When stacking vocals start with a good lead vocal comp, then add unison vocals, harmonies, and full high and low octave unison vocals to fill out the frequency spectrum. Then manually edit their positions to align the performances, and use de-essing, compression, tuning, and reverb.
In this video, we’ll cover some good practices when recording vocals, how to edit multiple vocal tracks, and then the mixing side of things.
Before we begin to layer vocals, we need to ensure that we have a reliable lead, since this will be the foundation for everything we add. For my projects, I’ll record about 10 lead vocal tracks, and then listen carefully, and select the best parts to create a vocal comp.
This comp will be my lead. So let’s listen to a raw vocal take, and then a comp and notice how the comp gives us a better foundation for layering vocals.
Next, I’m going to record 2 to 4 unison vocals on top of it, in other words, vocals that are as similar to the lead as possible. The main difference will be the sibilance and transients, which I’ll want to be a little softer to help these vocals blend in.
If the transients of the vocals aren’t perfectly aligned, we’ll easily hear timing differences, so having the vocalist hold back on these will help a lot. Due to small timing differences, we can simply pan these hard left and right and mild phase cancellation will handle the majority of our stereo imaging.
Let’s listen to these unison vocals included.
Just like before, we’re going to record in 2 to 4 harmony vocals to fill out the frequency spectrum and add some complexity to the vocal. I’ll want these to be even subtler than the unison vocals, so I’ll reduce their level to blend them in.
There’s no exact answer as to what to record - here you’ll have to be creative and try some different harmonies out. Again, I’ll pan them, but maybe not full left and right to help differentiate them a little from the unison vocals.
Let’s listen to these added in.
Right now we should have somewhere between 5 to 9 vocal tracks, but let’s add 2 more - a low and high octave vocal that’s nearly identical to the lead in everything except pitch. We could also simply duplicate the lead and shift it down and up an octave.
I’ll record these octaves instead to create a more natural sound, but how you handle this is up to you.
Let’s listen to how these fill out the frequency spectrum even more.
When editing we can use time alignment or flex pitch functions in our DAW, but to my ear this adds a lot of unwanted artifacts, so I typically choose to align them manually. To speed up this process, find the cut tool in your DAW and isolate each aspect first.
Then go back and start realigning each segment. Once this is done, listen closely and start adding crossfades when needed.
While we’re editing, we can also find sibilance either by listening, or zooming into the waveform and noticing where we have higher frequencies - which will be represented as dense clusters or waveforms. With clip gain, we can reduce these, or just use a de-esser later on.
As we covered last chapter and in chapter 2, we’ll want to control sibilance and aggressive transients, so let’s use a de-esser to attenuate these aspects in all BGVs. In addition to helping with timing, more sibilance than what’s present in the lead really isn’t needed.
Let’s listen to our BGVs with and without de-essing and notice how it improves the quality and the timing. Also, I’m going to include processing to my lead to make the demonstration a little more indicative of a real session.
Another thing that is going to affect the timing and cohesion of our BGVs is the dynamics - for BGVs we really don’t need them to be dynamic. So let’s compress them heavily and with smooth settings like optical emulation, soft-knee, and a longer release.
Let’s listen and notice how the BGVs become denser, more cohesive, and blend with the lead much better.
Similar to using heavy compression, we can add aggressive tuning and chorusing without the 2 effects seeming too aggressive to the listener. I’ll use the Slate Metatune to accomplish both of these tasks, but feel free to use whatever processors you want for this stage.
Tuning the unison vocals will be especially helpful in making the lead sound more in tune. Let’s listen to our BGVs with this added.
If we want our unison vocals and harmonies to blend in more, let’s use medium-length reverb, and isolate the reflections to the mid frequencies. Then I’ll reduce the pre-delay to reverberate the full signal, further reducing transients, and add some modulation somewhat similar to a chorus effect.
Let’s check out how this reverb further blends the vocals in, and let me know if you prefer to use more drastic reverb effects on your BGVs.
For this last chapter, I’ve created a vocal bus for my lead which includes some additional processing like mild compression and some EQ. I’m also going to route all of my BGVs through this vocal bus to show how combining the lead and BGVs helps make them cohesive.
This is especially true of compression - when we send our BGVs and lead to the same compressor, we affect their timing in the same manner due to uniform attenuation of all signals.
Let’s listen to this final step and notice how the very distinct lead that we started with now blends seamlessly into multiple and complex additional vocals.
Then we’ll listen to the full A B with no processing, and then all the processing we’ve covered.