Before compressors existed, levelling a vocal was done the only way possible: someone kept a hand on the fader through the whole take, lifting the syllables that dropped and pulling down the ones that jumped. It's still the most transparent way to control dynamics, for a simple reason: it doesn't change the sound, only when it's louder. A compressor, however well set, alters the envelope. Pulling a fader down two decibels alters nothing.
The real difference from a compressor isn't quality, it's information. A compressor reacts: it detects the signal crossing a threshold and acts afterwards, with no idea whether that syllable mattered or whether a shout was coming. You anticipate, because you know the song. That's why a professional mix doesn't choose between the two, it chains them. With the coarse work done by hand, the compressor barely has to act — and that's where it stops being audible.
Here a track plays and you ride the fader in real time, with one goal: keep the output pinned at −12 dB. There are 11 scenarios, because every source runs away differently. The levels don't change the exercise, they change how much you're forgiven: the green zone goes from ±6 dB down to ±0.6. Anticipating a peak scores higher than correcting it late — which is precisely the advantage a compressor doesn't have.
The next step in the dynamics chain is understanding the compressor, and that's in the compression guide. To decide when it's a fader job and when it's a compressor job, mix diagnosis.