> Other than stars etc., I don’t think know any genuinely omnidirectional light sources. Even other substellar objects emit non-isotropically because they aren’t spherically symmetric. And at most frequencies, light doesn’t diffuse much.

In what cases does it matter that stars are not symmetrical enough to emanate light omnidirectionally? Stars are pretty close to isotropic.

That’s my point. They — stars and some stellar remants — are basically the only thing which are isotropic. Everything else is directed, has some gain over isotropic in the peak direction/s.