I understand that normally when two waves combine destructively this happens only in certain places; the energy doesn't disappear but is redistributed to other places where the interference is constructive instead. In a sense, unless a noise cancelling speaker has the exact same origin as a noise creating speaker, there will always be an interference pattern with darker/quieter balanced by brighter/louder areas.
But what if instead of trying to have two speakers or a 0º phase laser and a 180º phase laser exist in the same place, we instead combined their two beams somewhere else:
signal A -+|+-+|+- ==> \/ <== |+-+|+-+| signal B, 180º out of phase
||
||
combined
Where would the energy go, if one could perfectly focus and combine two sound or RF/light sources in an interfering fashion?
I am curious about this for both the acoustic and electromagnetic cases, but if there is no analogy then it's the photons I'm most interested in! I found some discussion of the latter in https://physics.stackexchange.com/a/601337/103805 (and perhaps as well in https://en.wikipedia.org/w/index.php?title=Beam_splitter&oldid=1076656730#Phase_shift) but I don't understand either well enough to answer my own question. One answer said:
If you really want to consider two waves with identical wave vectors, then you cannot emit a second photon that is 180∘ out of phase with the first. Generating an electric field with that phase is actually absorbing the first photon, not emitting another one.
An ideal "beam combiner" would seem to get around this idea of emitting two photons of opposite phase right at the same place; is it simply physically impossible to make such a thing?
But if perhaps one can only come close — perhaps the incoming "beams" aren't perfectly collimated or the combiner isn't perfectly balanced or whatnot — so the interference isn't 100% destructive in practice. But of course even if only 80% of the energy, or for that matter 1e-8% of the energy simply "disappeared" that would still be amazing, no?! So what prevents an arrangement like the beam combiner above from ever working even partially?