The subjective perception of volume, with regard to sound. It is well known that the human ear's perception of a sound's loudness depends on both its frequency and volume; at low volumes, the ear is most sensitive in the 1000-3000 Hz region, somewhat less sensitive at higher frequencies, and considerably less sensitive at lower frequencies. At higher volumes, the curve "flattens out" and the variations in loudness perception with frequency are less severe. (This is one reason that listening to music at high volume is often more satisfying than listening to the same music at a lower volume.) Many consumer audio amplifiers have a "loudness compensation" circuit that applies equalization, based on the setting of the volume knob, to try to compensate for this effect.
In electronic music, patch programmers will sometimes, when building a patch intended to be played in both the bass and treble ranges, build loudness compensation into the patch, without being conscious of doing so. For instance, a common paradigm when using subtractive methods is to have a VCF that tracks with the note being played in order to shape the timbre, but the patch programmer, intentionally or not, adjusts the VCF tracking so that the filter is more "open" when lower notes are played. This makes the lower notes come out of the synth at a higher volume, and a timbre including more higher frequencies, which creates the perception that the bass and treble notes are equal in loudness. It can turn into a problem when the patch programmer is, in the process of creating the patch, listening at a lower or higher volume than what is used in mixing the track, or by the listener ultimately.