The received wisdom is to record "flat, dry and clean".
"Flat" of course refers to the frequency response.
"Dry" as above but also without dynamic
control or reverb.
And "clean" means at a low enough level such that there is no possibility of any overload. This means an average level of -18, even -20dBFS (24bits assumed). The main
reason for all this is that none of the above FX can be perfectly undone*.
It is true that in the bad old days of tape some "sweetening" was used going in, especially compression because of the poor dynamic range of tape and the fact that post processing caused a build up of noise. Us digital folk have no such problems! Even so, the "classical" boys tried to stay "F,D,C, as far as practically possible.
The point has been made of the technical limitations of cheap mixer EQ. Well they are actually not that bad if you avoid extreme boosts and in practice MOST post EQ tends to be cuts? It is however true that the comps', reverbs and EQ to be found in Cubase lite or
Reaper are likely to only be bettered by several 100 $ of hardware even if possible at all.
Of course you CAN have the best of both worlds! Take a single channel recording of an E guitar? DI or amped cab? Well you can do both. DI channel one and mic up into ch 2. You might need to "slide" one waveform to line them up perfectly if going for a mix. It does not take much imagination to extend the technique to have both dry and processed channels?
*But, wow! Just a few days ago I read of some "De-reverbing" software!
Dave.