I'm not sure if I'm happy with having read this book, i.e. if the read was time well spent or not. The book is far too long for its content, and a bit hit-and-miss. I couldn't really relate to many of the bands that Levitin was referring to (Sting, Eagles), except for some fun facts about the Beatles. Also, some of his personal anecdotes are really boring and didn't help explaining the topic he was discussing.
There's an annoying mind/brain dualism in this book. Despite discussing Wittgenstein in a part of the book, Levitin should study him more carefully next time and try to understand the point he was trying to make with his later philosophy. Levitin writes about pitch that it "is the end product of a chain of mental events that gives rise to an entirely subjective, internal mental representation or quality". I would say that this is wrong. First of all, pitch is partly the product of the frequency of the sound itself (the physical, out-there-in-the-world property). Second, pitch is not entirely subjective, since pitch is in a great extent part of a cultural heritage. He writes himself that "[f]or reasons that are largely cultural, we tend to associate major scales with happy or triumphant emotions, and minor scales with sad or defeated emotions." The pitches in a specific scale is largely a cultural product, but also depends on the intrinsic harmonious properties of the certain sounds themselves, such as overtones (how and how much is still a subject to debate). So the subjective, "mental" part of pitch is caught up in this inter-subjective reality, so to speak, even though there are idiosyncratic variations of pitch cognition between listeners.
Another example of Levitin's dualism is this: "Our brains can estimate the size of an enclosed space on the basis of the reverberation and echo present in the signal that hits our ears. Even though few of us understand the equations necessary to describe how one room differs from another, all of us can tell whether we're standing in a small, tiled bathroom, a medium-sized concert hall, or a large church with high ceilings." Of course, the computational process in the brain is a part of what makes it possible for us to experience different reverberation, but it is misleading to say that "we" don't learn about reverberational properties but our "brains" do. Certainly, after being exposed to different reverberation all your life, you learn how to distinguish between them; you know how your voice sounds when you're talking inside a church. When you experience a certain reverberation and understand it, then the reverberation that you are conscious of and the computational functions in your brain that make this experience possible, are not two different things but two sides of the same coin.
Levitin's attempt to explain why music moves us is slightly interesting, but, again, he confuses us with his dualism. "As the music unfolds, the brain constantly updates its estimates of when new beats will occur, and takes satisfaction in matching a mental beat with a real-in-the-world one, and takes delight when a skillful musician violates that expectation in an interesting way". What on earth do we make of this? Because our brains like it? This just creates another question; why do our "brains" like music - why does the cerebellum find "pleasure in adjusting itself to stay synchronized"? There's no "find pleasure" without a subject, but who is Levitin's subject? The brain? I would say this: Brains do not find pleasure in doing things, PEOPLE find pleasure in doing things! People like to have expectations about the music, people like to match rhythms and follow melodies. The brain is what makes our experience possible. It is not an aswer to the question "why do we like music?" to say "because the brain likes music!", because it just creates another question: "But why does the brain like music? Who is in the brain liking this music if not the person in which the brain is situated?" Levitin is not alone in presenting this flawed explanation; we see it everywhere in contemprary neuroscience.
It doesn't get better when Levitin tries to tackle complexity in music. To an adult, he says, there is music that is too simple to be challenging enough, so s/he will not like it so much. As the music gets more complex, s/he will like it more and more, but then there's a peak, a personal preference about how much complexity can be tolerated, and after that peak, the pleasure s/he gets from the music is falling again until the music is so complex that s/he will simply hate it. This is an inverted-U graph, Levitin tells us, is "intended to acount for [the] variable" why we might "like or dislike a piece of music ... because of its simplicity or complexity". I just want to say that the graph is simply wrong. Think about a beautiful fugue by Bach. Most people would like it a lot, and yet it is very complex. The reason is that people can like it DESPITE its complexity; i.e. you don't NEED to understand the complexity in order to like the piece. Still, it is also possible to penetrate the complexity and try to get a deeper understanding of the piece. There's pleasure on the surface as well as pleasure beneath it!
The final chapter about music and evolution is the most interesting one. I'm not sure I agree with Levitin when he says that a lot of birdsong is recursive; I would say that that statement is certainly a question that is still up for discussion. However, Levitin does pretty well in disproving Pinker's thesis that music has no adaptory function, and I would say that for those not interested in penetrating the whole book, the last chapter will probably do just fine on its own.
There are much better books on music out there. Anthony Storr's Music and the Mind is one, and also, half-way finished with Philip Ball's The Music Instinct, I would say that one is so far more substantial and less confusing.