Just to clarify. From the above it seems the RX7 Azimuth tool fine tunes the timing relationship between the two channels by shifting one channel with respect to the other by some number of samples. With the number of samples of the shift being determined either by direct input from the user or via its "learn" analysis function.
In that case, it is essentially moving one channel forward or backwards in time with respect to the other, by small amounts.
Think of the correction not as changing the spacing between two microphones (it can't change that), but rather more like moving one microphone slightly closer or further away from the source in comparison to the other. In that way its more like adjusting the rotation of a pair of spaced omnis slightly. But not exactly as it is actually analogous to moving the microphone closer or farther from all sources of sound around it, not just those arriving from the front.
More detail-
Coincident stereo microphone configurations are intended to produce no significant timing difference between the two channels. The Azimuth tool applied to them can correct (eliminate) a timing difference between channels that shouldn't be there.. or impose one if that were desired for whatever reason.
Spaced configurations (both near-spaced and wide-spaced) will produce timing differences between channels that vary depending on the location of the sound source. Sound sources on the medial plane (those sounds arriving perpendicular to the imaginary line between the microphones, or in other words, those arriving from directly in-front, above, behind, or below) are intended to produce no significant timing difference between channels, the same as a coincident arrangement. However, sounds arriving from off-center will produce timing differences that increase proportionately depending on how far off center they are. The largest timing difference will occur for sounds arriving from 90-degrees directly to the left and/or right. Sources less far over to one side or the other will produce less timing difference. This basic relationship between angle of arrival and timing does not change regardless of how far apart a spaced pair of microphones are spaced. What does change with spacing is the overall range of the timing difference.
Good to be aware of this when manually identifying timing offset between channels. Since the arrival angle in the stereo panorama determines the offset, try to use the impulse of a sound you know is supposed to be centered with respect to the array.
Azimuth adjustment won't change that basic relationship, but rather shifts the overall timing between the two channels. The Azimuth tool can correct (eliminate) a timing difference between channels that shouldn't be there for sounds arriving perpendicular to the spaced array.. or can serve to intentionally shift the angle at which sounds arrive in unison away from perpendicular to some angle to the left or right of directly ahead (above, behind or below).