Arguably the most important element of the song, the lead vocal is one of the hardest things to get just right. Often times I’m hired to mix entire projects because the producer, or band member doing the mix was struggling to get a solid vocal sound. So, I want to share my process with you and help you get better, more natural, clear, and open lead vocal tones.
The human voice is incredibly dynamic, and covers almost the entire frequency spectrum. It can also change or cross the spectrum quickly. Take the word “slow” for example. It starts with an S, which will be sibilant and need to be dealt with as an element on its own. The S starts high in the spectrum - anywhere from 4K and up, sometime as high as 12/13k. The middle of the word slow usually holds in a slightly closed “ah” sound, rounding off to a hard “oh” at the end. The changes in those vowel sounds will each happen at different frequencies, and the W will cause resonances down near 100-200hz. That’s just one word.
I like to start by controlling the S’s. I use the Waves DeEsser as the first insert in the chain to really knock down any sibilant sounds. I set it to split mode in this early stage, only reducing the sound above 5k or so. I’m really aggressive at this stage because once we add compression and brighten the vocal up with EQ, those reduced S’s are going to be nice and even and smooth. Don’t be afraid to hit the DeEsser hard at this stage, even if it sounds a bit lispy, well come back to it.
Next, the first EQ. I like a nice graphic eq like Fab Filter Pro Q2, because we’ll be making some tight cuts and the visual display of the frequency response is really helpful. Don’t really on your eyes alone, but if it helps track down the things you hear quicker, don’t feel like your cheating by using it! I’m listening here for anything harsh or piercing. Usually there’s some harsh resonances around 3.6k, 6k, and 8-10k. It’s okay to boost and sweep but make sure your only cutting things that really need leveling. The key word here is leveling, not removing. For example, I find that most microphones, specifically the Neumann u87, have a hard resonance at 3.6k. If I notch that down 4db it should balance that frequency when a note resides there but a vocal with virtually nothing in 3.6k will sound artificial and scooped.
Next, a tone shaping EQ. Here we’re looking to shape the general curve of the vocal and usually add some brightness and sheen. I like the hardware models in Steven Slates Virtual Mix Rack, as well as the waves API and SSL eqs. I’m focused here on getting the vocal to hit in some core spots. 1k for closeness, 2-5k for bite and a bit of closeness, 6k and up for presence and sheen. I start by boosting 1k until the vocal sounds a bit too midrangy. I’ll reduce a bit, but not all the way. I’ll balance that midrange with a high shelf boost at 10k. This creates a scoop from ~2-6k which can be nice to smooth out the vocal. If it feels too smooth or scooped I’ll push a bit in that area as well.
Next up is compression. The goal this far has been tonal leveling to get it feeling balanced before compression, now let’s balance the dynamics a bit. Compression is very audible on vocals and can make or break a song depending on genre. A rock or metal song is going to be fine with a heavily compressed, forward, and aggressive sounding vocal. The same settings would sound out of place if used on a softer acoustic or piano song. I like to compress in stages, adding only a little bit at a time. Keep in mind that ratios stack in multiples. For example, let’s set up compressor A and B in serial, A into B. Both have a 4:1 ratio. The sound coming out of compressor b is seeing a 16:1 ratio (4x4). So, a little here little there actually goes a long way. I start with a slow attack and work my way faster with each new compressor. The first compressor may take off up to 10db if need be, so I keep the attack slow to not hit too hard - unless that’s what the song calls for. Next, a faster compressor taking off much less - maybe 2-4 dB. The goal here is to level out the inconsistencies leftnover from a slow attack time on the first compressor.
At this stage the vocal should be balanced and level. Check back to your DeEsser to make any final tweaks. Any new resonances driving you nuts that the first graphic eq can handle? How’s the overall tone? Bright enough? Too bright? All of your tools will interact with each other and it takes a lot of practice to learn how each new tool will impact the last. For example I know how hard I can DeEss in the first stage because a lot of the S’s will come back after eq and compression. I know how bright I can get away with making the vocal because the compressor coming up is going to bring back some of the low end etc. Use your ears, take your time, and bypass regularly. If you prefer the bypassed version, start over! We will dig into FX in part two of the vocal mixing article. Stay tuned!