with Wave Lab of Steinberg you can "normalize" - that's the word you are looking for - the amplitude (volume) of the selection or the whole song.
You can use the Loudness Normalizer to achieve a specific loudness.
Increasing the loudness to a specific value can provoke clipping. To remedy this, a peak limiter (Peak Master plug-in) can be part of the process. The Loudness Normalizer raises the loudness and limits peaks in the signal at the same time if needed, to achieve the wanted loudness.
This process happens in several stages, first an analysis and then the final rendering.
Mind: normalizing can have a range of meanings. My approach comes of statistic science and not of music technology.
*norms are considered as norms concerning
a) ethical values, (ideal norm),
b) functional norms and
c) statistical norms.*
So level normalisation could be a functional norm: this would hinder a damage of your ear (most brass band and pop concerts are playing much too loud. normalisation would be helpful for the people preventing to become deaf!
Loudness noramlisation would mean a statistical sence to normalise extrem values of piano and forte: lower the forte tones and raise the piano tones to the average. But why then write or play music with various loudness?
The ethical norm in this context is concerning my comment to the OP's question:
I don't know why I tried to help to give answer to this question!
1. Why listen to music when reading a book? (unless you want to practice superlearning and in this case you should listen to quiet and not exciting music, as e.g. the Air of the Suite in G of Bach or the Adagio of Albinoni.)
2. If someone want's to listen at music for paying attention to a composition he should stop talking, stop reading. He would be paying some respect to the composer and the performers to concentrate to this work. To this aspect I can tell you how annoying it is to me when people are asking me to turn the music "a little bit louder" or a little bit lower" because there is just a short piano or forte passage. (If we think how many "takes" it needed and the conductor wanted until he was satisfied by the level of loudness: "the trumpets have to play softer at bar 299", "the trombones were too loud", the string section has to be stronger in section B" ... etc. and you want to "normalize" or "compress" this recording!!!???
3. If you really want/need to listen to music while reading a book or the newspaper or also if you have friends coming to see you - you just have to decide right in advance what song or CD you want to listen at - before you start reading. So you don't have to programm your equalizer or buy a software for 500$ or more.
And if you find the music is too loud or too soft, you can change the CD or turn off the radio!
This is what I would understand of "normalization" - in an ethical or cultural way.