I think this is very interesting how it can show the fluid moving in real time. From what I understand, the particles are either smoothed particle hydrodynamics or position-based fluid. The interesting part of this I that it calculates the movements and interactions of particles over time.
https://www.youtube.com/watch?v=rSKMYc1CQHE&t=66s
This video by Sebastian Lague talks about how we can build a water simulator and it is worth checking out.
spegeerino
@sparky-ed I read through some of the preprint of the paper and it looks like they used position-based fluid. You're right; they do simulate a lot of particles and computes their movements and interactions (applying forces to particles when they get too close, and things like that). The innovation in this paper seems to be the idea of forcing constant density throughout the fluid through some additional pressure forces, which allows the fluid to more convincingly look incompressible, like liquids in real life.
maxwelljin
I am wondering if the method needs to calculate the movements and interactions of particles, how we do it in real-time. To achieve high resolution, it requires many fine-grained particles. How do we accelerate the simulation process?
Refangs
I'm also curious how fluids are efficiently modeled and rendered in real-time, such as in a video game. Maybe the algorithms themselves aren't that computationally expensive, or there are some shortcuts?
theflyingpie
For simulating fluids, I would imagine the number and size of particles might play into the viscosity and perhaps other properties of the liquid. To render a complete, continuous body of fluid, is there some kind of interpolation performed between neighboring particles?
muuncakez
Similar to capping the count of light ray bounces, would simulating fluids also require a cap the count off particle clashes as waves will crash into itself, "other" waves, and any other object on the surface (like a rock) which could cause the water particles to "bounce" and crash back into other things? And how small is that cap before the waves look kinda funny? and what physical object can waves be deduced to to make simulating the waves easier (like cloth as springs) ?
I think this is very interesting how it can show the fluid moving in real time. From what I understand, the particles are either smoothed particle hydrodynamics or position-based fluid. The interesting part of this I that it calculates the movements and interactions of particles over time.
https://www.youtube.com/watch?v=rSKMYc1CQHE&t=66s
This video by Sebastian Lague talks about how we can build a water simulator and it is worth checking out.
@sparky-ed I read through some of the preprint of the paper and it looks like they used position-based fluid. You're right; they do simulate a lot of particles and computes their movements and interactions (applying forces to particles when they get too close, and things like that). The innovation in this paper seems to be the idea of forcing constant density throughout the fluid through some additional pressure forces, which allows the fluid to more convincingly look incompressible, like liquids in real life.
I am wondering if the method needs to calculate the movements and interactions of particles, how we do it in real-time. To achieve high resolution, it requires many fine-grained particles. How do we accelerate the simulation process?
I'm also curious how fluids are efficiently modeled and rendered in real-time, such as in a video game. Maybe the algorithms themselves aren't that computationally expensive, or there are some shortcuts?
For simulating fluids, I would imagine the number and size of particles might play into the viscosity and perhaps other properties of the liquid. To render a complete, continuous body of fluid, is there some kind of interpolation performed between neighboring particles?
Similar to capping the count of light ray bounces, would simulating fluids also require a cap the count off particle clashes as waves will crash into itself, "other" waves, and any other object on the surface (like a rock) which could cause the water particles to "bounce" and crash back into other things? And how small is that cap before the waves look kinda funny? and what physical object can waves be deduced to to make simulating the waves easier (like cloth as springs) ?