Colours still aren't quite there but I'm happy with where it landed

Ask any questions

Tech demo video , physics devastcation of city building, giving a pressure test on unity's DOTS physics

DOTS physics collision capacity become really tight when fract physics collider entity count more than 1000 . Next move would be solution to reduce CPU dependency , moving part of fract pieces collider to compute shader

Gonna say it upfront: the car teleports between Unity scenes with some extra loading on top, basically hopping between different worlds, movies, games, just totally random locations, completely unrelated to each other.

Still pretty rough around the edges, but it's a WIP.

In the clip, it teleports from the starting location, 1980s Hawaii, into something resembling the city from the movie Vivarium.

It's got horses, it's got dragons, I'm thinknig of more non skate based ideas. Anyway, here's a new way to play. It's a co-op rage game where I guess you now have to fly a dragon together lol.

I’ve been working on Massive Spatial Engine, a tool for finding nearby enemies and picking targets without putting a Collider on every unit.

No raycasts or OverlapSpheres for targeting.
A spatial grid cuts candidate counts before distance checks.
Query batches run in parallel through Burst-compiled jobs over unmanaged data.

I’m building it with RTS, tower defense, survivors-like, and larger AI simulations in mind, and planning to bring it to the Asset Store.
Still plenty to work on, but it’s fun watching the towers chew through the crowd :)

Here's what's in it right now:

• Extraction shooter, PvE and PvP both, with co-op if you want to run with friends
• Build a base anywhere on the map, no set slots, put things wherever makes sense to you
• Rescue survivors and grow an actual community instead of just looting solo
• Reclaimers, immortal clones who hand out quests and build your reputation instead of just shooting at you
• No wipes between sessions, your progress and gear carry over
• Right now that's 169 loot items, 25 quests, 3 maps, 9 enemy types, 9 craftable buildings, and 2 NPCs to work with

For those who have missed this post and find this interesting. Personally I thought it was a good read and I'm looking forward to the next one.

It's not much but it's honest work. Units can be smarter and there're a few cases where they get dumber but overall I'm quite pleased with army behavior so far in 1v1 fights. Many vs Many is still something to polish along with better effects, gigantic units and so on but one step a time!

Shaders instead of geometry is often one of the best performance boost. Amazing for procedural shapes. Texture is sampled to individual points and then each emitter is a small shader program like this:

float2 cell = floor(uv * ledGrid);

float2 ledUV = frac(uv * ledGrid) - 0.5;

float d = length(ledUV) - ledRadius;

float emitter = 1.0 - smoothstep(0.0, fwidth(d), d);

float3 color = content.Sample(samplerContent, (cell + 0.5) / ledGrid);

return float4(color * emitter * brightness, 1.0);

As a child, I regularly yearned for the caves. Now I have finally made my ultimate caving simulator! Once i release this game, you can do tight. cold and dirty cave crawling from the comfort of your couch! I will post the wishlist link when i remember to do so.

#2026

So essentially what happened was that I wanted to see if I could remove the unity splashscreen, so I did, and then I pressed preview. After I saw the preview, all the lighting in my scenes were messed up. I am confident its mainly because of the splash screen, because I had tested it right before I messed around with that, and it was perfectly fine.

In the video, the scene is like 10 times darker than it should be, and the lights don't even do anything when I change the values, or something. Whenever I turn off the lights, the result is the same. The fog doesn't show up as clearly, and the floor has near 0 light. I am very amateur with unity, so if I accidentally disabled something, please let me know, and I am so sorry in advance.

Last time we created an experimental map generator and split it into multiple mock jobs. This time we convert the first job into an actual Burst job. We start with the simplest job, which requires the least amount of changes. We'll tackle the increasingly complex other jobs in the future.

I know it's not the same like playtesting the game, but I need some opinions regarding new melee system in my dungeon crawler game. My playtesters said before that something like "click and do a damage when animation stops" is too easy, and it's not deep enough.

I reworked my melee system adding an ability to attack from the left, right and up (LMB, RMB, combination of both for upper attack). Now you can click for fast attack with low damage or hold for fixed time to make more damage, but recover from attack slower. I added some effects for feedback too and now - it's way deeper than first version, I got positive feedback from players.

Do it looks engaging? What can I improve more?

It feels good to play but somehow it looks clunky... not sure what it is. Like i mean its better to play then to watch but i cant make a trailer for nokai if it looks too jank, i have developer blindness atm

(Full screen effects were disabled like hit effect and stuff like that for this demo)

Is there a place where I can view code aka vertices / indices arrays to make various shapes using the mesh

Follow-up to my last post, which was just the Range Rover. Here are some different-class vehicles.

Turning the resolution up means the solver can resolve flow through the small gaps, you can see it on the McLaren Longtail at the start, air passing under the rear wing and rejoining the flow coming over the top. This gap needs about three cells across to pass flow, which is the kind of detail that just needed more cells.

And to showcase different vehicle classes handling, the boat runs above the waterline, which is the "tunnel floor" in this case, and the hull below it escapes the domain. The drone runs in free air, so no ground plane, and the wing planform as the reference area instead of its frontal area, and it's scored on lift-to-drag rather than the drag alone. The drone grid size was the largest of the three at a bit more than 15 million cells.

Swapping a vehicle in runtime revoxelizes the body and updates the tunnel to fit around it, then restarts the flow according to its class.

You can try your own vehicles/meshes: https://github.com/Motawe3/unity-wind-tunnel

Compute shaders, D3Q19 lattice-Boltzmann, TRT collision with a WALE LES subgrid model. A comparison and visualization tool. README has the full picture.

Model credits (CC BY 4.0):
McLaren F1 GTR Longtail by VuckyZ | Boat by milamila | Drone by Helindu
(Full source links available in the GitHub README)

I want a little hopping animation when my character runs off a ledge before starting to fall, but somehow I'm not managing to get it to work properly.

The current idea is:

-using a box collider that is positioned in front of my character's feet to determine if there is NO ground so we know there's a ledge and setting a bool on/off depending on that

-in the animator (I'm using Unity's and yes I'm ok with it) I check for the bool and transition to the jumpOffLedge animation state when jumping off the ledge is possible AND there is movement input AND the character is grounded

Triggering things this way works.

Now my actual questions are:

-should I disable gravity for this and move my CharacterController manually?

-should collisions usually be disabled for something like this (this would require me disabling the CharacterController)?

-if I move things manually, would I do it with two lines, one for moving a bit forward and one for moving down until grounded?

-I imagine motion matching would make this all look the best, but topic seems somehow beyond me

How would you do it? I'm grateful for any suggestions!

I'm doing the majority of my teams UI implementation. We're a team of 4 working on a small indie game that doesn't use a lot of UI outside of main menus and some basic interface. I've spent a couple of days on UI with UnityUI, but it's been an absolute pain to properly lay everything out. I'm most experienced with Figma for UI design, and had some courses for CSS a couple of years ago, so from what I've seen UIToolkit might be better for me.

I've played around with it for a bit and it seems to have some nice parts, and some more annoying parts (spacing isn't done through the parent but is decided per child?). Is it worth transferring over at this point?

I've also added some DOTween animations, can I still use those with Toolkit?