Ideas cannot be copyrighted, only ideas that have been "fixed".

Since the advent of songwriters and songs in general, the definition of
what is "fixed" has involved less and less notes put down on paper.

Copyright allows the concept of "derivative works", but only with
permission.  Creating a derivative work without permission is
infringement, even if the circumstances can be shown to be "innocent."
Proving this kind of infringement involves proving that the work being
challenged contains enough original material from the copyrighted work.
The easier that is to prove, the more likely the challenge will be
sustained.

Registered songs, rock and otherwise, are ideas fixed in their most
primitive state.   The bare outline of single notes with chord
markings, no fleshing out of the ideas into true "compositions", just
the bare bones of the tune.   The sheet music of a rock tune that you
see in record stores is probably not the form that has been registered
with the copyright office, but rather an arrangement ("derivative
composition") created by permission (a work for hire).

Art songs, such as those by Ned Rorem or Alan Hovannis are probably
registered as the whole composition, not a stripped down version to get
maximum coverage.

So if rock star 'A' comes up with a riff that sounds like that of rock
star 'B', that is an infringement unless rock star 'A' can prove that
he himself did not base his riff on someone else's, living or in the
public domain.   More importantly, if the registered idea is primitive
enough, then four notes can be an infringement since anything more
elaborate could be shown to be "derivative."

Get the point yet?   Register the song in a form sufficiently vague so
you can challenge as many people as possible, but sufficiently fixed so
you can show that those being challenged have created "derivative"
works.   It's an art in itself.

I read recently where the court denied copyright protection when it was
shown that the fixed form of the idea being challenged was not
sufficiently original (Apple vs. MSWindows for instance, where it was
shown that both had copied from the Altos).   Even more interesting is
those cases where it has been shown that the inventiveness of the bulk
of the challenged work so overwhelmed the part that was in question
that infringement had not taken place.  Also, that there was a logical
progression of ideas from previous compositions from the challenged
composer, and that the composer could prove that (s)he did not have to
rely on theft for success.  Very nebulous terrain indeed.

All of this is horribly expensive to prove, which is why you mainly see
this type of "innocent infringement in a derivative composition" case
in predatory circumstances.

George Bogatko