Within each fixed unitary irreducible representation, choose the selected right singular vectors to be orthonormal. Their corresponding left singular vectors are also orthonormal, even when a singular value is repeated: choose an orthonormal basis in each eigenspace of and put . Therefore, when ,
This is (iv), with the unnormalized Hilbert-Schmidt inner product and the Kronecker delta. No orthogonality of differently shaped matrices is being asserted.
For the unheaded continuation, partition into blocks . Then
The Schur averaging of rectangular matrices formula is
for any matrix . The average is an intertwiner; the Schur lemma makes it zero for inequivalent irreducible representations, and a scalar multiple of the identity for the same representative. The trace determines that scalar in the latter case.
Apply this with . The Hilbert-Schmidt inner product normalization just proved gives
Expanding the squared Euclidean norm now yields
Thus the final identity follows from averaging, although itself need not be the identity. The whole construction is the spectral inverse theorem for the matrix-valued U2 quantity.
For chosen inequivalent unitary irreducible representations of a finite group and a matrix , the average is zero when . For , it is . This follows from the Schur lemma, since the average intertwines the two group representations. Equivalent representations in different bases require the corresponding intertwiner; the identity-matrix formula assumes literally the same representative.
For inequivalent chosen unitary irreducible representations of a finite group, uniform expectation gives
The Kronecker deltas in the first case refer to entries in the same chosen basis. Averaging an intertwiner and applying the Schur lemma proves the formula; summing diagonal entries gives character orthogonality.