True, partially. My perspective is one of a live sound professional. A vocal mic, with a vocalist in front of it, can be a very dynamic source as the performer moves around, and sings various parts of their song. They cannot be constantly in one place when singing, and sometimes vary their distance to the mic intentionally in order to manipulate the proximity effect of some mics. -- However, as they move, the signal level changes.
We use compression to level out those peaks, and to generate a more consistent level for our mix. This is because in order to have consistent relative levels between our sources, say an electric guitar running through an amp modeler, and the aforementioned vocal mic, we must ensure that they have a semi-consistent dynamic-ness.
So since the vocalist is very dynamic, we must compress them to allow a consistent relative level to our guitar.
And then, once we have a consistent mix between our vocals and our guitar, we can add some reverb into the mix to give the vocalist a more "full" sound.
But this is not on the topic of designing speakers to reproduce live music.
I will add one anecdote: I recently did an event with a performer, an electric guitar, and some backing tracks. Obviously I had a PA system, comprised of good quality speakers, which were powerful enough to hit my SPL target in the area I wanted to cover.
Despite me compressing the vocal mic signal, and using reverb and EQ, it still sounded "real". My theory is that it sounded real since I still had plenty of headroom, and because I left the vocal relatively un-processed (at least compared to most studio mixes).