As a musician and electrical engineer I like to think about the Fourier transform as a piano.
The signal is me humming a C4. We see the overtones C5, G5 and C6 in the plot.
The same signal (me humming a C4), but the horizontal axis has been scaled logarithmically. Consequentially, the octaves are now equally spaced, much like on a piano!
Now we can easily project the power spectrum on a piano backdrop.
Analysis of the Matlab built-in "train whistles" sound byte. We see that it is composed of an inverted D minor chord
Using this visual support, it becomes pretty easy to transcribe the cluster chords:
Audio from Gbmaj7#9#11no3.