Some tracks sound best at a certain volume for the simple reason they were mastered to be that way. This is due to our hearing not being linear but corresponding to equal loudness contours mapped by Fletcher and Munson. We do not hear all frequencies equally well at all volume levels. Low bass and high treble in particular need to be at a fairly high level to be heard at all. This is why the Loudness control used to be so common. All audiophiles should know this but whether they do or not all recording engineers certainly do.
Because of this, in mastering the engineer has some decisions to make. A frequency response that sounds flat at one volume won't sound right at a lot higher or lower volume. So they master for what they consider the most likely level of the more serious listeners. Obviously then playing it back will sound best, or at least have the intended balance, at that same level.
This same feature by the way is why we can't rely on meters to EQ properly. Meters do not "hear" equal loudness contours. Meters measure pressure waves. What we want is good sound quality at whatever level we like to listen at. This may or may not be what the mastering engineers thought. May or may not be what the EQ police think. What we like is what we like.
Gosh now that I think about it I guess an awful lot goes into how I listen.
