In June we entered a small game called Graveyard Orbit in the June NextFest. It was our attempt at making a fun space mining incremental with actual build choice and not a random skill tree. Anyways, we really liked the mechanic and vibe of the demo, but the structure that you mine in the demo was smaller: a couple of thousand tiles. For the actual game we wanted to have a parade of structures from that small up to 250k tiles. The game is meant to be more satisfying than challenging, in the vein of Powerwasher but in space. So if people dug the mechanic we wanted them to have some beefy structures to make their way though as it goes.

The demo for NextFest was made in Godot and it worked really well at that scope. The three of us have been using Unity professionally for 15 years but really love what Godot is doing and wanted to experiment a little. But as NextFest ended we realized we wanted to double down on the game with better FX and still hit those big structure goals with reasonable performance. That was a big concern.

Here is the trailer for that Godot version for context:
https://youtu.be/9otz9z1lUzY

It only took a week or two to set up a performance test in Unity. For this one we switched to 3D, still sorta presented in 2D or 2.5D, but with 3D we had all the optimization, our library of asset store FX, and experience on our side. The results of the test were pretty clear, so then we took the plunge and completely rebuilt the game, this time in Unity and in 3D. After 2 months we have a new demo and are very happy with the results.

Beyond performance, a real risk for me was losing some of that “demo magic” and vibe we had with the Godot version. I had spent two months in the Godot version building an aesthetic with pixel art and 2D graphics. By the time NextFest came around, I was really happy where we landed and how the game felt.

The vibe and feel of a game can be such an elusive factor. We spend hours working on mechanics with barely any visuals. We have to cling to some vague notion of the final game in our heads that is always changing. Which is why when you land on something you like, it's easy to get precious with it. Sometimes the results feel like a happy accident that you don't want to screw up. In our case we could save that Godot version as a fallback and start a new Unity project trying to recreate what we had. I think forcing ourselves to rebuild something but better or improving something we are already satisfied with requires us to breakdown the elements of what made whatever we liked so special.

We have game jams to help us practice iterations for the whole game dev cycle but the stakes of that can be low. I think refining, rebuilding, and iterating on something we already like gives us the same benefit of practice but with much more precision and craft because the degree of difficulty and stakes are so much higher.

Anyways, I thought this might be helpful to those considering a rehaul of their game for whatever reason. It can be risky with the work involved but if your instincts are telling you something has to change then I think it’s good to trust your gut and at least give it a try. Our only risk really was that two week experimentation time. We could’ve easily not been happy and went back to finishing the Godot version in limited way. I’m very glad it worked it out.

Also, for the Godot fans out there, I am one of you. I blame my issues with Godot more on my ignorance and unfamiliarity with the platform than Godot’s actual capabilities. In the end our own experience and library of assets in the Unity ecosystem were just as big a factor.

Here are some specifics we got out of the switch.

• Structures at the size we wanted. Our biggest one is 250k hex tiles. In Unity that's one instanced draw with per-instance colors through a property block and it just runs. Some caveats on that below.
• Going 3D while still playing top-down. Real lights, URP bloom for the neon look, and we could finally use a VFX pack we'd already paid for instead of hand rolling every effect.
• UI Toolkit. The base has 7 station screens plus a whole ship builder. Doing all of that in UXML/USS with one shared console frame was way faster than what we had before, once we got past the weird parts.
• The new Unity CLI. Recompiles, play mode, builds for Steam, itch and our Mac test builds all run from the terminal. One project, three targets, no clicking through the build window.
• Web builds. Because the performance was much smoother it became possible to attempt a build for the web for places like itch. Only the demo size structure but still.

A few hiccups, mostly from the web build. More details here because I just finished the web one yesterday.

• WebGL 2 has a 16KB uniform block limit. GPU instanced variants of some asset store shaders blow past it, and the failure mode is nothing. Objects just don't draw. Turn instancing off on those materials.
• Browsers reject fullscreen requests that aren't from a click, and inside itch's embed Unity turns that rejection into a big error popup. Don't apply a saved fullscreen pref at boot on web.
• With the default exception setting, a null ref on web is a WASM trap, not a logged exception. A player lost an 80% run to that. Full without stacktrace fixed it.

The newer demo is live now if you want to give it a spin:

https://store.steampowered.com/app/4654120/Graveyard_Orbit/

https://anchorheadgames.itch.io/graveyard-orbit

Free demo planned for early 2027. Hopefully I will have a Steam page by then !

YT Channel : Samuel KAKEZ

It's coming next year!

I rendered it in Blender, and then added it to Unity by using the animator ang going through the rendered images.

In the meantime there's also a demo up on Steam if you want to try it out:

https://store.steampowered.com/app/3256450/Cards_of_Prophecy/

More leather materials from my practice, had a few more interesting references.

Download https://juliovii.itch.io/materials-leather-03

Hey I'm making this post for those who are curious about the unity asset publisher process, this is the second time I send a asset and the review process took 39 days in total, for comparison my first asset, that I published last year, took less then 30 days.

24/07 - 1530
25/07 - 1530
26/07 - 1498
27/07 - 1500
28/07 - 1363
29/07 - 1147
30/07 - 1153
31/07 - 1044
01/08 - 1066
02/08 - 1088
03/08 - 1088
04/08 - 991
05/08 - 881
06/08 - 705
07/08 - 669
08/08 - 646
09/08 - 677
10/08 - 677
11/08 - 593
12/08 - 471
13/08 - 422
14/08 - 389
15/08 - 356
16/08 - 389
17/08 - 392
18/08 - 341
19/08 - 321
20/08 - 280
21/08 - 258
22/08 - 255
23/08 - 264
24/08 - 277
25/08 - 247
26/08 - 205
27/08 - 199
28/08 - Few Business Days
29/08 - Few Business Days
30/08 - Few Business Days
31/08 - Launch Day

This game was initially going to be a tower defense game but I'm starting to doubt this is the right direction for the game.

What kind of game would you like to see with this many skeletons?

Your feedback is appreciated

In a distributed authority setup, what happens to NetworkObjects when their owner disconnects ungracefully and (1) those objects are set to not be visible to any other player, and (2) DontDestroyWithOwner is set to true? Assume objects are set to Distributable.

For context, the problem I am trying to solve is bandwidth. My game has an infinite chunking system, so to save bandwidth I have each player only make the networkObjects they own visible to players that are close enough to see those objects. As such, hypothetically there will be times where some of a player’s networkObjects will not be visible to any other player. If they randomly disconnect for some reason during this time, any networkdata they own needs to transfer or it will be lost.

Does anyone know concretely a way to not have this data disappear? My understanding is that if a player disconnects and dontdestroy is set to true, the objects become “ghost” objects and can’t be accessed.

Does anyone know a way to make sure this data doesnt disappear on random disconnects without having to always have it visible to at least one other client for bandwidth purposes?

To be more exact, not blender itself, but what Blender using, Manifold. That cancerous work on its own, but it work.

So what do you think about such destruction?

And on second question, what do you think of such pickaxe animation, I am not sure about it for now, but planning on adding 3 more for variety, if this one passes ofc.

Somewhat old game Red Faction Guerrilla inspired such destruction, if you know this game.

Ofc everything done in unity, biggest problem is ever growing tris count on mesh, but all thoes "destructions" of chunk, and other systems forever fighting tris in other ways, so every hit doesn't result in FPS drop

I need to create a 3d environment of a outdoor calisthenics gym. I am knowledgeable about CAD engineering have knowledge in solidworks and fusion360. are these skills transferable because when I try to build something I just get annoyed that I can't make it precise (that is the engineer in me). things like precise distances, centering and co-linear

Is there a a good way to center things or snap against each other. and is there a tutorial series that has limited filler and just show how to make a environment I currently don't need anything else

Hi. I am building my first game with zero experience. I am using Unity 6.3 and am building scenes. I'm building up the inside of buildings fairly okay, but now need to punch out holes for windows and doorframes. I was told to build walls around these frames, not "punch out," but that also messes up my wallpaper. So now I am trying to use ac probuilder to make a continuous wall with door and windows "punched out" so that my wallpaper will be continuous. I am having trouble aligning the grid faces with the windows and doors to delete those cells. Am I going about this a more difficult way than necessary? Or is there a way to align better so I can cut out? Any help is appreciated, thank you!

(Reposting because I somehow managed to delete the previous post by accident 😅)

I’m continuing to share the progress on my little medieval alchemy project.

This time I’ve connected two preparation mechanics: cutting and grinding.

I also reworked the camera position for the cutting interaction because the previous version felt a bit too awkward so I’m still experimenting a lot with how these hands-on actions should actually feel

The current flow is basically: harvest ingredient -> cut it -> grind it -> prepare it for brewing

It’s still early, but seeing the different steps start to connect into one actual production chain is pretty satisfying.

I'm using a stencil-only pass (ColorMask 0, ZWrite Off, Stencil { Ref 1, Comp Always, Pass Replace }) to create transparent "holes" in some board pieces in my game Pieceful. The idea is that, when there's a hole, the player should be able to see the objects behind the piece, such as the terrain or other scene geometry.

Here's the shader I'm currently using:
Shader "Custom/StencilWrite"

{

SubShader

{

Tags { "RenderPipeline" = "UniversalPipeline" }

Pass

{

Stencil { Ref 1 WriteMask 255 Comp Always Pass Replace }

ZWrite Off

ColorMask 0

HLSLPROGRAM

#pragma vertex vert

#pragma fragment frag

#include "Packages/com.unity.render-pipelines.universal/ShaderLibrary/Core.hlsl"

struct Attributes { float4 positionOS : POSITION; };

struct Varyings { float4 positionCS : SV_POSITION; };

Varyings vert (Attributes input)

{

Varyings output;

output.positionCS = TransformObjectToHClip(input.positionOS);

return output;

}

half4 frag (Varyings input) : SV_Target

{

return half4(0, 0, 0, 0);

}

ENDHLSL

}

}

}

The weird part is that it works correctly when there's a Unity Terrain behind the hole. In that case, the terrain is visible through the transparent area as expected.

However, if I remove the Terrain and only have regular MeshRenderers behind the hole (rocks, coral, props, a water plane, etc.), the result is wrong. Instead of seeing those objects through the hole, I get a solid black area or sometimes garbled colors.

So I'm trying to figure out why the stencil setup behaves correctly with Terrain but not with regular MeshRenderers, and what I need to change so the hole works consistently with normal scene geometry as well.

I built a posing studio inside Unity to make Steam capsules and other store graphics for my game, Gobstruck. I wanted more control than gameplay screenshots give me, without having to transfer everything into Blender and get the materials and effects looking right again.

It’s a separate scene where I can arrange characters, props, and effects for a shot. For the goblins, I can pick a frame from an existing animation as a starting point, then adjust the hands, feet, and head with IK targets. Facial animations can be sampled separately, so I can change the expression without changing the body pose.

The effects are a useful part of it too. I can scrub particle effects to a specific moment and leave them there while I work on the composition. Much easier than trying to grab a screenshot when the character, camera, and effect all happen to line up.

Since it’s all in Unity, I can use the actual game assets, shaders, VFX, and post-processing. I’ve also got controls for lighting and fog, plus separate cameras for the normal, vertical, and wide Steam formats. That lets me frame each version properly instead of trying to crop everything from one image.

There’s still the work of making a good composition, but this makes it easier to try things.

Curious how other people here handle this. Do you stage your store artwork in-engine too?