The same goes for other instruments. Very often I must automate instrument levels just to get them to make their important statement then back off a little to make room for others.
Once they’re under control, I can fine tune sounds and then levels. Sometimes I’ll automate instruments for creative reasons (so the dynamics of the instrument follow and exaggerate what I perceive to be the dynamic of the song or vocal). Remember the breathing, shifting and dancing mentioned in step 4.
7. Back to the vocal. Vocal processing and riding is very important. A small ride change can make a big difference. I ride vocals for several passes on a variety of loudspeakers in order to be sure of how things feel.
8. I then re-ride all of the background instruments to support and exaggerate the automated-vocal’s enhanced expression.
9. Next, I re-ride the solo instruments to work around the automated-vocal.
10. Almost done! The next step is to tweak the vocal rides by going through them carefully and improving what I can based on the automated background instrument tracks. This includes background vocal rides.
11. After an ear break and listening for obvious things to change on different loudspeakers and in different listening situations (next room, car, etc), I make any final changes I think the song needs. Usually at this point I’m listening for problems rather than creating new images. Sometimes I’ve made drastic changes at this point.
12. Finally, the fade. I find the places where the song feels like it should start to go and where it should already be gone. I then go back and try to start the fade so people do not notice (due to either the fade start time or the gradual initial slope of the fade) and fade the song out in a way that it continues to pull people in while it drifts off. Sometimes I’ll change some rides due to the fade so certain parts are the very last things heard.
Then I send it off and hope that the client does not disagree with the direction I went in during step 1. Sometimes I think that the song needs to go in a direction that is obviously different from what the artist had intended. In those cases I’m obligated to give the artist “Take 1” in which I use everything in the obvious way and also “Take 2” in which anything goes. In all my decades of mixing, “Take 1” was chosen only a few times. “Take 2” usually involves going back to step 1 and making drastic changes. Although sometimes you can salvage work from Take 1, often you have to start over…even with the vocal sound.
Mixing is an art, and can be an emotional rather than technical process. Mixing is about creating an illusion, regardless if the illusion is of a band on stage or something never heard before. I know when a mix is going well if I’m believing the illusion as I’m building it, and each new thing inspires the next.
Final word: Although different styles and songs will require completely different approaches, never forget that most listeners will never get past “hearing” only the vocal. The entire mix should support the vocal in every way possible, regardless of whether that means making a hard rhythm to push it or a lush rolling carpet for it to ride on.
Bruce A. Miller is an acclaimed recording engineer who operates an independent recording studio and the BAM Audio School website.