no. 16 · Tuesday 17 November 2026
Entropy and frame measures for self-affine measures
Abstract
Let $M$ be a real expanding $d\times d$ matrix and let $\mathscr D=\{d_0,\ldots,d_{N-1}\}\subset\R^d$ consist of $N\ge2$ distinct vectors. We study the existence of frame measures for the self-affine probability measure $μ$ generated by the maps $S_i(x)=M^{-1}(x+d_i)$ with equal weights $1/N$. Extending the entropy obstruction for self-similar measures with a common scalar contraction, we prove that $μ$ admits no frame measure if the conditional entropy of the first symbol given the spatial point is positive. In particular, this holds whenever $N>|\det M|$, without assuming absolute continuity or imposing arithmetic conditions. When $M$ has integer entries, $\mathscr D\subset\Z^d$, and $N=|\det M|$, the measure $μ$ admits a frame measure if and only if $μ$ is normalized Lebesgue measure on its attractor.
- fields
- Functional Analysis