Equivalence classes of Parseval frames (Q2680742)

The paper studies the notion of ``projective-permutational unitary equivalence'' between two Parseval frames for finite-dimensional Hilbert spaces. Theorem 2 gives a characterization of projective-permutational unitary equivalence between two Parseval frames using their Gram matrices by certain operations on matrices. The notion of equivalence of frames for Hilbert spaces (including infinite dimensions) started from the paper [\textit{R. Balan}, Proc. Am. Math. Soc. 127, No. 8, 2353--2366 (1999; Zbl 0922.42025)]. One of the interesting results of this paper is that if an equiangular Parseval frame with \(2d\) vectors exists in \(\mathbb{R}^d\), then \(d\) is odd. We wish to remark that the existence of \(d^2\) equiangular lines for \(\mathbb{C}^d\) for every \(d\) is the celebrated, still unknown, Zauner conjecture, see [\textit{M. Appleby} et al., J. Phys. A, Math. Theor. 52, No. 29, Article ID 295301, 26 p. (2019; Zbl 07630978)] and [\textit{S. F. D. Waldron}, An introduction to finite tight frames. New York, NY: Birkhäuser (2018; Zbl 1388.42078)]. Theorem 4 (resp. Theorem 6) shows that upto projective permutational unitary equivalence, there exists a unique equiangular Parseval frame for \(\mathbb{R}^5\) (resp. \(\mathbb{R}^7\)) with 10 vectors (resp. 14 vectors). An algorithm is given that determines when a Parseval frame has a full spark.