Figure 1
The iterative SMAR phasing algorithm in four stages. (Upper right-hand corner) Initial (or updated) φ phase estimates belonging to set Φ are combined with observed |E| values to obtain ρ and ρ(Φ)sρ = |ρ(Φ)| [the superscript 1) indicates that ρ is stored]. (Upper left) The FT of |ρ(Φ)| is calculated to get the new set χ of α phases as well as the calculated |ξ| values. (Lower left corner) The new α values are combined with the experimental ΔE = |E| − 〈|E|〉 and Fourier transformed to obtain δM(χ). The mρ mask and the sρ signs are derived from the stored ρ(Φ), and the δM(χ)sρmρ = ρ′ product is carried out. (Lower right) The FT of ρ′ supplies the updated φ phases [the superscript 2) indicates that when ipp is applied, ρ′ is further modified to ρ′′ before the FT operation] as well as the calculated values. |