Sean, Flavio,
It may be a very moot point wrt light as well, as you're hopefully finding in my essay.
As the source of sound is NOT also the detector the only way to get a relative speed is by calculation, of c over the flight distance.
Of course it's exactly the same with light when calculated with that same data. That's where the paradox is, because that gives c + v, yet the detector at the end of that flight path, but in the same frame as the emitter, measures 'frequency' (a derivative of effective wavelength on detection) and finds the speed at c/n wrt his lens, not c + v!!
I should say that's where the paradox 'was'. I believe it has now gone using DFM dynamic logic. Because all detectors are made of matter, and all matter is a medium, then all detectors MUST find c/n, because light MUST change speed on arrival to the new local c comply with the constant refractive index n and SR in all frames. (It works the same in a vacuum with scattering off diffuse plasma).
That may need you to lie down with your eyes shut and think through a few dozen times. I've failed abysmally to falsify the model, and found all the apparent issues with it evaporate like the SR paradoxes. Except the human factor of course, as few can overcome it's unfamiliarity. Working at Perimeter may just allow you to not reject it before full testing (I have a stack of test data).
I look forward to your comments or questions (and maybe the look on Lee's face when it's explained to him).
Thanks, and best wishes
Peter