Are sprites which player never see a good idea?

[goosetron]

Location in my game looks like flat cube as floor and long+high cube as wall and I need to add materials to them to have specific sprites on them. The issue is shown on next 2 screenshots - while player will see location as I want, sprite images are shown on all sides of cube, even which are never visible by player.

So my question is what is ‘bestpractice’ in such cases? Is there any way to make unity draw sprite on one cube side only and will it significantly improve the perfomance? Or its ok to have sprites on all 6 sides of cube even if only one is visible?

[Kurt-Dekker]

The back side of geometry is generally not rendered due to backface culling

Let me assist you in nailing down your terminology: nothing above looks like a Sprite… it just looks like Textures on Materials being Rendered on a Mesh…

Terminology primer for assets:

Convert sliced sprites into materials for quads

Generally nothing above can be “converted” between each other. They are all distinct things. To put an area of a Texture2D on a quad, generally you need to manipulate the UVs of the quad. Otherwise, just use a Sprite and render it with a SpriteRenderer Just to help you keep your terminology straight: a Texture2D is an image (think JPG or PNG) a Sprite is a set of coordinates (box or otherwise) outlining something interesting on a Texture2D SpriteRenderers render Sprites UI Image components …

Finally, if you intend to sprinkle actual sprites all over this surface, which is a great idea, keep this in mind:

Three (3) primary ways that Unity draws / stacks / sorts / layers / overlays stuff:

orthographic camera rendering order?

Unity draws things a few different ways, regardless of ortho vs perspective. The default 3D Renderers draw stuff according to Z depth - distance from camera. Sprites draw according to Sorting Layer and Sorting Depth (properties of the Sprite Renderer) UI Canvas Renderers draw in linear sequence, as if you were pasting papers atop each other When you mix and match you can get these issues. There are workarounds for all combinations. Google up mixing sprites with 3D objects, and mixing 3D obj…

In short, as far as the Standard Rendering Pipeline,

1. The default 3D Renderers draw stuff according to Z depth - distance from camera.

2. SpriteRenderers draw according to their Sorting Layer and Sorting Depth properties

3. UI Canvas Renderers draw in linear transform sequence, like a stack of papers

If you find that you need to mix and match items using these different ways of rendering, and have them appear in ways they are not initially designed for, you need to:

• identify what you are using
• search online for the combination of things you are doing and how to to achieve what you want.

There may be more than one solution to try.

For instance, you may even use multiple co-located cameras (along with different Layers and LayerMasks drawn to different-depth Cameras) to more-explicitly control apparent draw ordering.

Additional reading in the official docs:

Unity - Manual: 2D Sorting

And SortingGroups can also be extremely helpful in certain circumstances:

Unity - Manual: Sorting Groups

Other rendering pipelines may have other ways of layering (HDRP and URP). Go see the latest docs for details.

There’s lots of third party open source packages available as well, such as this:

GitHub - mob-sakai/ParticleEffectForUGUI: Render particle effect in UnityUI(uGUI). Maskable,...

[pertholdth]

the best practice is to dont use a cube, use a plane/quad, and then you can make it so that it only displays the texture on one side, like a paper page but one side is invisible

its indeed worse for performance to have cubes, you might say that its 5x worse, because you have 5 faces of each cube that are useless but are being rendered, not only the bottom but the sides aswell

[PhiMue]

@pertholdth

the best practice is to dont use a cube, use a plane/quad, and then you can make it so that it only displays the texture on one side, like a paper page but one side is invisible

its indeed worse for performance to have cubes, you might say that its 5x worse, because you have 5 faces of each cube that are useless but are being rendered, not only the bottom but the sides aswell

In theory, yes, but in practice there really isn’t any difference because the faces are all part of the same draw call and it doesn’t matter if there is one, or six, or six hundred faces in a draw call. For the GPU it’s all the same. By the way, the default “Plane” object has more faces (triangles) than the cube, 200 to be precise. But as I said, that really doesn’t matter. Especially with backface culling.