You walk along the fence - relative to you, the posts have speed c and the frequency you measure is f=c/L, where L is the distance between the posts.
Now you start running along the fence - relative to you, the speed of the posts shifts from c to c'=c+v and the frequency you measure shifts from f=c/L to f'=c'/L=(c+v)/L.
Sane minds apply this model to all types of waves (posts become wavecrests and L becomes wavelength) and infer how the speed of the waves and the measurable frequency vary with the speed of the observer.
Bingo the Einsteiniano admits that the fence model is valid for waves other than light waves. When it comes to light waves, however, Bingo the Einsteiniano says different things: The speed of the wavecrests relative to the observer does not vary with the speed of the observer, there is no shift from c to c'=c+v, the frequency does shift from f=c/L to f'=(c+v)/L but the speed of the wavecrests does not shift from c to c'=c+v, never, help, help, no shift in speed, shift in frequency only, the speed of the wavecrests relative to the observer remains the same, absolutely the same, c'=c, not c'=c+v, Divine Einstein, yes we all believe in relativity, relativity, relativity.