CONCERTO MPI optical concept. The main beams, B_p and B_s, described in Eqs. (3) and (4), enter the first input port of the MPI, while an unfocussed disc (D) enters the second port. A first polarizing grid, P₁, is at the entrance of the MPI. A second grid, P₂, divides the light between the two MPI arms, each ending at a rooftop mirror (two plane mirrors at 90 deg. angle), M₁ or M₂. Mirror M₁ is fixed, whereas M₂ oscillates along the horizontal axis, providing a variable OPD of 2Δd. Grid P₂ then recombines the light and sends it into the cryostat. The light is analysed by a third polarizing grid, P₃, which provides the two MPI output ports feeding two KID arrays, LF and HF. Each spectral component (ν) of the incoming radiation interferes constructively or destructively, depending on the moving-mirror position (Δd). The total output signal on one detector, S (Δd), is the integral over the entire wavelength range of the input spectra, modulated by the OPD between the two arms. By performing a Fourier transform of this signal as a a function of Δd, the spectral intensity, I_v, of the incoming radiation can be recovered, enabling precise spectroscopic analysis.