[flow_default] Transcription: 01-Substeps Drag Constraints.json

transcriptions/01-Substeps Drag Constraints.json

@@ -0,0 +1,7 @@
+{
+  "audio_file": "01-Substeps Drag Constraints.wav",
+  "text": "So the reason this is failing right now and the solver isn't able to compute all of the elements is we have too much points right now the solver is solving points so it's calculating how every point moves and how it collides with the next point as well as how it collides with the grain, so not also the bag, everything is being solved simultaneously all together at the same time. And as I've said, we only have one sub step right now. And what this means is every one of these points is being solved just one time over every frame. So every frame, we have like 20,000 points right now that includes the bag and the greens and the solver runs them through and computes the movement only one time every frame. So it makes sense that the solver is failing because these are too many points to be solved only once. So let's try to crank this up to say five sub steps and see what this gets us. So as you can see by raising the sub steps to 5 instead of 1, we now have a much more stable simulation. The grains come down, interact with the bag and the bag resists. So it results in this bouncing motion up and down. So 5 sub steps worked here because that means in every and each frame the solver instead of going through all the calculations one time it now goes and calculates everything five times in each frame so it's literally being smarter in the way it calculates every point and how they should move so the more sub steps our solver has the smarter the solver is and much more capable of producing natural results. So this will keep bouncing up and down forever until we run out of frames. So what we need to do is introduce some kind of force that inhibits the velocity of our bag. So it reaches zero over time and it loses its velocity as we should expect to happen in the real world. So we have two ways to make this happen. Either use this velocity damping parameter right here, so we can give it an extreme number at first to see what it does. So this is a vertical number that I typed here, 0.5 is too much, but it will allow us to see the result of the velocity damping option, which is literally the bag is losing its velocity over time so this is the result and it's a very radical result. You can use this if you want to it would be a sure way to make your object lose the velocity over time but it's really it really will look natural realistic so let's do the other thing now, which is using a pop drag. So let's get a pop drag node, hook it to the particle forces input. And let's get rid of this ignore mass parameter since we have added our masses manually. This ignores the mass of all the objects involved in the solver. So this irresistent parameter, which is resisting the movement of her bag and the sand is being forced over everything that we have in the solver equally but since we have some elements that are heavier than others so we don't want to ignore our masses we actually we actually want to affect our masses in a natural way so things that are heavier are harder to get to stop. So let's take this down from 1, one is too high right now, let's say 0.25 and see what this gets us. So what we have right now is even more stable than before, as we can see the grain comes down and the back bounces as before but it slowly loses its momentum over time. so in 72 frames it loses so much of its momentum and if we keep going it will eventually result in a complete stop which is what we want and what we would expect in a real scenario like this in the real world but we still have too much of a bounce from my taste and let's see what we can do to combat this even more. So let's go to the Solver and let's go to the Advanced tab. Let's scroll down to the Green Collisions tab. And we have this parameter that's called Repulsion. So Repulsion controls how strongly the particles are kept apart. So the higher the repulsion, the less objects bounce. And this doesn't just control the grains this actually controls anything in the solver since the solver is solving any element that's fed into it as points it doesn't care about your wireframe or anything it deals in points basically so this controls any bounce between any point and the next so I like to abuse this actually. So let's crank it very high like this. And we can boost our sub steps even more. Let's go with 10 sub steps. And while we're in the iterations tab, let's boost our constraint iterations as well. What this is is literally, you can think of it as sub steps, but only concerning constraints. So it goes over just our constraints and for each sub step we have 100 constraint iterations. So now our constraints are being solved a thousand times for each frame, a hundred times for each sub step and a ten sub step for a frame. So a thousand times every frame and more stiff constraints require higher values and we do actually have stiff constraints since we made our BendStefness.01 which is considered somewhat stiff. So let's take this up to 150 and see what this gets us. So as you can see, we now have a much more stable result. We can always go with more sub steps if we want and I suspect we will do that before the end of this tutorial. Things don't bounce as much in the real world as you would expect or as you would get normally with a 3D simulation. So this is a very realistic behavior we have here we have the bag loses its momentum very quickly before even two seconds and it comes to a complete stop which is what you would expect in a realistic scenario. So since we now have a very stable simulation, let's start working on tearing our bag apart so the sand can escape.",
+  "language": "en",
+  "confidence": null,
+  "duration": 442.94
+}