Speaking from the recording side it's much harder to play/sing with too little or too much reverb.
Not really. Musicians regularly sing/play under conditions where at their position there is very little reverb, and occasionally too much as well. They are used to that. It is much more difficult if there is delayed reverb/echoes at play, or you do not hear yourself at all (that is where you find a lot of musicians using headphones or IEM).
If you've ever recorded a live performance from your normal audience location you've probably noticed that the reverb that sounds wonderful live, coming from all directions, sounds artificial when reproduced from a pair of speakers in your living room (or with headphones).
That is mostly because the directional information of all parts of the reverb gets lost with conventional stereo recordings, so all the reverb from all angles, which sounds natural in the concert hall, gets summed up from the stereo speakers and that feels wrong to our brain. So it is clearly a problem of loudspeaker stereophony.
If you repeat this experiment with either a dummy head/binaural mic and playback on calibrated headphones, or a multi-channel main mic arrangement, for example for Dolby Atmos, the result is very different. It sounds surprisingly ´realistic´ in terms of imaging (which does not make a perfect recording, though).
Pro live recordings usually have the microphone around the conductor's location/distance or closer. O with rock music everything is close-mic'd and multitracked
You are referring to a main mic arrangement for classical recordings. This is meant to capture a ´total picture´ of the reverb soundfield, and is usually placed far away from the floor in order to maximize first reflection delay, and at a place to ensure a proper ratio of direct and indirect sound. Which in many concert venues is further away from the instruments than the conductor (the conductor actually has a pretty distorted balance with instruments like violins, celli and brass overly dominant).
Close-mic´ing and multitrack is in classical recordings as common as in rock. These are called spot microphones, and they are gradually added to the signal of the main microphone to ensure clarity and restore balance between the instruments and between direct sound vs. indirect sound alike.
Somewhere I read that listening in an anechoic chamber can give the perception of the sound coming from inside your head,
Have tried conducting listening tests under anechoic conditions and cannot really confirm this. In some cases of very ´dry´ recordings this can happen, but with a minimum of meaningful reverb in the recording, you usually get a stereo imaging even under anechoic conditions, it is just very constrained in width and depth, with drastically reduced perceived proximity. So it is pretty likely that voices sound as ´coming from directly in front of your nose´, not really from inside your head like with IEM.
Notice that listening to live acoustic music performed outdoors sounds meh, weak, disappointing.
Cannot really agree to that. I have heard numerous un-amplified classical performances under free-field conditions, with the most extreme being a performance of Gluck´s ´Orphée et Euridyce´, featuring a contralto, in a courtyard of a Rokoko palace, as well as Orffs´s ´Carmina Burana´ in a gorge (!), and was pretty amazed.
You have to get used to it, though, as our brain notices a contradiction between reduced SPL with almost no lower frequencies (which usually indicates sound sources further apart) and close to no reflections at all (which signals our brain that instruments and voices are very very close to our ears).
