Very interesting video!
Coming from the physico-chemical science community I had a similar phenomenon in one of my lectures about Fourier transformation.
If you mix two waves, where one wave has the double frequency of the other (i.e. one octave higher), you have to take into account their relative phase because you might get exactly what this guy is talking about:
Here we have two waves with frequency 2 and 4 (arbitrary units). You can see the individual oscillations as grey lines. They start both at a maximum, i.e. their phase difference is 0. As you can see the resulting sum (blue line) is peaking at +2 amplitude (again arbitrary units) but only at -1 in the negative direction.
Here is a second example. Just as above, but now the oscillations are phase shifted by pi/2.
As you can see, the maxima of the individual oscillations are now at different positions and the resulting wave is symmetric around zero peaking at about +1.71 and -1.71.
Looking at the frequency spectrum (right side) in both examples position and peak height are identical, but the peaks in the second spectrum look a bit more symmetric.