The plain spectrum tells you that for any one sample of the song, the following frequencies were present. It doesn't tell you exactly when.
This is the spectrum of me pressing the "G" key on my piano. It represents all the energy during the entire key press.
http://kozco.com/tech/audacity/piano_G1.jpg
This display does tell us loudness up and down and pitch side to side accurately. You would want the logarithmic frequency scale to make up the difference between the mathematics and our ears. Half-way up the frequency range is obviously about 10,000 Hz, but any human will tell you a middle tone is about 400Hz. A 440 is the oboe tone at the beginning of the orchestra. 400Hz is one of the two common broadcast test tones. The arithmetic is linear and the musicians log. Same music.
The ear works the same way with loudness. Half audio level is clearly -6dB on any voltmeter, but people hear half volume at about -18dB.
I'm never used that display, but it appears to tell us frequency distribution up and down against the performance time side to side There are no other ways to display data, so the loudness is relegated to colors.
I would be using this display to tell me when in the orchestra performance the area of interest is, and then use the spectrum -- log to tell me more accurate information about that note or phrase.
Koz