It is possible, I suppose though it seems unlikely. An ordinary, running, car engine is roughly the same loudness as human conversation. So we are talking about the ability to hear someone talking at 20 miles away.
I judge it on the edge of plausibility that when measured at the ideal frequency spectrum for a dog, under conditions of complete acoustic shielding a difference of this magnitude at the maximum sensitivity level for dogs might just be discernible by a dog (probably with a high error rate).
The ability to do reliable recognition of such a sound at that level in a natural, noisy environment, with many almost identical sounds seems very implausible to me. It might be worthwhile, though, for someone who is a little less rusty on acoustic physics run the numbers and see what this claim actually amounts to. |