Zed Scaling - Upcoming Feature Revealed by Ashley (video)

Video released by Ashley earlier today (Dec 19, 2018): youtube.com/watch

So, with this new update about an upcoming feature, I'm beyond thrilled. I'm like a kid meeting the REAL Santa.

The feature, as shown in the video, focuses on literal 3D zed-axis scaling/zooming/parallax (not sure which is the most accurate term there).

I've actually worked with this kind of concept by taking layers and scaling them at different intervals to give a sense of depth when working on a top-down perspective in my games. The effect was very nice, albeit, limited due to everything being FLAT. (This IS a 2D engine, after all.)

But Ashley has revealed that this zed-axis feature won't be limited to just layers, but will apply to individual object instances, too!

This brings a few questions to mind, which I will TRY to articulate clearly and succinctly...

1. Since this is zooming (I'm just gonna stick with that term) on the zed-axis, the objects are not actually getting bigger or smaller...they're literally just closer to or farther from the camera. With this understanding, it is presumed that objects that overlap "zoomed" objects will still factor collisions and such. At the same time, since this is dealing with real 3D space (per Ashley in the video), will there be conditions and actions that take into account objects that are "zoomed" versus objects on the same zed value?

2. Although this is still explicitly 2D we're working with, the addition of zed/depth will add new quirks, no doubt. Similar to how the shadow behavior has a "height" property to determine how far a shadow will be cast, will objects (namely sprites, but hopefully most the rest, as well (like background tiles, 9patch, etc)) be given a "depth" property to take into account the zed nature being introduced? Again, this is keeping in mind we're dealing with strict 2D in all other matters, but with zed now factoring in, it might be interesting to ponder when things like solid behavior comes into play... An object that is, say, a height of 3 pixels can pass under/behind a 5-pixel object so long as those 3 pixels are not "overlapping" at the same 1-3 zed pixels. Anyway...you get the idea, I'm sure (I hope).

3. Despite the repeated strong emphasis on this NOT being a move into 3D engine territory, how will movement behaviors take to the new zed-axis? (SHORT PARAGRAPH! Who knew I was capable of that?! HA!)

Thank you, Ashley for this awesome new reveal and your work and the rest of the Scirra Bros team!

1. The idea is all the gameplay is 2D, but some of the graphics can be Z elevated. So collisions will still work as if everything is on Z=0.

2. There's no concept of depth, and I don't think it makes sense to add it for 2D objects with Z elevation - they essentially have no thickness.

3. Due to the first answer (all gameplay is 2D), we don't have any plans to update behaviors to use Z at this time.

This looks like a pretty cool feature just because it means you can run all your different parallax scrolling layers off a single layer.

The zoom stuff is visually very nice too.

Awesome! Thanks for the clarification, Ashley! I can't wait to make use of this!

Hey, really awesome, ive been doing this in events since the beginning of my game

eg old video not really representative of my current game drive.google.com/open

but it requires a "for each" for everything on the screen to be scaled and then z sorted every tick

which takes like 40% cpu on i5 and pretty much rules out mobile

im assuming i should be able to transfer the z coordinates in my game into this pretty easily and, Ashley maybe get some significant performance increase?

Glad to see this long missed feature return. Thanks!

No need anymore for putting paster objects on different scaled layers and then paste objects accordingly ;)

This would also help my raycaster engine to do the depth sorting/scaling.

Which begs the question, do you plan an option to separate the scaling from positioning?

So e.g toggling:

Z-Height options:

(X) Scale object

(X) Position object

I know the applications are somewhat limited, but I imagine there might be scenarios where you want to have it separated, like a starfield simulation where you want the stars always be 1x1 pixels etc.

Cheers!