Distribution of power in music

Back when I owned my own studio, I acquired a real-time spectral analyzer (referred to by Academic Power Dudes and Dudettes as an RTA) which I set up on the meter bridge of my console as one of my primary monitoring displays. Over about a four year period I got an extended exposure to the physical display of the audio spectrum as I recorded, overdubbed and mixed multitrack and stereo recordings of all sorts of music. One of the primary things I noticed was that the bulk of the power resides in the bottom three octaves, except when the rhythm section is not playing. In a typical pop-rock song, the electric bass and kick drum combine to generate the power level that pretty much defines the overall upper limit of the final mix level. During intros and verses, high and mid frequency levels are generally well below the bass levels, except for the loudest moments of lead vocals, which will generally come up to a level approximately equal to the bass levels. During loud, climactic sections, particularly with lead guitar lines (which are usually both sustained and loaded with distortion-based high-frequency harmonic content), the overall spectrum becomes much more “flat.”

	+enlarge
	Figure 10. Typical power spectrum for a pop-rock recording during the first verse. Highly variable and rapidly changing over time.

	+enlarge
	Figure 11. Typical power spectrum for pop-rock recording during the loudest chorus, including harmonic content from guitar distortion and heavy cymbal action, and vocals rising above this in Octave 8.

There are several generalizations to make about these examples. First is that bass takes the bulk of the power. With the kick drum, it will determine and limit the overall level of the tune. I’ve found that if you get them together showing a summed level of about -5 VU on the mixdown meters, you usually come out about right, with the loudest material just about hitting +3 VU and the general level of the mix (this is for pop-rock, remember) cruising between 0 and -3 VU on the mixdown meters.. The second generalization is that vocals tend to dominate the upper octaves. Finally, the two outer octaves tend to have comparatively little power and there is reduced power in the middle octaves, relative to the bass and the vocals.

Drums and bass provide the motor energy for music, energy which is particularly critical for popular music. Because the kick drum and bass are so important, and require most of the power, the setting of their levels is critical. Vocals, meanwhile, are the most important part of the song, and simply must be audible. Their audibility and intelligibility are generally a function of power in Octaves 7 and 8, where the spectrum of consonants lie. The mid range is a fascinating region. If you boost energy in Octaves 4, 5 and 6, the mix will become thicker, fuller, richer, heavier, muddier, darker, and/or fatter (whatever these terms mean). If you cut energy in these octaves, the mix will become, conversely, thinner, more open, more transparent, lighter, and/or clearer. Which is better depends on your particular goals and production values in this particular case. The outer two octaves are not well reproduced by most speaker systems. The bottom octave, particularly, is a problem. Gary Katz, producer for Steely Dan, once told me he deliberately gets rid of content in the bottom octave simply so that he won’t have to be concerned about the variability it causes in playback. By the same token, it is a common practice for producers to boost the top octave so that at least a little sound in that octave is audible when played back over mediocre speakers (this is sometimes politically incorrectly called “ghetto mixing”). Such mixes are a little hard to listen to over flat or bright monitor systems.