Non-spectrality of some piecewise smooth curves and unions of line segments
Mihail N. Kolountzakis, Chun-Kit Lai · 2025
Read Non-spectrality of some piecewise smooth curves and unions of line segments by Mihail N. Kolountzakis, Chun-Kit Lai (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-00581.txt.
Introduction
Mihail Kolountzakis and Chun-Kit Lai study when curves can serve as a base for sound like signals in 2025. The math question is called spectrality. A shape is spectral if a set of pure tones can rebuild any signal on it. The authors build a general method based on a tiling equation.
They prove that arc length on all closed polygonal lines is not spectral. In plain terms, even the edge of a square fails the test. Two crossed line segments fail as well. Smooth convex curves with a few self crossings also fail. The work closes some long open cases. It ends with new open questions on singular measures and curves.
Catalog record
| CREATOR | Mihail N. Kolountzakis, Chun-Kit Lai |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-00581
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Non-spectrality of some piecewise smooth curves and unions of line segments Mihail N. Kolountzakis, Chun-Kit Lai Abstract . We develop a systematic study about the spectrality of measures supported on piecewise smooth curves by studying the support of the tempered distributions arising from the tiling equation of some singular spectral measures. In doing so, we show that the arc-length measures of all closed polygonal lines are not spectral. In particular, the boundary of a square is not spectral. We also show that the “plus space” (two crossing line segments) is not spectral. Furthermore, our theory also shows that the arc length measures on smooth convex curves with finitely many transverse self-intersections are not spectral. Finally, several natural open questions about the spectrality of singular measures and piecewise smooth curves will also be discussed. 2020 Mathematics Subject Classification 42C15, 42C30 1. Introduction In this paper, a measure \mu is always referring to a compactly supported probability measure in {\mathbb{R}}^{d} . We say that \mu is a spectral measure if there exists a countable set \Lambda\subset{\mathbb{R}}^{d} such that E(\Lambda):=\{e^{2\pi i\lambda\cdot x}:\lambda\in\Lambda\} forms an orthonormal basis for L^{2}(\mu) . A measurable set \Omega with finite positive Lebesgue measure is a spectral set if the normalized Lebesgue measure on \Omega is a spectral measure. The study of spectral sets of was first initiated by Fuglede in 1974 [ Fug74 ] , who famously proposed the conjecture that spectral sets are equivalent to translational tiles on {\mathbb{R}}^{d} . The conjecture was disproved in both directions on {\mathbb{R}}^{d} with d\geq 3 [ Tao04 , KM06b , KM06a , FMM06 , FR06 ] , but the research effort to explore the exact geometric relationships between spectral sets and translational tiling has not declined. In one recent major result, Lev and Matolcsi [ LM22 ] showed that Fuglede’s conjecture holds for convex domains. Readers can refer to the survey [ Kol24 ] for different aspects about the Fuglede’s conjecture. The generalized concept of spectral measures gained its popularity when Jorgensen and Pedersen discovered that the self-similar Cantor measure generated by dividing [0,1] into 4 sub-intervals of equal length and keeps the first and third one is a spectral measure, while the standard middle-third Cantor measure is not a spectral measure in 1998 [ JP98 ] . This work was followed up by Strichartz [ Str00 ] and Łaba and Wang [ ŁW02 ] . There is now a vast literature in this direction. Readers can refer to the survey [ DLW17 ] for more recent advances about fractal spectral measures. The connection with tiling appears to be weaker, but it exists. Gabardo and Lai [ GL14 ] showed that if the pair of measures (\mu,\nu) “tiles" [0,1]{{}^{d}} in the sense that \mu\ast\nu={{\mathbbm{1}}_{[0,1]^{d}}\,dx} , then both \mu and \nu are spectral measures. In the same paper, they also proposed a version of the generalized Fuglede’s conjecture for singular spectral measures. A verification of this conjecture in a special case of Cantor-Moran measures was recently given in [ ALZ25 ] . The purpose of this paper is to study the spectrality of singular measures supported on piecewise smooth curves. This first appeared in Lev’s paper [ Lev18 ] . Iosevich, Lai, Liu and Wyman showed that boundary of the unit disk never admits a Fourier frame (generalization of the orthonormal basis of exponentials), hence it is non-spectral [ ILLW22 ] . In the same paper, they also showed that the boundary of any polygons always admit a Fourier frame. This naturally led to the question to determine if the boundary of a square, equipped with the 1-dimensional Lebesgue measure, is a spectral measure. Notice a well-known fact that the square domain is spectral, thus it becomes an interesting question to determine if the spectrality of the domain can induce a spectrum on its boundary. We will answer these questions in this article in the negative. Lai, Liu and Prince [ LLP21 ] initiated a new study about the spectrality of symmetric additive spaces with measure supported on two orthogonal lines defined (1) \mu=\frac{1}{2}({\mathbbm{1}}_{[t,t+1]}~dx\times\delta_{0}+\delta_{0}\times{\mathbbm{1}}_{[t,t+1]}~dx). They proved that \mu is a spectral measure when t=0 . However, the case where the two lines overlap, i.e. -1/2\leq t<0 , is much more subtle. With a series of works by Ai-Lu-Zhou [ ALZ23 ] , Kolountzakis and Wu [ KW25 ] and recently by Lu [ Lu25 ] , it is now shown that all are not spectral when -1/2\leq t<0 . See Fig. 1 . Figure 1. The measure that is arc-length on two equal-length line-segments. Some of the measures on the left are spectral, depending one where the starting points of the segments are compared to their length. In the crossing case on the right they are never spectral. The symmetric crossing case, t=-1/2 , is called the “plus-space”. Apart from spectra, our results also deal with tight frames of exponentials, a slightly more general notion of basis. We say that \mu is a tight-frame spectral measure if there exists a countable set \Lambda\subset{\mathbb{R}}^{d} , which we call a tight-frame spectrum , such that E(\Lambda):=\{e^{2\pi i\lambda\cdot x}:\lambda\in\Lambda\} forms a tight frame for L^{2}(\mu) in the sense that there exists a constant A>0 such that (2) \sum_{\lambda\in\Lambda}\left|\int f(x)e^{-2\pi i\lambda\cdot x}~d\mu(x)\right|^{2}=A\int|f(x)|^{2}~d\mu(x),\ \forall f\in L^{2}(\mu). Clearly an orthonormal basis is a tight frame. But the converse is not true. For example, we can take union of two orthonormal basis to form a tight frame of constant 2. For general frame theory, readers can refer to [ Hei11 ] . For tight frame spectral measures, general theory was established in [ HLL13 ] and [ DL14 ] . In [ DL14 ] , it was shown that if \mu is tight-frame spectral and absolutely continuous with respect to the Lebesgue measure, then \mu must be a constant multiple of the Lebesgue measure on its support. The consideration of tight-frame spectrality allows us to deal with spectrality of curves by focusing on a subset of the curve (see Corollary 1 below). We say that a finite union of line segments is (tight-frame) spectral if the naturally induced one-dimensional Lebesgue measure is a (tight frame) spectral measure. Our main result in this paper is to settle the spectrality questions we mentioned. Moreover, we determine non-spectralities for much more general classes of line segment collections. Theorem 1 . (1) A finite union of line segments that forms a closed curve, self-intersecting or not, cannot be tight-frame spectral. (2) A finite union of line segments containing three lines that starts at the same point and points in distinct directions cannot be tight-frame spectral. In the second part, we remark that the theorem also includes cases in which two of the three vectors are in opposite directions to each other. Hence, it shows that the “plus space” corresponding to t=-1/2 (refer to Fig. 1 ) cannot be tight-frame spectral, providing another independent proof after Lu’s proof [ Lu25 ] . Some more examples following from our Theorem 1 are shown in Fig. 2 . Clearly, the first part of the theorem showed that the boundary of a square cannot admit a tight frame of exponentials. Closed polygonal curves can be generalized to certain smooth curves of positive curvature. Figure 2. Some