Gain Staging sounds like a fancy term, like some kind of weird platform to put something on. In reality, gain staging is an audio engineering term for getting levels right so your signal-to-noise ratio (the difference between your voice recording and your room noise) is the best it can be. When you set up to record and adjust the gain or microphone volume knob on your interface, you are gain staging. Again, when you master your audition, or read and output it, you are gain staging. And these can be separate things; for instance, you do not need to worry about ACX levels when recording for ACX—you will set those levels, that gain stage, when you output your finished files.
This is all part of setting a proper recording level for your voice; the signal (your voice) to noise (your room noise) ratio should be as good as it can be, and this is simple to accomplish: just turn up that gain knob. Ideally you will see your peaks hitting around -9 to -6 without hitting 0. Your volume will go up and down, of course, as the human voice is a very dynamic instrument, but with your peaks (the highest or loudest points in your recording) hitting around -9 to -6, you can rest assured that your levels are where they need to be, which is as far from your noise floor as they can be!
Once you’ve set a good level and recorded your script, you can go about editing your recording. Once that’s done, it’s time to prepare your recording for output, the second gain stage. Often people use the normalize function (peaks set to -3 is most common) to get their levels dialed in, but if they gain stage properly on the way in, they may actually see levels go down a bit when normalizing. This is because the peak levels were already higher than -3, even if just a bit. When you set a proper level, your signal-to-noise ratio is in the best place, and your room noise is as far from your signal (your voice) as it can be. If you do not gain stage properly prior to recording and then you use the normalize function, you are also bringing up the noise, and that noise floor can be the difference between acceptance or rejection of your recording. This is the main reason I am not a fan of normalizing the way it is typically used in the voiceover world; it has become a bit of a crutch for folks who have not put in the time learning to set levels properly.
To sum it up, when you record, you want your room noise to be as far from your voice as possible—this is the signal-to-noise ratio. To accomplish this, you should record your voice at a high level and aim for the peaks (the loudest parts) to be right around -9 to -6. Don’t stress if they go a bit higher, but don’t let them hit 0. Then, when you finish editing and are ready to output your file, you can utilize the normalizer to get your levels to wherever they need to be, but don’t be surprised if your levels actually go down slightly!
Lots more to learn, including how to use compression and EQ to not only master but enhance your recording so it can be the best representation of you. If you’re curious, sign up for my 5-week engineering course or book a private tech class.