But technically this is a "voxel editor" - because originally "back in the day" - voxel engines (those few that existed) rendered everything with pixels.

With a bit of googling, you can find some of these engines - here's a start - Ken Silverman's VoxLap (if you don't know who he is - then I haz a sad):

http://advsys.net/ken/voxlap.htm

Note that these are not "small 3D cubes" - they are literally just pixels. It was a technique that was explored in the 1990s during the whole DOOM era, but machines weren't able to render fast enough, and 3D cards just kept getting better - and here we are today.

Again - you can find more than a few of these examples from the time, but the tech of the day just wasn't fast enough.

Editors for voxel objects didn't exist - but they didn't really need to, because regular 3D editors could create the objects.

Which makes me wonder how well those old engines would work today, if this "sprite stacking" thing works well. As I recall back then, the demos I recall seeing weren't terrible, they just weren't up to the performance levels needed for a game engine...

EDIT: Some more detail and history:

https://killscreen.com/articles/meet-the-voxel-pixel-cousin-...

Isn't the stack then just the third dimension of the original two dimensions?

It is, e.g. I think tools like MagicaVoxel can construct a voxel object based on a stack of sprites, as you describe.

However, not every engine supports 3d models / voxels, so sprite stacking is commonly used in combination with engines that only support 2d, by simply drawing each layer on top of the previous one but one pixel higher on the y axis.