Idea for sprite rendering with and without depth testing

[GordonTombola]

• zIndex of the Sprite3D should return and modify the internal projection sprite zIndex
• zIndex value dictates whether we want to render sprites before/after other scenery without depth test

Not 100% certain that this would be the best way of handling this, thought it'd be best to get your thoughts on the premise at least

Why?

• this means we can render out sprites without depth testing to get the 3D position but render them out as a form of screen space effect example draw behind: stars rendered behind everything else to give the impression of moving through space example draw in front: rain particles that may be constrained to the camera in 3d space and be rendered in front of everything else

[jnsmalm]

Hello! I like the idea of this but I'm a bit busy right now - I'll take a look later. Just some quick thoughts:

• The NO_DEPTH_ZINDEX seems a bit random and may be difficult to understand from a API perspective.
• Maybe it would just be simpler for a Sprite3D to have a property "depthTest", and all of those sprites gets rendered last after the meshes and other sprites?

[GordonTombola]

Hello! I like the idea of this but I'm a bit busy right now - I'll take a look later. Just some quick thoughts:

• The NO_DEPTH_ZINDEX seems a bit random and may be difficult to understand from a API perspective.
• Maybe it would just be simpler for a Sprite3D to have a property "depthTest", and all of those sprites gets rendered last after the meshes and other sprites?

I agree, the zindex determining depthtest is a bit arbitrary, there would still need to be some combination of negative zIndex being used as a way to draw before meshes

[jnsmalm]

In what circumstances do you think this level of control is needed? Could you give an example?

[jnsmalm]

@GordonTombola While thinking about this I got another idea which can be used for all objects, both mesh and sprite. What if mesh/sprite had an "layer" integer property. All objects with same layer is rendered, if another object exists with a different layer then it will clear the depth buffer between those layers. So everything with a higher layer number will always be on top of stuff with a lower layer property. With this solution there is basically no restrictions in which order you want to render objects.

What do you think?

[GordonTombola]

@GordonTombola While thinking about this I got another idea which can be used for all objects, both mesh and sprite. What if mesh/sprite had an "layer" integer property. All objects with same layer is rendered, if another object exists with a different layer then it will clear the depth buffer between those layers. So everything with a higher layer number will always be on top of stuff with a lower layer property. With this solution there is basically no restrictions in which order you want to render objects.

What do you think?

Sorry I've been busy past few days, that would make more sense, keeping a common property between the two types

[GordonTombola]

Closing this, will investigate the possibility of an ILayerObject interface that could be implemented to sprite3D and Container3D which could determine the depth buffer clearing