonna start creating our microsolvers that we're gonna animate the whole thing. I'm gonna deactivate my, maybe it will work like this. So, before adding any microsolvers, we're gonna go back to our subcontext where we created the ground and we'll start building our main attribute VOP. I'm gonna do it, the main for that will create lock motion is going to be inside an attribute VOP. As I told you before in the previous lesson, we're not going to spend time simulating inside Vellum and creating this force, tweaking it inside because I already know how it's supposed to be. So I'm going to create this attribute VOP. In the first input I'm going to connect the master transform and in the second the ground force. This is going to be just for testing purposes and for building purposes and you'll see why. Then we're gonna transfer it into the DOPs. So yeah, the force search algorithm is based on a simple PC Open, which is a point cloud open. So the points from our creature are going to seek points from the ground. So each point will seek the closest point or a closest cluster of points that are defined inside the PC Open, which is 10 points. So each point here in the creature is going to be searching points here on the ground, like the average of 10 points in a radius of 0,1, which is 10 centimeters. But we're going to increase this radius to 0,5 because we need some search radius. So we need a lot of radius for the points to find anything on the ground. What we're going to do with this PC Open, we're gonna find points on the ground and see what are their attributes, especially the velocity vector and the normal vector. So we're gonna find those vectors with the PC Filter, and the PC Filter goes, the handle of the PC Filter goes into the wind cloud open, and we duplicate this in the first one, we're gonna search for the velocity vector, and the second one we're gonna search for the normal vector. we're going to normalize these vectors. Normalizing vectors means that you're basically fitting the magnitude of the vector into a range of 0 and 1 between 0 and 1. And this is for precision purposes. We need to have normalized vectors whenever we're using these kind of forces, because we don't know what's the magnitude. I don't want to have any additional magnitude and have chaotic control over those forces, so I want to have vectors that have components fitted between 0 and 1. As you remember, we had a CD attribute and we had a CDR attribute. So we either import them with bind. The CD attribute is already here, so we're going to do a vector to float. So we're going to just take one component of the CD attribute and we're going to import the cdr attribute which is also a vector. I could have done a float attribute but it's not a lot of overhead because we don't have a lot of attributes there. cdr attribute is a vector attribute. We're gonna, or maybe not, I don't really remember. We'll have to go back see what we've done with the cdr attribute. Oh, so yeah, I created a vector of cdr attribute. So let's make it a float. fcdr is great. Here we're going to initialize it to the fcdr and instead of setting a vector here we're going to just set it to 1 to 0 and the same for this one. So we're gonna have floats. cdr is a float, cdr is a float, cdr is a float, everything is good. We're gonna come back here, we have cdr as a float here and the color is a vector attribute. So this is a local test attribute WAP. So the CDR is going to be a float. Let's save on those useless components of the vector. It's going to be a float which is lighter to read and right and everything else. So yeah. So how does it work? This is going to be our creature. This is just one of the tentacles. This is going to be the tip of the tentacle. the bulk of the body. So we have a frontal force that is going to be applied to the front, which is going to be inherited from the ground. So the ground is right here and it has a horizontal z-axis force which pushes the tips of the legs forward in this direction. But we also have a normal force which is going to go up or down. So whether it's up or down, it's going to be defined by the CDR attribute. So the CDR attribute is a color that turns from 0 to 1, but we're going to map it to something like minus 1, let's say n1, so it's going to be remapped from 0 to 1, between 0, like it was 0 to 1, and it's going to be remapped from minus 1 to 1. We're going to have a normal attribute that is going to be multiplied by minus 1 or by 1. So in the beginning there's going to be a normal force applied up and then a normal force that will be applied down. So the in-between is going to be looking like this. So the vector, the resulting vector is going to be forward plus up so it's a diagonal line and then the CDR is going to be multiplied by minus one and it's going to go down. So the tip will be here and then we'll end up here. So the The final movement of this tip is going to be a curved motion like this. So the curved motion will be defined like this because the CDR attribute is going to be turning on and off and its values will be mapped between minus one and one. So either the normal will be added as a positive value or a negative value and then a negative vector. So it's going to result in a curved motion over time. you'll see how it works in our system. So in order to implement this, we're gonna just take the CDR and fit it, fit range, between the zero and one, and then the final will be minus one and one. So now we're gonna multiply, because this is, we're gonna multiply the normal vector by this, by the CDR. So the up and down is controlled by the CDR, and we're gonna add this to our velocity vector that we inherit from the ground force that we created earlier. So we add this to the normalized vector here. And we're going to multiply the overall force by this color because we want to turn on, oh yeah, I forgot to connect the normalized version of the normal vector. So we're going to multiply the result of this addition of two vectors by the color. Because we want to buy one of the components of the color, so the red channel, for example. So we're gonna have a resulting vector which turns on only when the tips are white because once it's white, the overall vector is multiplied by one. Once it's black, it doesn't multiply by anything, it's zero, so the vector will be zero, so nothing will happen with these tips. And the product, like the safest way to add forces in Vellum, in my mind and how I've done it before, is adding any force that you create, you add it to the existing force. So if there's any residual forces or previous attribute VOPs that you've done before, this is how you add your additional forces on top of it. You take the existing force, which is cumulative. If you only have any forces existing in the system already, you take the previous force and you add your additional forces to it and you add it to the force. So you don't connect it to the velocity because you'll have unexpected results there. And I'm gonna press Shift S just to change the type of lines that are connecting our nodes. So here we go. This is the force. And we're gonna also create a bind export. I create temporary force attribute, which is gonna be called FRC. I do this attribute I create, this attribute always. Every time I create this attribute is for debugging and seeing what the force does. Also, I've never understood the difference between tree floats and the vector, because they're both vectors, they end up being like three float, vectors created by three floats, three components. I use either of them, it's the same for me. So yeah, we're gonna output FRC, and I'm going to debug this FRC force to see what it does. So the FRC is there. Hide the deer arm and just add a marker. It's gonna be a vector trail. Yeah, and make it kind of blue, let's say. So yeah, we have the CD, we have all the good stuff here. And oh yeah, I forgot something, I forgot to make a overall multiplier. So we're going to multiply, like you first most click on the input tree, make a promote parameter, double click on it. We're going to rename this. MoldOverall. This is going to be the the overall multiplier that adds forces to this creature. Let's put it to one, let's say. Let's check that we have normals here. Oh, okay, yeah. You see the normal is a vertex attribute and we have to go back and make it a point attribute. So, as you can see, once it turns white, because we're going to apply this only to the tips, we have to apply this only to the tips. Let's see the tips. As you can see, what happens is the forces applied up or down. So every time there's a color turning on, there's a vector that's applied on it. So let's see what happens when we go into the solver. Let's do our first simulation test. The multiplier is going to be really, really low. Let's start with the value of 0, 0, 0, 5. And it's a little bit barely visible, but I'm going to debug this unmultiplied force, because I want to really see what happens with the vectors. We're going to save this. And we're going to dig into the valence solver and do preparations for the attribute transfer. One thing to mention, which is really important, is that any attributes that come here into the valence solver are not updated automatically. So anything that goes into the solver, you have to update it. If you have any live attributes that you want to feed into the geometry during simulation, you have to do it inside the valence solver. You usually do it in the sub solver. So we connect the subsolver to the output, to the force output, and we're gonna dig into the subsolver and you'll see what we do here. You have already an output here. We're gonna object merge our rest attribute, or a rest null that we had before, out rest into the specific object, and this is this object. So we have our rest static geometry with live attributes, with all the attributes that we need. We're gonna attribute copy. Because the monster is the only thing in the simulation, There's no other geometry objects that are simulated. Where we can do the attribute copy because the object merge will only, like, dop geometry matches the object merge. These are, this is the only object that exists. In the simulation, we have one object and the object that we merge is one object and it's the same object. Sometimes you have emitting geometry and you have several objects in the geometry. You have to do groups to attribute copies, specific groups to specific groups. Otherwise, it's just gonna say that the point number is different on the Dov geometry and on the rest. So we're going to copy CD and CDR because those are live attributes and let's also import the ground because we have the ground outside the Dobs, we just import the ground force and we're gonna copy and paste our test lock emotion inside this sub-solver. This is where the forces will be applied, we'll do it the same way in the first input you have the geometry that comes in and in the second you have the ground and shift s just for the lines to be look neat and yeah for now it says that we don't have any tips group but you'll see that usually this is what it does in the first frame of the simulation don't be alarmed by this or villain does this often it just doesn't evaluate the first frame properly sometimes or some Some groups appear dynamically during the simulation. It depends. There's many scenarios where it's going to say invalid group, but once the first frame is simulated you'll see that everything is fine. So let's see what it does. So let's press... I didn't change any settings in Valem first, like before. But you'll see that something happens. You see that one arm is extending, but it extends too much. So these guys were... The creature falls down. We're gonna disable the vectors here. You'll see that once one of these tips goes white, it's going to just be pushed up forward and down. So yeah, we see that. Okay, but the thing is as you can see that it's really a really soft body and the arms they extend too much. This is because We don't have enough sub steps here to satisfy our constraints, so in order to satisfy those constraints we'll add more sub steps. In my first simulation I used the five sub steps for tetrahedral constraints you'll have to aim for five tetrahedral constraints, five sub steps I mean for the kind of sub steps. I usually change only the sub steps, I don't touch the constraint iterations, because these are more like it's more of a secondary thing. The the sub steps are the main parameter that defines the quality of your simulation and also the satisfaction of your constraints. Because sometimes you'll find yourself setting some really high values here, the higher you go, the more sub steps you need to satisfy those constraints. If they're unsatisfied, they will behave like constraints that have less stiffness than what you've set. So you have to keep in mind for this. In the forces tab of our Valum solver, I'm going to call it Valum main, I'm going to set the static threshold to 0.01 and dynamic scale to 0.01. I'm trying to set the static friction and dynamic friction as low as possible. This is over or multiplier over the dynamic and static friction. The gravity is alright, everything is fine. And that's kind of it. And you'll see that once we change the sub-step to 5, the arms are not going to extend as much. The shape, the stiffness of the creature shape is going to be higher. So the arms are not going to extend as much. They'll try to restore their initial volume, even if the force is kind of strong. Yeah, so you see it simulates, it's much slower than before, but because of the sub steps. But believe me, it's much, much faster. It's like 10 times faster than the FEMS solver. We will use the FEMS solver later on for wrinkles, and I'm going to build a system that creates wrinkles on top of the existing valum simulation. And you'll see the differences between the solvers. We have here like 250,000 tetrahedral primitives, which is not a small number for a creature like this. But so this is it will take some time to simulate. So but it's not a big deal because we can always launch wedges which is a new system in Houdini a relatively new system the top system where you can launch a lot of like different variations of your simulation. And yeah it's a we can do it overnight and launch a bunch of simulation overnight. So as you can see it's more rigid than before, the legs don't extend as much. Now we're just working on the tips side of the dynamics, you see the tips are turning white and they're kind of pushed, pushed up and forward. So for me it's, it's alright. The only thing I will change here is go into the sub-solver and we're gonna change a little bit the multiplier, 0, 0, 0, 0.06. We increase the force a little and we're going to go into our locomotion test here and we're gonna fit the the CDR attribute to something like or maybe not we're gonna leave it like this. We may probably We'll leave it like this, I'm not going to change the fitting here. The fit in the CDR section defines how much of a multiplier you apply to the normal component of the final force when you want it to go up or down. So basically if you want the arms to attach faster and fall down onto the ground faster, the tips will fall down faster, you want to increase this value. minus 2 is going to make the tips really go down and attach to the ground faster than before. Also if you want to have more of a vertical jump for those tips, you can increase this value to 2 or higher and you'll see what it does. You can experiment with these values, but I found that these are kind of balanced, so this works fine for this scenario. Now we'll have this simulation here. So the base locomotion system is done. It's just the tips, but we have much more, many more Microsofters to add here. The subsoil is the first one. And we'll see about that in the next lesson, because I think we've already passed the length threshold for this tutorial. I try to make it faster, but there's too much information to deliver. So I think I'll split it more for the next lessons. See you in the next one.",
+  "segments": [
+    {
+      "text": " So in this in the previous lesson we stopped at the master transform, I'm gonna call it master transform, and I want to just clarify some things. Yesterday I mean the previous lesson we I told you that we had to create these two attributes the density and stretch stiffness, but if you notice the max density mult and min density mult are inverted so instead of having the min density multiplier here and the max density multiplier right here so they should have been swapped but I did it on purpose because I want to have more density where if you visualize the ramp here I wanted to have more density where it's black and less density where it's white so in order to for these arms to pull the body they have to be strong enough to pull it so I wanted to compensate the weight, like the weight of the body itself by lowering its density. Because otherwise if we leave it as is, the weight of the body will be so high that the arms will not be able to pull it forward. So it's kind of a fake, it's a solution that kind of fakes the dynamics of this creature. Basically the white area is going to be lighter than the black area. So this is why we have have a magnitensity multiplier here, so the zero values of this ramp are going to be mapped to this magnitensity value, and the values that are higher will be mapped to this one. The same applies to the stretch stiffness, so I want the stretch stiffness to be higher when there's black areas and the opposite for the white areas. So we're going to deactivate this visualizer and set properties, creature. We have the master transform now, we have the ground set up in the previous lesson. In order to switch between the viewports as I do here, I press space and then press 1, 2, 3, 4. This switches the views. It's really a handy shortcut for navigating the scene. In this lesson, we're gonna be tackling one of the most interesting parts of this simulation, which is the vellum solver itself. I usually prefer to use the dot version of the vellum, basically reconstruct the solver from scratch with all the nodes, with all the microsolvers in it. But we're just gonna, just for the sake of convenience, we're gonna use the subversion of the vellum solver. And we're gonna end up in the same kind of result the same result. It's going to be easier to dig into the solver and add some microsolvers and additional nodes there. So let's start by adding a null which is going to be called outrest. What I call outrest is usually the static version of our creature and you see the difference in performance here because of the attribute blur. So it's fastest to drag the timeline when you have the attribute blur disabled. It adds kind of a strain to the performance of our timeline. So keep in mind that you may bake this in advance. Usually this is what I would do, but I don't know if I'm going to change the timing later on or not. So let's keep it like this. Maybe we'll cache it before simulating. We'll see. So the alt rest, no. It's pointing to the master transform the geometry with all the attributes, blinking attributes active, the live attribute transferred to the tetrahedral mesh. So yeah, well we're gonna check just in case we're gonna check that the geometry is tetrahedral. We have like a third mouse like you click with the scroll wheel on the null you'll see that it says polygons and tetrahedrons. So we have a polygons on the surface and the tetrahedrons inside of it. If we clip our geometry here, let disable the... hide the grid that we had and let's clip it. Yeah, so we have to clip it in the positive direction. So as you can see earlier I was testing this geometry. You see that the attributes are transferred properly inside the volume. The trihidrons they get their attributes also, which is how it's supposed to be working. So yeah, we have a volumetric mesh. So let's create the first valent constraints. Let's say valent cloth first. The valent cloth constraint is going to be applied to only one single group. And the group is going to be the surface triangles that we had created earlier, with the tetrahedralize that conform. As you see, we added the surface triangles. So we're applying these constraints only to the surface triangles primitive group. What I'm going to change here is the stiffness of the stretch and the bend stiffness are going to be really minimal. So yeah, the cloth constraints. I'm going to set the mass to unchanged and the same for the thickness. I'm not going to change these, but the stiffness is going to be really really low. Something like this. really low and I'd want to the skin to really stretch and bend in all its glory and the band stiffness is going to be really minimal. We're gonna also I don't like leaving these output groups by default this is why I always name my VELM constraints with a CSTR prefix and you'll see why because we'll have a bunch of constraints that we have to use inside the VELM solver and you'll be able to recognize easier the names of these constraints. So one is going to be called constraint clove stretch and constraint clove bend. We're not going to change anything else here. And we're going to duplicate this valem constraint node. And if you want to connect, you see that all valem constraint nodes, they have three inputs, three outputs. So the easiest way to connect these is press J on the keyboard and it's going to turn a cursor on in a kind of a needle and you just cross both nodes like this and it's going to connect both nodes with all the inputs into the outputs. So instead of the development cloth here we're gonna change to tetrahedral stretch and here we're gonna click preserve volume and use the non-linear A-Rap. The non-linear A-Rap is a constraint from the previous versions of Houdini, when the tetrahedral constraints were introduced in Vellum, and A-Rap means as rigid as possible. So this type of constraint will try to really preserve the volume of your tetrahedrons. And if you have enough sub steps in your solver, it's gonna try to preserve the volume of these constraints. So let's see what we do with our mass. mass. So for the mass we're gonna choose 200. This is a proven kind of value that I used before and in the thickness we're gonna use 0.32. I calculate uniform. So yeah the mass here is gonna be calculated depending on the size of the tetrahedron and the size of the edge between the points inside the tetrahedrons. So the mass is gonna be varying depending on that. And we're gonna scale this mass by an attribute that we created earlier, which is density, the one that we created right here in the set properties of the creature. So you see the F density, this is where it's going to go. So this 200 value is going to multiply by the density attribute that created earlier, and the thickness is going to be uniform for all the points. So if you zoom in and visualize the thickness, you'll see the radius, the collision radius of these points. we're gonna hide this and for as for the stiffness we're gonna just the stretch stiffness we're gonna set it to 0 0 0 1 and scale it by attribute which is going to be the stretch stiffness the default for the damping ratio we're gonna set to 0 0 15. The damping ratio is a parameter that will try to stop any unwanted energy inside the tetrahedral constraints so any jittering and any unwanted movement or like really fast movement, it's going to try to calm it down. It's like an internal friction force. And this is a value that we'll play around with a lot. I mean, I play around with it a lot because it helps with achieving really stable simulations. And we're going to rename this constraint with CSTR, that's stretch. So our CSTR, the tetrahedrons will be, tetrahedral stretch constraints will be saved in this primitive group. Also, we apply this to surface triangles. We have to delete this. And we're gonna apply it to the GRP tests because we are applying this to the internal tetrahedrons. Now as you can see, if you dig into, go into inside the geometry, you can see that like you have kind of a web of primitives that are constraining all those tetrahedrons together and making one single soft body. Yeah, we're done with this tetrahedral constraint. And we're also going to create, let's say, a Vellum constraints, press J, drag the cursor like this, and go here and create a tetrahedral fiber. And the tetrahedral fiber constraints is a special one that's really underrated, one of the most underrated constraints in Vellum and in Houdini in general, because this is what creates constructions expansions of muscles and it's used a lot in the new muscle system inside Vellum. If you click, if you just search the muscle for the muscle nodes you can see that you have muscle properties, certified, solver, Vellum. These are all related to the new muscle system for the Vellum solver and it's quite a new thing and it's not, it's still kind of in beta but it's really promising and I think in the future I'm gonna be using it really extensively. But yeah, maybe we'll explore it in another tutorial, but this is a really complex one. So let's get back to our dead fiber constraints. And that fiber constraints will be applied to the legs primitive group. And you can see that the legs are really only the legs are constrained here. And the thing is we are not going to apply this to the primitive group, but to the point group. The one where the tips are not included. So if you remember we had, if we scroll like middle mouse click onto our vellum cloth, vellum constraints, you can see that we have a GRP legs which is a point group and a GRP legs primitive group. We're gonna use the one that is in the point section, so which doesn't contain any tips. We want to compress only this part of the leg. Along the vector that we created earlier, if you remember we created the strider arm. This is the vector that is going to be used as a compression vector for the tetrafiber constraints. The tetrafiber constraints are really kind of a dumb type of constraints, it just compresses these muscles in this specific direction. It uses, by default, it uses the material w attribute, which is a vector attribute, and it's going to define the direction of compression. Also, we're going to set the stiffness to zero because we don't want to to have any initial stiffness or initial compression and we're going to set this to 1. Also for the damping ratio we don't want vibrations or jittering so I'm going to increase it to 0, 0, 3 and I'm going to create an output group which is going to be called CSTR, that's fibers. This is it, this is done and now the core of our simulation, the Veyron solder. Press J, drag to the vellum constraints, and let's rename this vellum fibers. vellum bits. vellumcloth. This is done. So now, I'm going to activate the vellum solver. It's going to think for a second. I'm going to hide the visualizer for the arm vector. I'm just cut this connection. In order to cut the connection, you press Y and just drag your cursor above this line, on top of this line, and it's gonna just cut it. I'm going to connect the ground directly here. So we have the ground visualized in our SIM. And yeah, now for the interesting part, because this is where everything is gonna get interesting. We're gonna dig into the solver. When you double click onto the development solver, you'll find yourself in a dopnet. Basically, any SOP version of any solver contains inside a top version. So it's not a surprise because you can go back, like if you want to go back to the origin where you were before, you have to go right here. But if you double click and go back to the dopnet, you'll see that it's a really simple, I mean, it's a regular dopnet with external static objects to the ground lane. Everything is linked together. You have the forces, you have everything. But yeah, we're gonna end up here. So the force output is for connecting any microsolvers and the source same for connecting microsolvers, but they're related more to the sourcing. So if you wanna emit any vellum objects, emit a bunch of spheres, a bunch of objects from the same point or any geometry, this is where you go. So we're gonna activate the force and we're gonna start creating our microsolvers that we're gonna animate the whole thing. I'm gonna deactivate my, maybe it will work like this. So, before adding any microsolvers, we're gonna go back to our subcontext where we created the ground and we'll start building our main attribute VOP. I'm gonna do it, the main for that will create lock motion is going to be inside an attribute VOP. As I told you before in the previous lesson, we're not going to spend time simulating inside Vellum and creating this force, tweaking it inside because I already know how it's supposed to be. So I'm going to create this attribute VOP. In the first input I'm going to connect the master transform and in the second the ground force. This is going to be just for testing purposes and for building purposes and you'll see why. Then we're gonna transfer it into the DOPs. So yeah, the force search algorithm is based on a simple PC Open, which is a point cloud open. So the points from our creature are going to seek points from the ground. So each point will seek the closest point or a closest cluster of points that are defined inside the PC Open, which is 10 points. So each point here in the creature is going to be searching points here on the ground, like the average of 10 points in a radius of 0,1, which is 10 centimeters. But we're going to increase this radius to 0,5 because we need some search radius. So we need a lot of radius for the points to find anything on the ground. What we're going to do with this PC Open, we're gonna find points on the ground and see what are their attributes, especially the velocity vector and the normal vector. So we're gonna find those vectors with the PC Filter, and the PC Filter goes, the handle of the PC Filter goes into the wind cloud open, and we duplicate this in the first one, we're gonna search for the velocity vector, and the second one we're gonna search for the normal vector. we're going to normalize these vectors. Normalizing vectors means that you're basically fitting the magnitude of the vector into a range of 0 and 1 between 0 and 1. And this is for precision purposes. We need to have normalized vectors whenever we're using these kind of forces, because we don't know what's the magnitude. I don't want to have any additional magnitude and have chaotic control over those forces, so I want to have vectors that have components fitted between 0 and 1. As you remember, we had a CD attribute and we had a CDR attribute. So we either import them with bind. The CD attribute is already here, so we're going to do a vector to float. So we're going to just take one component of the CD attribute and we're going to import the cdr attribute which is also a vector. I could have done a float attribute but it's not a lot of overhead because we don't have a lot of attributes there. cdr attribute is a vector attribute. We're gonna, or maybe not, I don't really remember. We'll have to go back see what we've done with the cdr attribute. Oh, so yeah, I created a vector of cdr attribute. So let's make it a float. fcdr is great. Here we're going to initialize it to the fcdr and instead of setting a vector here we're going to just set it to 1 to 0 and the same for this one. So we're gonna have floats. cdr is a float, cdr is a float, cdr is a float, everything is good. We're gonna come back here, we have cdr as a float here and the color is a vector attribute. So this is a local test attribute WAP. So the CDR is going to be a float. Let's save on those useless components of the vector. It's going to be a float which is lighter to read and right and everything else. So yeah. So how does it work? This is going to be our creature. This is just one of the tentacles. This is going to be the tip of the tentacle. the bulk of the body. So we have a frontal force that is going to be applied to the front, which is going to be inherited from the ground. So the ground is right here and it has a horizontal z-axis force which pushes the tips of the legs forward in this direction. But we also have a normal force which is going to go up or down. So whether it's up or down, it's going to be defined by the CDR attribute. So the CDR attribute is a color that turns from 0 to 1, but we're going to map it to something like minus 1, let's say n1, so it's going to be remapped from 0 to 1, between 0, like it was 0 to 1, and it's going to be remapped from minus 1 to 1. We're going to have a normal attribute that is going to be multiplied by minus 1 or by 1. So in the beginning there's going to be a normal force applied up and then a normal force that will be applied down. So the in-between is going to be looking like this. So the vector, the resulting vector is going to be forward plus up so it's a diagonal line and then the CDR is going to be multiplied by minus one and it's going to go down. So the tip will be here and then we'll end up here. So the The final movement of this tip is going to be a curved motion like this. So the curved motion will be defined like this because the CDR attribute is going to be turning on and off and its values will be mapped between minus one and one. So either the normal will be added as a positive value or a negative value and then a negative vector. So it's going to result in a curved motion over time. you'll see how it works in our system. So in order to implement this, we're gonna just take the CDR and fit it, fit range, between the zero and one, and then the final will be minus one and one. So now we're gonna multiply, because this is, we're gonna multiply the normal vector by this, by the CDR. So the up and down is controlled by the CDR, and we're gonna add this to our velocity vector that we inherit from the ground force that we created earlier. So we add this to the normalized vector here. And we're going to multiply the overall force by this color because we want to turn on, oh yeah, I forgot to connect the normalized version of the normal vector. So we're going to multiply the result of this addition of two vectors by the color. Because we want to buy one of the components of the color, so the red channel, for example. So we're gonna have a resulting vector which turns on only when the tips are white because once it's white, the overall vector is multiplied by one. Once it's black, it doesn't multiply by anything, it's zero, so the vector will be zero, so nothing will happen with these tips. And the product, like the safest way to add forces in Vellum, in my mind and how I've done it before, is adding any force that you create, you add it to the existing force. So if there's any residual forces or previous attribute VOPs that you've done before, this is how you add your additional forces on top of it. You take the existing force, which is cumulative. If you only have any forces existing in the system already, you take the previous force and you add your additional forces to it and you add it to the force. So you don't connect it to the velocity because you'll have unexpected results there. And I'm gonna press Shift S just to change the type of lines that are connecting our nodes. So here we go. This is the force. And we're gonna also create a bind export. I create temporary force attribute, which is gonna be called FRC. I do this attribute I create, this attribute always. Every time I create this attribute is for debugging and seeing what the force does. Also, I've never understood the difference between tree floats and the vector, because they're both vectors, they end up being like three float, vectors created by three floats, three components. I use either of them, it's the same for me. So yeah, we're gonna output FRC, and I'm going to debug this FRC force to see what it does. So the FRC is there. Hide the deer arm and just add a marker. It's gonna be a vector trail. Yeah, and make it kind of blue, let's say. So yeah, we have the CD, we have all the good stuff here. And oh yeah, I forgot something, I forgot to make a overall multiplier. So we're going to multiply, like you first most click on the input tree, make a promote parameter, double click on it. We're going to rename this. MoldOverall. This is going to be the the overall multiplier that adds forces to this creature. Let's put it to one, let's say. Let's check that we have normals here. Oh, okay, yeah. You see the normal is a vertex attribute and we have to go back and make it a point attribute. So, as you can see, once it turns white, because we're going to apply this only to the tips, we have to apply this only to the tips. Let's see the tips. As you can see, what happens is the forces applied up or down. So every time there's a color turning on, there's a vector that's applied on it. So let's see what happens when we go into the solver. Let's do our first simulation test. The multiplier is going to be really, really low. Let's start with the value of 0, 0, 0, 5. And it's a little bit barely visible, but I'm going to debug this unmultiplied force, because I want to really see what happens with the vectors. We're going to save this. And we're going to dig into the valence solver and do preparations for the attribute transfer. One thing to mention, which is really important, is that any attributes that come here into the valence solver are not updated automatically. So anything that goes into the solver, you have to update it. If you have any live attributes that you want to feed into the geometry during simulation, you have to do it inside the valence solver. You usually do it in the sub solver. So we connect the subsolver to the output, to the force output, and we're gonna dig into the subsolver and you'll see what we do here. You have already an output here. We're gonna object merge our rest attribute, or a rest null that we had before, out rest into the specific object, and this is this object. So we have our rest static geometry with live attributes, with all the attributes that we need. We're gonna attribute copy. Because the monster is the only thing in the simulation, There's no other geometry objects that are simulated. Where we can do the attribute copy because the object merge will only, like, dop geometry matches the object merge. These are, this is the only object that exists. In the simulation, we have one object and the object that we merge is one object and it's the same object. Sometimes you have emitting geometry and you have several objects in the geometry. You have to do groups to attribute copies, specific groups to specific groups. Otherwise, it's just gonna say that the point number is different on the Dov geometry and on the rest. So we're going to copy CD and CDR because those are live attributes and let's also import the ground because we have the ground outside the Dobs, we just import the ground force and we're gonna copy and paste our test lock emotion inside this sub-solver. This is where the forces will be applied, we'll do it the same way in the first input you have the geometry that comes in and in the second you have the ground and shift s just for the lines to be look neat and yeah for now it says that we don't have any tips group but you'll see that usually this is what it does in the first frame of the simulation don't be alarmed by this or villain does this often it just doesn't evaluate the first frame properly sometimes or some Some groups appear dynamically during the simulation. It depends. There's many scenarios where it's going to say invalid group, but once the first frame is simulated you'll see that everything is fine. So let's see what it does. So let's press... I didn't change any settings in Valem first, like before. But you'll see that something happens. You see that one arm is extending, but it extends too much. So these guys were... The creature falls down. We're gonna disable the vectors here. You'll see that once one of these tips goes white, it's going to just be pushed up forward and down. So yeah, we see that. Okay, but the thing is as you can see that it's really a really soft body and the arms they extend too much. This is because We don't have enough sub steps here to satisfy our constraints, so in order to satisfy those constraints we'll add more sub steps. In my first simulation I used the five sub steps for tetrahedral constraints you'll have to aim for five tetrahedral constraints, five sub steps I mean for the kind of sub steps. I usually change only the sub steps, I don't touch the constraint iterations, because these are more like it's more of a secondary thing. The the sub steps are the main parameter that defines the quality of your simulation and also the satisfaction of your constraints. Because sometimes you'll find yourself setting some really high values here, the higher you go, the more sub steps you need to satisfy those constraints. If they're unsatisfied, they will behave like constraints that have less stiffness than what you've set. So you have to keep in mind for this. In the forces tab of our Valum solver, I'm going to call it Valum main, I'm going to set the static threshold to 0.01 and dynamic scale to 0.01. I'm trying to set the static friction and dynamic friction as low as possible. This is over or multiplier over the dynamic and static friction. The gravity is alright, everything is fine. And that's kind of it. And you'll see that once we change the sub-step to 5, the arms are not going to extend as much. The shape, the stiffness of the creature shape is going to be higher. So the arms are not going to extend as much. They'll try to restore their initial volume, even if the force is kind of strong. Yeah, so you see it simulates, it's much slower than before, but because of the sub steps. But believe me, it's much, much faster. It's like 10 times faster than the FEMS solver. We will use the FEMS solver later on for wrinkles, and I'm going to build a system that creates wrinkles on top of the existing valum simulation. And you'll see the differences between the solvers. We have here like 250,000 tetrahedral primitives, which is not a small number for a creature like this. But so this is it will take some time to simulate. So but it's not a big deal because we can always launch wedges which is a new system in Houdini a relatively new system the top system where you can launch a lot of like different variations of your simulation. And yeah it's a we can do it overnight and launch a bunch of simulation overnight. So as you can see it's more rigid than before, the legs don't extend as much. Now we're just working on the tips side of the dynamics, you see the tips are turning white and they're kind of pushed, pushed up and forward. So for me it's, it's alright. The only thing I will change here is go into the sub-solver and we're gonna change a little bit the multiplier, 0, 0, 0, 0.06. We increase the force a little and we're going to go into our locomotion test here and we're gonna fit the the CDR attribute to something like or maybe not we're gonna leave it like this. We may probably We'll leave it like this, I'm not going to change the fitting here. The fit in the CDR section defines how much of a multiplier you apply to the normal component of the final force when you want it to go up or down. So basically if you want the arms to attach faster and fall down onto the ground faster, the tips will fall down faster, you want to increase this value. minus 2 is going to make the tips really go down and attach to the ground faster than before. Also if you want to have more of a vertical jump for those tips, you can increase this value to 2 or higher and you'll see what it does. You can experiment with these values, but I found that these are kind of balanced, so this works fine for this scenario. Now we'll have this simulation here. So the base locomotion system is done. It's just the tips, but we have much more, many more Microsofters to add here. The subsoil is the first one. And we'll see about that in the next lesson, because I think we've already passed the length threshold for this tutorial. I try to make it faster, but there's too much information to deliver. So I think I'll split it more for the next lessons. See you in the next one."
+    }
+  ]
+}