Collisions v2 Blender Geometry Simulation Nodes Tutorial How-To

Рет қаралды 9,478

Күн бұрын

💗 LIKE, COMMENT, SUBSCRIBE - THANK YOU! 💗
💥 ChrisP's Feature-Packed Startup File for Blender: • ChrisP's Feature-Packe...
Realistic Material Packs for EEVEE+Cycles: chrisp.zone/blendermarket
🔥 Note: "Gold Monkey" Patrons get free access to all existing and future material packs!
Patreon / chrisp
Blendermarket blendermarket.com/creators/ch...
Brave brave.com/chr471
Twitter / chrisprenn
Insta / chrisprenn
Home chrisp.zone
0:00 What to expect
0:07 Blender 3.5 Experimental Build
0:28 Fix GeoNodes
1:24 Collisions using Raycast node
5:57 A closer look
8:37 Outro

Пікірлер: 38

@chrisprenn Жыл бұрын

I finally got around to publishing this video - a follow up to the last tutorial simplifying the collision calculations using the "Raycast" node in Geometry Nodes 💪

@xaviervicens5217 Жыл бұрын

You are on fire Chris. Almost a video x day. Happy to see all of them.

@shmuelisrl Жыл бұрын

you can still get more accurate without sub-frames by using the length of the hit and the hit position, but I don't think it is worth it (not that I think it would be particularly hard, but) because it is also perfect the way it is now. (just pointing out that you can probably get actual perfect collision detection without sub-frames)

@chrisprenn Жыл бұрын

Think so too 😁

@InterruptingOctopus 2 ай бұрын

how would you do this? because i do actually want to get more accurate collisions. a bucket filling animation wont work well if theres holes where there shouldnt be after all

@zboy303 Жыл бұрын

Another great method! Interesting to see how differently similar results can be achieved.

@TheKrisnag Жыл бұрын

Superb Chris.....How to display velocity text / Height text on the or any data text on the moving points ??

@johnsmith56920 Жыл бұрын

Done! Thanks.

@thedarkmatterplanet 10 ай бұрын

Geonodes REALLY needs some quality of life improvements. I'm happy and appreciative about what's possible, but node setups like this are what is driving me toward Houdini every day. Hopefully Blender Foundation continues the paradigm they started when they included the Geonode setups for the new hair system, but I won't hold my breath.

@SjonSjine Жыл бұрын

Thanks Chris, like your tutorial! Man those Simulation Nodes / Geometry Nodes - imagine that this is only the beginning of something real big! I was wondering, is self collision between particles/ real instances possible and between two emitters??

@KDawg5000 Жыл бұрын

Very nice! tnx

@FarukAhmet Жыл бұрын

Oh for crying out loud... I never use the Add menu, my right mouse click is set as a shortcut for node search instead, but the Simulation Output doesn't have the "add this first" warning in the title in the search menu so I 've been banging my head against the wall trying to make it work for the last two days. No sim frame, files crashing... Thank God I saw this haha. Perks of alpha software I guess.

@chrisprenn Жыл бұрын

I like using menus especially for tutorials 😬

@KDawg5000 Жыл бұрын

I had the same issue a couple days ago. I thought the Dec 10th simulation nodes were broken, until I saw a vid where someone said "you have to add the output node first". Arghh! 😄

@Unleashed75 Жыл бұрын

Wouldn't the RayCast Length give you a measurement too? So there must be some math to say where it hits in 3d space and where to bounce from, instead of just bouncing from the frame it's currently on. Then you should be able to calculate how far it has to travel before it Hits and how far is left after. I hope that makes sense. Great video again and much simpler node setup GG

@dhw9406 Жыл бұрын

The math should be something like newVelocity = energyLoss * reflect(oldVelocity, hitnormal) newPosition = hitPosition + newVelocity * ((length(oldVelocity) - hitDistance) / length(oldVelocity)) basically, the ((length(oldVelocity) - hitDistance) / length(oldVelocity)) is the amount of 'motion' left in the particle after it's traveled to the hit point in the subframe; this should be a number not a vector, hence the length calculations. the direction comes from the vector newVelocity

@themostchillbro Жыл бұрын

Amazing tutorial! I am trying to have this collision inside another object so they bounce around inside, but the points jitter at the boundary until ultimately "breaking through" and flying away. Seems like the reflection may not be working but unsure how to adjust.

@chrisprenn Жыл бұрын

Not 100% sure now with the raycast method but did you try flipping the normals of your object your particles are "inside" of?

@themostchillbro 3 ай бұрын

@@chrisprenn I will try, thank you!

@alexchazavi8244 Жыл бұрын

Hey chris, amazing video as always. I have a question, it's possible to make particles system in 3.4 or I need 3.5?

@chrisprenn Жыл бұрын

Simulation nodes are only experimental and in 3.5

@ruslandad365 10 ай бұрын

GREAT SOLUTION!!! And subframes are not needed at all Ideal particle system!!!

@seanmurray170 Жыл бұрын

Can node simulation work for particles that detect collision among themselves?dropping beads into a bowl that bounce off each other and collect at the bottom? Basically like a rigid body replacement?

@chrisprenn Жыл бұрын

for spherical particles you could calc that but for other shapes you need more sophisticated physics sim

@seanraymond7343 Жыл бұрын

Okay.. this works and all.. but how can I replace the points with an Object? Blender 3.6 crashes / stalls every time I try..

@chrisprenn Жыл бұрын

what kind of monster object are you trying to instance onto the points? ;)

@seanraymond7343 Жыл бұрын

@@chrisprenn I fixed it, I was trying to put a realized ico sphere on to the points before the gravity effect. I finally realized that it should be done after the simulation. now.. how do I get a point to spawn new reflected points after the ray cast calculation? hrrmmm

@RoyMendezCastro 9 ай бұрын

Piensa en hacer un tutorial de simulación de túnel de viento para diferentes modelos. No he visto ni un solo tutorial que lo abarquen, podría ser con simulación de humo y generar lines de velocidad con flechas para visualizarlo mejor. Tiene aplicables de diseño aerodinámico y de análisis de ventilación de diseños de arquitectura

@MrTomyCJ Жыл бұрын

Don't you find it annoying to work with the new viewer node? Most of the time I do not want to view the connected geometry on the viewport, just the spreadsheet. For that, now I need to do a couple clicks more, and I often need to change the domain in the viewer node, it's really annoying :(

@B-2-Z Жыл бұрын

About the threshold problem: couldn't you treat that manually? Simply subtract the distance from your velocity and reflect what's left? That should give the proper reflection without any threshold.

@davidbenoit919 Жыл бұрын

Does that mean the raycast node calculates if there is a collision in between frames? I don't see any node calculating if the distance is less than x, consider collision

@chrisprenn Жыл бұрын

yes - kinda - all collisions happen "between frames" since you will never have a frame be exactly at the right time where the particle is exactly hitting the surface. So for each frame - before we move the particle - we calculate if it hits, and if it does, then we set the particle to the position of the reflection vector.

@davidbenoit919 Жыл бұрын

@@chrisprenn Thanks for your answer Chris. I just realized while writting a reply that this distance is actually the ray length that you plug into the raycast node. Now everything makes sens!

@shmuelisrl Жыл бұрын

yes. it casts a ray in the direction it is moving and the distance it would travel in this frame and if it hits (meaning it will hit in this frame) then we reflect it. we don't need a comape node because the ray length does that for us already p the is hit ( for @Chris P : so if you compare the distance of the hit with the ray length you can change the velocity (strength/ magnitude) according for the reflect to get it more accurate, and set it to offset from the hit position instead of what it is now)

@chrisprenn Жыл бұрын

@@shmuelisrl yeah but keeping it simple and with the scale to .7 it looks just fine 😉

@shmuelisrl Жыл бұрын

@@chrisprenn I mean scaling the velocity before the reflection not considering the dampening. (still you wouldn't gain that much more accuracy than you already have)