But you're talking about two different things here.
@JoachimStrobel spoke about reproducing actual acoustic instruments and music. Then there's a genuine debate about how to record and reproduce it - which position in the hall, at the stage, close-micing or not, etc. You're talking here about live music over PA, if I understand you correctly. That's a whole different can of worms. As far as I'm concerned talk of reproducing PA concerts has nothing to do with high fidelity at all. What sense is there in trying to reproduce what is essentially already a reproduction? With electronic music, the original "event" is the studio-mixed record. It has no real-life correlate. With acoustic music, the original auditory event we may try to reproduce is the actual performance - and one may discuss what may be the best way of reproducing this event in our listening setups.
I was actually making the same distinction but you are further confusing the issue by misunderstanding the reference to PA. It is useful to think of the audio system for recording and for amplifying the sound inside the theater as two separate but parallel systems even if they use the same mixer and mics. Also, let us not further confuse the use of amps and speakers required to get
any sound out of an instrument like an electric guitar or the keyboards from the PA system to amplify the stage for the audience at a distance. You may mic the former to feed to the PA system but typically never mic the latter if you can help it, partly to avoid feedback if you are using the same system for recording and room PA (so there is no recording of a recording as you put it).
Think of those required electronics for some instruments as part of the instrument itself. Keeping these conceptual differences is key to understanding how audio engineering works for live performance or recording.
Talk of the PA system here is a red herring. Hypothetically, think of a live performance without ANY amplified audio inside the room that is good enough for an audience (other than any amps needed for electronic instruments which have no sound otherwise). Now, let us think of how to record it so it can be heard at home.
Basically, it boils down to how to mic it and how to mix it.
There are two different issues - how to capture the quality of the sound and ambience and how to represent the stage in terms of locating the listener relative to the stage.
These are two separate issues. These have nothing to do with whether there was a PA system (for stage amplification different from instrument required transducers) in the location or not. This must be clearly understood before proceeding further.
Any mic positioning for recording is at best an approximation to the live performance. But you are not going to get a recording without mics. So, a whole part of audio engineering is to determine the position of the mics so that it captures the sound quality of various instruments and/or vocals. Think of this as just raw data before mixing. You can also mic to capture the ambience to input into the mix. So far, it has nothing to do with how many channels in recording output. These are all independent cables coming into the mixer from various positions and have no relation to the speaker channels at home yet. As I have mentioned, there are limitations to capturing the timbre and ambience of a performance perfectly but that is a given. These have nothing to do with if there was a PA system in the theater or not.
Now, once you have the raw feeds coming in, there is the question of how to mix it for the final output into a recording whether stereo, or multi-channel. But regardless of the latter, there was a discussion earlier of whether the position of the listener being represented in the recording is as being located in the audience or in the midst of the performers. This should not be confused with the other issue of quality of sound capture. You can use the exact same mic positioning and their raw feeds that gives the best quality you can get to mix it for either of those listener positions. You can do it in stereo or multi-channel for either.
My first point was that mixing the output from the perspective of a listener in the middle of the performance has no basis in reality of a listener being in the middle of a stage in a live performance. The latter in live performance is not good because of lack of balance between instruments on stage (with some exceptions as I have mentioned). So, you can create a virtual world from the raw feeds coming in and mix it as if the sounds were coming from all around you rather than in front of you. This is what some of the multi-channel SACDs do when they remix from original independent tracks even of a live performance. The goal here is never one of capturing the actual performance in terms of positioning even for a live performance for reasons mentioned above. Conceptually, they are not very different from some of the post-processing surround modes that try to approximate it from stere recording outputs.
Second, there is a misunderstanding that recording from an audience location perspective is to be done by recording what comes out of the PA system. In general, this is not true (with some exceptions for ambience or in budget/amateur gigs). It isn’t. It is created by mixing the same incoming mic feeds that are fed to the PA system if one exists. In budget/amateur settings, the two mixes might be the same but in most professional settings, it isn’t. The PA system isn’t typically just 2 speakers. The PA mixing goal here is to amplify the sound so that ideally it is as if the stage is playing louder to be heard in the audience NOT that the saxophone is moved to the left speaker. This is achieved via a combination of strategically placed speakers and mixing feeds in right proportions between them.
The mixing for recorded output from the audience position is conceptually different although it has the same goal as the PA system - to make the stage be heard as a stage in front of you. Just to be clear again, this is typically
not done by mics in front of the PA system to capture what the audience is hearing except in budget/amateur productions.
If both are done correctly and professionally, the image of the stage projected with PA amplification in the performance is the same as the stage projected by the recording at home for a listener at the center of the audience (or wherever the mixing engineer’s reference point is), in reality a decent approximation. So from that perspective, it is capturing the live event as heard in the audience (but
not by placing mics in front of the PA system, I keep repeating this because this seems to be misunderstood).
Multi-channel output capability can enhance either of the listening position reference points. In terms of the virtual world of being on stage to listen, it is almost a necessity. In terms of the reference point of being in the audience, it can be used to enhance ambience and the “live nature” in terms of audience presence if that is required.