Welcome to Dave's Vocal Processing Guide.
The general goal with vocals is you want them even and clear, free of unnecessary artifacts like mouth noises (clicks, tongue slaps, lip smacking etc), and present and bright, without being strident. Vocals are very much a garbage-in garbage-out thing, quite likely the most so out of any musical element. A lot of dark magic is possible, but at the end of the day a better vocal take is always going to outweigh any processing.
A typical individual vocal channel's plug-in chain will look like this: Noise gate -> EQ -> Compression. This is the basic bread and butter processing, but it is also the most important.
A noise gate's utility is largely in cutting out extraneous noises and increasing the dynamic range of a vocal track. Even in quiet rooms where there is not much ambient noise, noise gates can still improve the quality of a track by creating a clearer signal for the remaining processors.
My settings for this are typically going to be using an expander (I don't actually use a straight up brickwall/infinite ratio gate most of the time):
- Ratio: Anywhere from 1.5:1 to 4:1, no knee.
- Range: Anywhere from 6 dB to 60 dB. This is a huge range and depends largely on how much noise I need removed. Start with 10 dB and see what that does for you.
- Attack time: Fastest possible, so 0.1 ms, 1 ms, 5 ms, 10 ms, whatever.
- Release time: Varies with the program material, but 100 ms is a good start. If it cuts off too tight, increase this. If it doesn't cut off fast enough, decrease this.
- Hold time: ~50 ms, which is usually enough to prevent chopping. If you're processing speech, you may need to increase this slightly.
All vocal EQ is subjective and there is no one-size-fits-all solution to it. That said, the most important EQ moves you will make with a vocal are these, in this order:
- Low pass filter: Just about all vocals are going to be in the range of E2 (~82 Hz) to E5 (~659 Hz), so you can almost always safely set a filter to cut anything below 80 Hz, and as high as 200 Hz sometimes. 12-18 dB slope is usually fine here, though you can also set a much higher point and use a 6 dB slope.
- High shelf: Most vocals like a 3-4 dB boost around 5 kHz. A 12 dB shelf is good for this; if you need more brightness, try increasing the slope before increasing the boost.
- Low-mid peak: For male vocals, a wide boost around 150-200 Hz can help if it is necessary. Q < 0.7 or so, 3-4 dB/to taste. For female vocals, the same but starting around 300 Hz.
- Mid cut: Sometimes scooping out a little bit around 600-700 Hz can clear things up if needed.
- High-mid peak: Sometimes adding a little poke around 1-2 kHz can brighten things up if needed.
The compressor is IMO the most important part of the vocal chain, so let's get right to it:
- Ratio: I will do one of two things here: 3:1 with no knee, or 100:1 (brickwall) with the knee turned all the way up. 3:1 is more precise and is more suitable for live vocals, 100:1 is generally more pleasant for studio vocals, and it is more "predictable" in that making moves "feels" like it works more how you expect it to.
- Threshold: This value will depend on the program material, and the value on the dial is not important. What is important is you set this such that you are getting between 10-20 dB gain reduction. I usually shoot for about 13 dB GR on the low side.
- Attack time: 5-15 ms. Faster attack will equal smoother vocals.
- Release time: ~175 ms. This appears to be a magic number and your results can vary. Increasing this makes the compression feel more obvious, decreasing this makes it feel more transparent.
- Makeup gain: Whatever is necessary to get the level back to where you need it. You can use the minimum gain reduction value as a good starting point, or set this to auto. I generally notice compressors will not give me enough level back when I use auto makeup gain on vocals, so I typically will either manually set this or add a few dB of gain on the output if my plug-in allows it.
I will usually set up 3 effects busses and I will run every vocal to them.
Typically a fairly rich hall reverb with a long decay (~2-3 seconds, maybe longer). This depends on the material. Always, always, always put a high pass filter at around 100 Hz after this. Usually set to -25 to -20 dB on the send, or to taste. Enough to make it feel rich without sounding like you're in a cave.
This will usually be synced to a 1/4 note. I will usually set the stereo channels to be slightly slightly offset (~20 ms or so) from each other. Sometimes I will do a straight up ping-pong configuration, in which case I'll set the time to be a 1/4 note and then set the second delay (the "pong") to be 150%, which makes it a dotted 1/4. I like a good bit of feedback with this (40% is my magic number). I'll usually also set up some EQ to lop off the top and bottom of the delay, so you only hear from like 500 Hz to 3 kHz. Usually set to -30 to -25 dB on the send. I usually like to set this in combination with the reverb so I hear the first 2-3 repeats, then have the reverb take over.
This is my secret sauce. I used to use a combination of a very short delay (~20-25 ms) with a moderate amount of feedback (40-50%), a chorus/doubler effect set to a short delay (~4-10 ms) and somewhat slow LFO (0.7-0.8 Hz), and a stereo image expander (150% width), or the Waves Doubler4 effect. Now I just do this with a delay that does all of it with modulation and ping pong (FabFilter Timeless 3). Usually set to -20 dB or so on the send, but this is a less-is-more thing. You want to set this so the vocals just "lift" out of the mix. Too little and the effect does nothing, too much and it sounds robotic.
On a vocal mix bus, I usually will have a de-esser and a saturation processor if I have anything at all. This section otherwise intentionally left blank.
There is no easy way to learn how to tune vocals. You must learn the tuning program, then try to apply it on every vocal that needs it. The only way out is through.
My typical recording flow for a vocal is to record a "main" vocal, then record a doubled vocal in choruses. I will do essentially identical processing for each vocal, but set the double anywhere from 0 to 10 dB lower than the main vocal. The more of the double you have, the more like Dave Grohl you will sound. If I record a single harmony, I am almost always going to record doubles of it and then pan them 50% to each side, then set them about 8 dB lower than the lead vocal. If I record more harmonies, I will do the same but increase the pans by 10-25% each time. Given a sufficiently tight performance, plus all aforementioned processing, this will get you a wide, full, lush vocal mix that you can tape to your refrigerator.
- My damn self
- FabFilter Beginner's Guide to Music Production playlist: https://www.youtube.com/playlist?list=PL6e8wfdmIuLGALV-6x3arKIK2Hw5Mjlxx
- https://muted.io/note-frequencies/