Video

I'm confused on the part where you reverse the distance and normal if the face is facing toward the origin. There is an edge case where if I have two cubes that are aligned on two axes and they collide, reversing the normal seems to be causing problems giving me the wrong normal. That's because the smallest normal becomes largest normal I guess, but I don't want to just cut the code out without understanding why it was there to begin with.

Can you explain why 3:06 for SphereToPlane you pass a *tb Transform pointer as Plane Transform struct ? I cannot find this part of engine on github?

What is the best way to resolve 20k objects' collisions?

This makes me feel like a "Welcome back, to the SandPond saga. Let's get started."

Good video, and I could understand it. But... However... Why are you speaking as fast as possible without any space between your sentences?

Where were you back in 2013 when I was writing my Terraria world analysis software that depended on computing biome extents using these algorithms... Thank you for such a clear explanation. This is amazingly informative and very useful.

looks really cool! you gonna release it on steam soon?

How did you tackle the rotation problem? That’s something I’ve been trying to solve on my own for quite some time!

This is SO cool

Wow. Thank you very much for this video!❤

how do you handle updating the dirty rect if two particles want to occupy the space but only one gets chosen?

Don't you also need the full set of contact points?

When you say that you iterate the direction D. How you do that? How many directions you use? Which values?

Please dive deeper into the maths behind the implementations.

are u still there? beacause there is not any video while two year

Pretty neat but a little info and code on optimisation would have been really good.

I tried to download the code from and GitHub and tried to run the build file but it throws some errror

It might not be stable, floating point comparisons <-

To implement the EPA, do we need to have all the possible Minkowski difference points?

That... explains a lot... A while back, was trying to figure out why all the collision detection stuff was reactive to objects clipping together. Apparently, this is the norm. I wonder if I can make it more proactive. Maybe with a raycast?

;)

Pretty sure there is an error in formula @2:20. We want max(D*A - D*B) => max(D*A) - min(D*B) => max(D*A) - (-max(-D*B)) => max(D*A) + max(-D*B). This is using the identity that min(f(x)) = -max(-f(x)). I think you just merged the minus signs on accident. What are your thoughts? Am I missing something?

...Is collision detection between primitives really as simple as separate functions for each possible interactions between all the different shapes? Is there not a better, more easily extended way to do this?

Very interesting... thanks

i love you

came back after 3 years to this channel, looks like you still going strong. looks very nice :)

This can benefit so much from ECS

Ugh. Too. Much. Math. I think i’ll just stick to iterating over my edges to find an overlap

been a while, wanted to say this looks really cool and would love to see more of it

I miss your videos!

Great video! But I have a question, how did you do the graphycs of the video? Like the dynamics title and those trancitions. I want to do videos like this, but I don't know how to make them visually apealing

A saga perhaps?

Thank you

I might sound silly but can we have a python version of this😅

how do you construct your vector3 union/struct/class?

The Cherno sent me here.

Don't forget quadtrees!

How to find direction in supporting point?

excited for the next devlog

Hello Winterdev. Thanks for the great Video. When checking what faces to delete you check if the normal of the triangle and the supportpoint are in the same direction. This only works for some cases. Since you want to check if the support point is on one side of the triangle you want to use the supportpoint relative to the triangle. So the check would look like this if (SameDirection(normals[i], support-polytope[faces[i*3]])

Wow!! Thank you for the easy explanation.!!!

Not to be a jerk but I think it would be valuable to clarify that this is really the sub-variant of the traditional GJK algorithm (GJK-SAT). As well as this, the number of cases in three dimensional space actually expands to considering all regions rather than just the faces of the simplex. I hate to be rude but I spent way too long figuring this out when attempting to implement this myself. Gino van den Bergen's 'Colllsion Detection in Interactive 3D Environments' documents the GJK as well as Gino's EPA rather extensively, and I think the full distance minimization implementation is insightful for those just figuring this stuff out. Either way, this video does a fantastic job explaining the essence of the support function and the configuration space obstacle. I can't wait to see what else you're up to!

Winterdev: Codes a physics engine in 5 minutes, provides glimpses of classes but doesn't provide the source code Me: Can I see the full source code, so that I can run and experiment with it? Winterdev: NOOOOO, don't you know how much TIME and hard work went into this??? How can you put such high expectations on me???

It is the first example of well defined problems with this kind of physics mechanics. Loved the way you explained them <3

great video ! Unfortunatly the full article is down :c

An alternative approach for handling particles in a efficient way is to maintain 2 collections : one is a 2D array that map a x/y position to a particle (same as what you describe in the video) and the other is a array of particles. Each particle contains a type and a x/y position. It can be a struct to avoid creating many objects. To update simulation, you simply iterate on the particle array. it automatically skip all empty grid cells and prevent you moving a particle more than once.

Going to make any more videos? How about a github link?

Definetly has a cool foundation. Are you still working on this? Can you upload the project to your git? :)

Hey man, I just noticed that your Description is not Algorithm Optimized. The first few words in your Description are used by CZcams to help people find your video (CZcams says this in the helper text next to the input box) So, I think it should have keywords and be descriptive of your video, such as: "How to Make a Falling Sand Game in Unity.". Something like that. Just thought I'd tell you!

Have you ever considered using compute shaders for sand logic simulation ?