From Scholarpedia:
Quasiperiodic oscillation is an oscillation that can be described by a quasiperiodic function, i.e., a function $F$ of real variable $t$ such that $$F(t)=f(\omega_{1}t,\dots,\omega_{m}t)$$ for some continuous function $f(\phi_1,\dots,\phi_m)$ of $m$ variables ($m\geq2$), periodic on $\phi_1,\dots,\phi_m$ with the period $2\pi$, and some set of positive frequencies $\omega_{1},\dots,\omega_{m}$, rationally linearly independent, which is equivalent to the condition $$(k,\omega)=k_1\omega_1+\dots+k_m \omega_m\neq0$$ for any non-zero integer-valued vector $k=(k_1,\dots,k_m)$.
My first question is: how one define the power-spectrum of $F$? I think that power-spectrum of $F$ should be defined as: $$S(\omega)=\lim_{T\rightarrow\infty} \frac{1}{T}\left|\int_0^T F(t) e^{-i\omega t}dt\right|^2$$ However, the real question is the following. I expect that $S(\omega)$ is formed of discrete peaks corresponding to the basic frequencies $\omega_1,\omega_2,\dots,\omega_m$ but I would understand if $S(\omega)$ is in turn a quasi-periodic function. Are there any study on power spectrum, $S(\omega)$, of quasi-periodic functions?
Bibliography references and answers are welcome.
