Unlike waveform scoping, the spectral one is idiosyncratic. Deciphering or fine-tuning the spectral visualizations can be subtler and more involved as to be explained here.
When scoping the spectra, we get a time-varying representation of them as fluctuations of energy per bin. The horizontal axis represents the continuum of the frequency range (typically from DC to the Nyquist frequency), divided in discrete frequency ranges (that is, the bins). The vertical axis stands for the magnitude of energy of the bins. Accordingly, each point in the graph represents the energy of a particular frequency range (and not that of singleton partials). The graph is constantly updated with respect to the FFT temporal window; once the FFT algorithm has analyzed the snapshot of our signal, the graph is updated to represent it. As already explained, the more time-accurate a spectral scope is, the less number of bins it can accurately represent.
As discussed...