The microphone technique used will depend on whether you want an accurate reconstruction of the soundfield at home (e.g. using ambisonics or even wavefield synthesis - WFS) or some sort of acoustically processed version, which is what the multi microphone method will give you. The multi mic version you describe is sort of old fashioned quad and that never really worked all that well if what was wanted in playback was to sound realistic and natural, hence the demise of quad except for its Dolby-type variants (5.1, 7.1, Atmos, ProLogic etc).

How about using a tetrahedral mic (e.g. Soundfield, CoreSound Tetramic, Brahma or even a home made version)? You can then mix the B-Format channels to make any virtual mic polar pattern you like and retain something closer to the soundfield at your preferred listening position. The B-Format archive retains the acoustic data, if I can put it that way, in pressure and velocity terms (channels). Spot micing or spaced mics can still be added to taste and mixed in, but you still retain the focus of the original listening position soundfield.

Meridian surround DSPs since the 565 support ambisonics in its UHJ 2-channel version and 3-channel B-format (both are horizontal isotropic systems).

G61RSL, G98-DH, G01, Quad 606, Quad 306 (x3), Quad-ESL63 (x 2 pairs, F&R), KEF LS50 (S), Quad 10L sub, Martin Logan Motif (CF) + miniDSP