The same procedure utilized above to eliminate a posteriori the quadratic phase was applied to recover phase maps for the RBCs when operated in the nontelecentric mode. Owing to the RBCs leaving large free areas to gather the needed information for the numerical compensation, the recovered phase maps for the nontelecentric DHM and telecentric DHM look alike, as shown in Figs. 6(a) and 6(c). Despite their similar aspect, in the nontelecentric case, the phase measurements are strongly dependent on the location in the FOV.31 To see that, the areas inside the pink squares are zoomed in Figs. 6(b) and 6(d). These closer views reveal the following. (1) The reported shift-variant behavior for the nontelecentric DHM is evident. In Fig. 6(b) two individual RBCs are zoomed in and their phase values quantified. As visualized from these zoomed images, there are important differences in the measured phase for the apparently identical RBCs but placed at different regions in the FOV. On considering that the measured phase is related with the thickness of the sample , the refractive index of the sample , and the surrounding via , the distorted phase measurement could mislead illness diagnostic, for instance. (2) The telecentric DHM is a shift-invariant QPI. The same RBCs analyzed for panel (b) are studied in panel (d). As , ,47 and , the measured phase can be converted in specimen’s thickness. Inside the experimental errors, which are in the order of , the measured thicknesses for the RBCs placed at different locations in the FOV are the same. (3) Apparent morphological changes are introduced by the nontelecentric DHM. As the zoomed individual RBCs are compared, those reconstructed from the nontelecentric DHM present an internal structure that differs from that seen in the RBCs reconstructed from the telecentric DHM. In summary, even though there is enough free area in the FOV to gather data for computing the compensating phase inherited from the nontelecentric DHM, the numerical a posteriori elimination of that phase leaves remains that perturb the QPI. This could mislead this to illness screening, specimen identification, or others diagnostics that resort on label-free imaging of transparent or semitransparent samples.