theorem
largeSieveMultConvLHS_le_combined
(N₁ N₂ M : ℕ)
(hM : 0 < M)
(f : Fin N₁ → ℂ)
(g : Fin N₂ → ℂ)
:
LargeSieveMultConv.largeSieveMultConvLHS N₁ N₂ M f g ≤ ((↑N₁ / ↑M + ↑M) * (↑N₂ / ↑M + ↑M)) ^ (1 / 2) * LargeSieveMultConv.l2Norm f * LargeSieveMultConv.l2Norm g
Combined bound used as a black box for the main theorem: the LHS is at most $((N_1/M + M)(N_2/M + M))^{1/2} \cdot \|f\|_{L^2} \cdot \|g\|_{L^2}$ (without an extra constant).
Large sieve and multiplicative convolution theorem: if f : [N₁] → ℂ and g : [N₂] → ℂ,
then f *_M g : [N₁ N₂] → ℂ, and
$$\sum_{p \in P_M} \|(\pi_p(f *_M g))_h^*\|_{L^\infty}^2 \lesssim \big((N_1/M + M)(N_2/M + M)\big)^{1/2} \|f\|_{L^2} \|g\|_{L^2}.$$