Actually though, I think a sphere is still in a way the right object. The weird donut is a representation of the amount that is emitted in different directions - I guess it's a contour of equal intensity. But the radiation is still emitted at the speed of light in every direction. So the wavefront at any given time is a sphere (until it is defracted or reflected).
(Edited to add) unless you're thinking of a really big dipole, so you can't treat it as a point source? Or close to it.
> But the radiation is still emitted at the speed of light in every direction.
I think that makes sense.
> (Edited to add) unless you're thinking of a really big dipole, so you can't treat it as a point source?
Honestly, didn't think of that issue either. You're right though, some antennas (especially AM antennas) are the size of Godzilla, and its absurd to even think of them as point-sources.
The term gain is just referring to the antennas directionality relative to a perfect spherical omni radiator that doesn't exist.
That explanation makes a lot of sense.
EDIT: I also know we aren't dealing with "spheres", but... I get the gist. Dipole antenna emits in this weird doughnut shape, but as a 3-dimensional object, we can estimate it to be kinda-sorta like a sphere :-)