The Mahler measure of the Rudin-Shapiro polynomials (Q295810)

The Rudin-Shapiro polynomials are defined as follows: \(P_0(x)=Q_0(x)=1\) and \[ P_{n+1}(x)=P_n(x)+x^{2^n}Q_n(x), \quad Q_{n+1}(x)=P_n(x)-x^{2^n}Q_n(x) \] for \(n=0,1,2,\dots\). In Theorem 2.1 the author shows that the Mahler measure of \(P_n\) and \(Q_n\) is at least \(c_1 \sqrt{N}\), where \(N=2^n=\deg P_n+1=\deg Q_n+1\) and \(c_1>0\) is an absolute constant. Setting \[ M_0(R,[\alpha,\beta])=\exp\Big(\frac{1}{\alpha-\beta} \int_{\alpha}^{\beta} \log|R(e^{it})| \,dt \Big) \] (so that \(M_0(P,[0,2\pi])\) is the Mahler measure of the polynomial \(P \in {\mathbb C}[x]\)), he also shows that \(M_0(P_n,[\alpha,\beta]) \geq c_2 \sqrt{N}\), where \(N=2^n=\deg P_n+1\), \(c_2>0\) is an absolute constant and \(\alpha<\beta\) and \(N\) satisfy \[ \frac{32\pi}{N} \leq \frac{(\log N)^{3/2}}{N^{1/2}} \leq \beta -\alpha \leq 2\pi. \] The same holds for \(M_0(Q_n,[\alpha,\beta])\).