When making a vocal chain, understanding your routing will affect how you process your vocal. For example, I always start with subtractive forms of processing to remove frequencies I don’t want later on, since later forms of processing will most likely amplify these aspects if they aren’t corrected first.
For this video, I’m going to show you the setup I use for personal projects, as well as explain why I use each processor so that you can make some informed decisions when creating your chain.
The first thing to understand is chain routing - in short, each processor is going to lead into the next in a top-down manner - so we’ll need to consider how we’re feeding our signal into each subsequent processor. After all of our inserts, are signal is sent to the pan-pot.
But we probably won’t use panning on our lead - it’s then routed to any sends we’ve created, and the channel fader simultaneously. Although, this send can be routed as post pan which is the default and what I usually stick with, pre-fader, or post-fader.
For the time being, we’ll keep the output as the stereo output, but I’ll change this later in the video and explain why. So let’s listen to the original vocal and the full vocal chain to get a better understanding of where we’re headed.
The first processor can be difficult to pick, but a subtractive EQ is a great choice - I also use this first when mastering to attenuate aspects of the frequency response that I don’t want. Since
some of my processing will be additive, I really want to avoid amplifying unwanted frequencies.
With that in mind, I’m going to attenuate the lows with a high-pass, and then dip some of my fundamental and second ordered harmonic but only by about 1dB - just to subtly increase clarity. Of course, what you attenuate is up to you and depends on the vocal, but this is a good option.
Let’s listen and keep in mind it’ll be subtle since this is the first processor.
Next, I want to attenuate my vocal’s sibilance, since I definitely don’t want this to get amplified by subsequent processing. A de-esser or a multiband compressor can be used for this, but I prefer a de-esser, since it uses less CPU, even if it gives me less control.
A split band setting will ensure that only the highs get attenuated, not the level of the entire signal. Let’s listen and notice how the highs become controlled.
Additive compression may seem oxymoronic, but, the purpose of compression is to ultimately add detail to our signal. By reducing peaks, we reduce the dynamic range or distance from our highest and lowest levels of signal - we can then push our overall signal, quieter parts and all, higher.
As a result, our vocal’s quietest details are made easier to hear. For vocals, a quick attack, quick release, soft-knee, a little lookahead, and auto-make-up gain will work well for this.
I like to follow downward compression with upward compression, which captures, compresses, and amplifies low-level signals - in turn amplifying the details of the vocal even more.
Let’s listen and notice how much detail we can hear now.
The last step on my channel strip is going to be some saturation or analog emulation - these processors do a lot. First, they slightly compress, but more importantly add harmonics that fill out the vocal, all while adding program-dependent processing, adding some needed complexity to the vocal.
For example, so saturators will compress the lows more than the higher or distort frequencies differently depending on which frequencies are present - resulting in a more nuanced vocal.
Let’s listen with all of this in mind.
With the processing on my channel strip done, I’m going to set up a send via a bus - on which I’ll add my temporal or time-based processing, but we’ll cover this next chapter. Additionally, I’m going to change the output of my channel strip to a new bus.
I’ll change the output of my send that includes my temporal processing to this vocal bus as well. So to recap, the signal flows through the channel strip, then gets split - one part goes to my temporal processing bus, which then gets sent to the vocal bus.
The other part goes directly to my vocal bus. This seems convoluted but serves a really important purpose that I’ll cover in the next few chapters.
On the temporal send, I’m going to insert a quick delay, and then a reverb. The delay is going to thicken the vocal and make it sound more impressive - then the reverb is going to do the same, but also give it a stylistic effect that’ll give it some identity.
I use temporal processing on a send instead of my channel strip since it lets me shape just the temporal processing if needed. For example, I could insert a high pass filter before the delay to ensure only the highs are delayed and reverberated - or maybe add a compressor after the reverb to really accentuate the details of the processing.
Let’s listen to what this temporal processing adds to the signal, and notice that I’m going to include both the EQ and compression I described to the signal.
On my vocal bus, which contains both the original signal and my temporal processed signal, I’m going to add one last compressor and very subtly compress. I like to use optical compression here and only achieve about a dB to 2, just to smooth out the sound and make it cohesive.
If you have additional BGVs you can send them to this vocal bus as well, and the compression will help blend the lead and other tracks.
Let’s listen to how this compressor subtly ties everything together.
The last processor I’m going to add to my chain is an additive EQ, with which I’ll amplify anything about the vocal that I want more of. This EQ also goes on the vocal bus, meaning it’s going to affect everything that came before it, the temporal processing included.
I’ll cut the lows with a high pass again, which may have been introduced by some of our processing. Then I’ll subtly boost some of 4kHz to add clarity and above 12kHz to add air.
Let’s listen to how this EQ subtly shapes the full signal.
When vocal tuning is needed, I find that the best place for it is either as the first insert or after my subtractive EQ. The reason is, additional processing is going to make it harder for the processor to measure and affect the pitch of the signal.
Let’s listen to this same chain, but with some tuning at the beginning, and notice how it only makes a subtle difference, which is what we want in most cases.