Thanks for all the interesting input to this thread.
My take on this is that even if you are taking a purist stance, there are some legitimate step that can be taken to capture the natural sound source. So, it may be that the vocalist has made a perfect performance but that he/she just swayed a few inches away from the microphone and the signal level has dropped for that brief moment. Remember with the inverse square law you only have to move a little to have a relative large effect on the SPL reaching the microphone. There are then three options;
You leave the take sounding naff ie you haven’t been able to capture the true performance.
You use compression whilst recording or during mixing (this is the sort of thing compressors were designed for)
You “correct” the take using judicious gain riding, thus offering the opportunity to use less/no compression.
If there are any other options that I’ve overlooked, please let me know but as far as I’m aware you have to make one of these three choices.
Just a couple of other points:
A) Remember that your eyes and ears are able to cope with a very wide range of intensities (from candle light to bright sunshine and from hearing a pin drop in a quiet room to a loud rock band). Cameras and microphones don’t have this dynamic range and so compression of the dynamic range is inevitable to create an intelligible end product. Then consider the environment in which the music will be reproduced. If it is in a quiet room you can use less compression, but if the music is played back in a car, then the loud sections may end up too loud and the quieter sections inaudible.
B) All reproduction of a sound source is artificial and there are many decisions that have to be made, even if you are a fanatic purist! Try recording a solo piano for example with a view to creating a stereo recording; are you going to record it from the players perspective or the audiences perspective? If from the audiences perspective, close to the piano, or far away? How wide will the stereo image be? Are you going to use a straight stereo technique or a Mid/Side technique that might replay better or mono playback? Will you use a high pass filter to get rid of rumble etc?
Modern music production has used signal processors like EQ and compression over many decades and I think most raw recordings sound pretty awful!