Hello, Per. Coincident and near-coincident stereo recording setups can sometimes seem strange and "unintuitive" because (apart from M/S techniques) neither microphone is aimed at the center of the direct sound source. But in stereo recording it is really the combination of the signals in the two channels that matters. One can hardly predict by listening to either channel alone how the stereophonic result will sound; it is a kind of system and must be considered as such.

To address your specific questions: A good small-diaphragm cardioid microphone will have nearly identical pickup quality across quite a wide angle in front of it--though directly on-axis, of course, it will be somewhat more sensitive. You can get a quantitative sense of this from polar response curves. Uniform frequency response at all angles of sound incidence across the front hemisphere is a key ingredient for stereo arrangements such as ORTF; they aren't nearly as effective with (for example) large, dual-diaphragm microphones.

Another important fact that can be seen from polar diagrams is that cardioid is just not a very "sharp" pattern; whoever decided to call it "unidirectional" was probably a marketing person, not an engineer. At 45 degrees a cardioid's sensitivity is still within 1.5 dB of the on-axis figure. And that is precisely why it's important to aim a stereo pair of coincident or near-coincident cardioids fairly far apart (e.g. 110 degrees for ORTF): to create enough difference between the left and right channel pickup for good stereo. With any narrower angle, more signal is picked up identically and simultaneously in both channels, creating more of a mono result. Any cardioid microphone is already "50% omnidirectional" so to speak--so having too much signal in common between the pair of microphones is really something to be avoided, or else the recording will lack all sense of spaciousness.

As to your remark about vocal recording: Actually, sometimes it is a very good idea to record a vocalist from 45 degrees off-axis, depending of course on the room and the location of any other sound sources. A cardioid microphone that is addressed from off-axis will have less proximity effect and less chance of picking up breath noise and "popping" from consonants. Especially in classical music recording (e.g. recording sessions for operas) you will often see a vocal microphone positioned above or below a singer's mouth rather than directly in front of it. The microphone is then picking the singer up from off-axis, since a directional pattern works in both the horizontal and vertical planes at the same time. So your "rhetorical" question was actually a very good practical question.

--best regards