To apply phase closure with an array of optical telescopes, each apert
ure must be phased. This sets an upper limit on the size of each apert
ure equal to the coherence area of the wave fronts (if no adaptive opt
ics is used). By recording the wave-front on each aperture and by post
-processing, it is possible to apply phase closure algorithms with ape
rtures sustaining an arbitrary high number of coherence areas, as far
as the combination remains coherent. This procedure increases substant
ially the signal-to-noise ratio in the visible and stands as a complem
ent when adaptive optics only provides a partial correction of wave-fr
onts.