But if we were listening to a baritone singer (range from 98Hz upwards) in an untreated room, would we wince at the coloration on his voice, perhaps finding it unrecognisable? Would we suggest he instead sang through some sort of PA system fitted with an EQ filter to quell some of the terrible resonances at source?
I don't think we would even notice the room unless it had a real humdinger of a resonance and as a result the EQ'd version would sound wrong. I think this is why Linkwitz suggests that any room in which you can have a comfortable conversation is good enough. The room forms a complete, coherent system where every reflection gives the corresponding boosts and cuts in frequency response, reverberation and resonance entirely consistently. If you EQ the source then you break the system. Can the ear/brain detect this, even if only subconsciously?
EQ is different from room treatment. Room treatment still gives you a complete, consistent acoustic system.
Yes, well, small rooms especially have problems that larger rooms do not have to the same degree. But, even some larger performance venues are deemed acoustically more preferable than others to discerning audiences. I think we can and do hear the difference in performance acoustic, live or recorded, small or large room, but some people are not as consciously aware of it depending on experience. They can learn to hear it more consciously, if they have an interest in doing so.
Yes, of course, we would not want your baritone singing through an EQed PA system, which would introduce colorations of its own, even if the EQ itself did not and it counteracted some room problems. So, your example is sort of specious, especially when applied to pure playback of source material recorded in a more ideal venue.
The playback room may be "complete and coherent" while it is still distorting the source, and this is almost certain if it is a small room. So, it is not ideal, even if Linkwitz says it is good enough when a "comfortable conversation" can occur with its limited frequency bandwidth. Yes, the listener's ear/brain can compensate to some degree, but most listeners prefer sound where most common acoustic problems have been minimized. And, some sounds at certain frequencies will be completely inaudible due to reflective cancellation at some locations in many rooms unless the acoustics are corrected.
The pattern of acoustic variation in any room is not consistent. There are variations inch by inch around the room - small ones but they become more significant foot by foot or yard by yard. Treatments can be good, but they do not eliminate acoustic variation unless the room is anechoic. Properly done, they might lessen that variation, but some still inevitably exists. Proper acoustic treatment including careful measurement is also difficult and costly, especially if room aesthetics are an issue. In the deep bass frequencies, huge structures are required to deal with the long wavelengths. It is also technically difficult to achieve best performance via treatments alone. EQ has been increasingly used by acousticians as a better solution, coupled with multiple subwoofers for best uniformity. EQ has been shown to be sonically quite transparent, especially sourced from digital recordings. The result with proper execution can be sound that is more consistent and consistently better.