Vocals can make or break a song. I’m sure you can recall hearing the intro to a tune and thinking, “Oh yeah, I like this,” but the first couple vocal phrases made you go, “Ugh, this is tragic. Now, I hate everything.”
There are many significant junctures capable of enhancing or undermining the quality of a song’s vocals. Let’s look at some common vocal production practices—some to use and some to avoid. Of course, it’s best to start at the source.
Audio engineers sometimes focus so much on their gear and techniques that they forget to pay attention to the artist. The best equipment in the world doesn’t mean much if the singer isn’t able to give a good performance. Amazing mics, preamps, and converters won’t make a bad performance great; they’ll simply produce a sonically respectable recording of a performance people don’t want to hear.
The DO list:
1. Do make sure the artist is comfortable. Consider temperature, lighting, and the general vibe of the recording space. Even the room size and materials matter. A small dry room has a very different feel and sound than a large reverberant one.
2. Do have some bottled water and lemon handy. All that singing makes for a dry mouth and a dry mouth leads to early fatigue and an increase in extraneous mouth noises—those annoying little clicks that plague so many vocal recordings.
3. Do have pencils, pens, and legal pads. Artists have ideas and need a convenient way to put them down on paper.
4. Do communicate. Talk to the artist, figure out how he/she is feeling, and keep him/her informed. If you need to put the singer on hold to change a microphone or cable, say something. Don’t just leave the poor soul hanging, wondering what’s going on and for how long it’s going to go on.
The DON’T list:
5. Don’t ignore the artist. Imagine going to someone’s house and being treated as if you’re either a nuisance or not even there. Not cool!
6. Don’t make the talent wait unnecessarily. If you and the client agree to start a session at noon, don’t in at 12:30 PM. You do not want the artist thinking, “What was the point of me showing up at noon?” An annoyed singer is less inclined to pour out a solid performance when trying to suppress the early signs of rage.
Microphone choice and placement
There are many variables in this territory and there isn’t one way that’s always right. However, there are some general tips and considerations worth noting.
The DO list:
7. Do pay attention to the sound of the room when setting up the vocal mic. Use your voice to make some noise in different places around the room. If you happen to hear flutter echo or odd room resonance, avoid those spots.
8. Do have a pop filter. Not only does a pop filter reduce strong plosives from “p” sounds, it also serves as a target for the singer. If you want to keep the singer from getting too close to the mic (to avoid proximity effect), place the pop filter a few inches away from the mic. I guarantee you that the singer won’t bust through that barrier.
9. Do select a polar pattern that suits the desired sound. Many condenser mics offer multiple polar patterns, which have a significant impact on the amount of indirect sound picked up by the microphone. For example, omnidirectional picks up sound from equally from all directions, while cardioid rejects a significant level of sound from behind the microphone. Between those two, cardioid would yield a more close, intimate sound, whereas omnidirectional would give you more room ambience.
The DON’T list:
10. Don’t pick a mic based upon price. Let’s say that you’re going to record a female vocalist who has a lot of sibilance and you pick the most expensive mic available, which happens to have a sizable high-end boost. It’s likely that it would exaggerate the already prominent sibilance. Maybe there is a cheaper mic—a ribbon or a dynamic for example—with a mellow top end that would impart a warmer, less sibilant tone. Check out the frequency response charts of different microphones to give you an idea of how each mic would “EQ” the sound.
11. Don’t forget to adjust the mic position appropriately for the singer. This is easy to overlook. You might place the mic before the singer arrives, focus on setting up your DAW session, then forget to check the mic placement when the artist shows up. If the artist is a little taller than you expected, and the mic ends up a few inches below his or her mouth, you’ll end up with a noticeably different frequency response. Once the singer is comfortable and in the desired spot, adjust the vertical position of the mic so the mic’s capsule is level with the singer’s mouth (a good starting point, at least).
12. Don’t forget to check the microphone’s pad and roll-off settings. Overlook this and you might be recording with a roll-off that thins out a vocal when it should be full. For example, the classic and widely used Neumann U87 has a roll-off, for which Neumann doesn’t list the frequency specification. One look at its frequency response diagram shows that the roll-off starts at around 500 Hz! Look out!
The recording signal chain
Preamps, compressors, EQs, converters, and clocks—this is where audio engineers tend to geek out and freak out. Sometimes it’s for good reason, other times it is misdirected anxiety.
The DO list:
13. Do document your signal chain. If you ever need to re-record a section at a later point in time and want to get the same sound, you need to know how you got the tone in the first place. Write down not only what gear you use, but also the settings on the gear. For example, document the mic, mic settings, mic position, cable, which preamp you use, the amount of gain, which compressor, the compressor ratio/attack/release/threshold, converter channel, sample rate, clock settings, etc.
14. Do use equipment that you know. You want to get results efficiently and predictably, so use gear that you have verified operates properly with no unintended distortion or crackling.
15. Do listen objectively. Sure, it’s good to check your meters, but listen closely to the vocal sound. Listen to it in solo and with the accompanying music. Make sure it works (sonically) in both cases.
The DON’T list:
16. Don’t record too hot. It’s especially common when referencing loud accompanying music to record the vocals hotter. Only pain and regret can come from doing that. It’s better to be conservative with your vocal recording level—somewhere in the range of -20 dBFS to -10 dBFS for example—and just turn down the reference music if it is overpowering the vocal.
17. Don’t overcompress. You may have a compressor such as an 1176, LA-2A, or Distressor that you absolutely adore for vocals; it’s great to have something you love. However, if you record with it, err on the side of caution. Too much compression can result in distortion and a loss of important dynamics. Keep an eye on the gain reduction!
18. Don’t waste the artist’s time needlessly experimenting. Imagine the following scenario: You want to get the best possible vocal tone with the equipment available. So, you have the artist sing, sing, sing, and sing some more as you try different preamps, compressors, compression ratios, EQ bands and modes. 30 minutes fly by and the singer has exhausted a valuable portion of his or her stamina. The singer will fatigue sooner and I’m willing to bet that he or she will be in a less inspired or creative mood. So, don’t do that.
The style in which you record the vocals massively affects the productivity of the session. What should you do: record whole takes from the song start to end? Record all the verses and then the choruses? Record line by line? Just wing it?
The DO list:
19. Do familiarize yourself with the arrangement. When the artist inevitably asks to jump to the next chorus, you don’t want to be fumbling around the session trying to find it like its your first time hearing the session.
20. Do discuss the approach with the singer. If the client has some recording experience, he or she may already have settled into a method that’s comfortable. A simple question like, “Would you like to take it from the top or knock out the verses or choruses first” can prevent you from starting down the wrong path.
21. Do record multiple takes. Having to record multiple takes or attempts is expected. It allows you to amass performance variations and piece together the best bits later. So, when recording multiple takes, please keep track of the take numbers. Five takes in, the artist might say, “What did you think of the second take? Let me hear it real quick." You need to be clued in and able to access the right take quickly. Have a pencil and pad handy to jot down take numbers, file numbers, and performance notes. Also, try to keep your recording chain consistent between takes. It’ll make it so much easier to piece the best parts together.
The DON’T list:
22. Don’t stop the singer’s workflow to edit. Even if you are obsessed with cuts and crossfades, don’t make the singer wait in front of the mic while you squint at your screen and tweak fade slopes (unless it’s requested or necessary). That stuff is low priority compared to recording the raw vocals.
23. Don’t force the artist into an unnatural workflow. Just because recording line by line worked great for the last vocalist you tracked doesn’t mean that’s how you should do it this time. Something that is standard practice for you could be completely foreign for someone else. Be willing to change your routine.
24. Don’t record destructively. Some DAWs allow you to destructively record over existing audio—just like it was with analog tape! Although the thought of erasing a bad take by recording a new take over it might seem attractive to those looking to avoid wasting drive space, the danger is obvious. There’s no undo for destructive recording! Nightmare scenario: The singer says, “Nah, that was trash” after a take, so you destructively record over it in the next take. Then the singer says, “Actually, that one was trash, but I think the last one was pretty good in a rough and raw way. Let’s use that.” Oops, no can do!
25. Don’t undo recordings unless they are confirmed worthless. In the spirit of keeping your session tidy, the temptation to undo a “bad” take may be great. However, it’s not uncommon to get ten takes in only to realize that the first two takes were the best. So, even if the client says, “Oh man, I can totally do that better,” don’t undo it—keep it!
Don’t act surprised when the vocalist wants to record layers or stacks, especially for choruses, hooks, and background parts. It’s a common technique that can yield a thickened sound or emulate a group of singers. How do you go about this?
The DO list:
26. Do reference the previous part. Typically, after recording the first layer of the stack, the singer will benefit from hearing it while tracking the second and other layers. It provides a timing and pitch reference to try to match. Note that there are situations in which the singer will not want this; accommodate the artist’s request.
27. Do create tonal variations between layers. This is so important for successfully making one singer sound like multiple voices. Consider varying the singer’s distance from the mic, microphone polar pattern, microphone, preamp, compressor, and EQ between layers. Otherwise, prepare for a possibly problematic build-up of similar frequencies. For example, if 1.2 kHz was already prominent in the first vocal layer, it will be obnoxious after a six-part stack with no tonal variation between the layers.
28. Do create positional variations between layers. In a real-life group of multiple singers, the vocalists aren’t sharing the same exact physical spot; they stand in different places. Recreate that reality in your DAW with panning. Pan them parts apart!
The DON’T list:
29. Don’t duplicate tracks to make stacks. A common amateur mistake is recording a vocal track, then duplicating it multiple times to create a stack. All this accomplishes is increased amplitude. It in no way creates the effect of multiple voices singing the same part.
30. Don’t rule out using other singers. If tracking stacks with the same singer isn’t producing the desired effect, it may be time to use another singer to lay down a few layers for the stacks.
31. Don’t assume that more layers universally equates to a bigger sound. There is a point of diminishing returns where having too many layers gives you a smeared, muddy tone. Plus, if the timing is between each track in the stack isn’t tight, it can be extremely distracting.
Comping and editing
Whether you recorded one take, multiple passes, or stacks and stacks, you’ve got some editing ahead of you. It’s easy to get lost in the details whilst peering at your audio under the microscope.
The DO list:
32. Do the comping first. Comping is the process of picking the best parts of each take and piecing them together to create one really good take. Rather than using the entire take 1 or take 2 or take 3, you can use the intro from take 2, verse 1 from take 3, chorus 1 from take 1, etc, etc, etc. Assuming that you recorded multiple takes for a vocal, do the comping before you do any tuning or automation; you don’t want to waste time tuning and automating vocal parts that won’t be used.
33. Do tighten up the timing. Singers will occasionally rush or drag certain words or phrases unintentionally. On just one vocal, this is a bit off-putting. On stacked vocals, it’s annoyingly distracting. Go through the vocal tracks and nudge, slide, or quantize as necessary. Remember, the mission is not to align the vocals perfectly to a tempo-based grid. The goal is to get the right feel or groove for the song.
34. Do “clean up” the vocals. There will always be strange little sounds that don’t contribute in a positive manner—things like mouth clicks, humming, and rustling between phrases, and headphone bleed are common culprits.
The DON’T list:
35. Don’t always edit out the breaths. Humans breathe, yes, even when in front of a microphone. Between vocal phrases, you’ll hear breaths. It’s normal to have those unless they sound unnaturally loud or awkwardly interrupt the flow of the vocal. Though you may need to turn down loud breaths and edit out awkward ones, only treat the problems (7-time GRAMMY-winning audio engineer Frank Filipetti would agree). Cutting out all breaths is typically a misuse of time in music production.
36. Don’t throw away unused portions. After assembling the best parts of multiple takes, it’s tempting to throw away the bits you didn’t use to keep things clean and orderly. However, you never know when you or the artist may have a change of heart about a word or line. Plus, those extra parts allow fairly easy creation of doubles and stacks.
37. Don’t listen only in solo. If you only listen to a vocal in solo while deciding what elements of it to edit, you’ll end up editing things unnecessarily. You’ll hear every tiny flaw of the vocal because it’s the only thing you can hear. A more time-efficient approach is to listen to the vocal and music together, then stop when you hear an issue such as a loud mouth click. At that point, you can solo the vocal to do the edit, but ultimately check it with the music playing.
You might be opposed to vocal tuning (pitch correction), but the standards of modern music production frequently demand it. So, leave your audio morals at the door.
The DO list:
38. Do check the vocal pitch against the music. Let’s imagine that a guitarist and bassist tune to each, then record the instrumental parts for a song. They’re in tune with each other, but not with any particular pitch source. They don’t realize it, but their instruments are actually flat from the standard A440 reference. Later, they record vocals to the music and need some vocal tuning to fix the singer’s occasional pitch issues. They slap a pitch correction plug-in on the vocal track to pull the vocal on pitch. They overlook the fact that the plug-in tunes according to the A440 reference, which their instruments are flat from. So, when the plug-in shifts the vocal to be perfectly on pitch, it will be sharp of the instruments. Yikes! Be sure to listen to the vocal and the music when executing pitch correction; you’ll be able to hear if the vocal matches the pitch of the instruments it will be mixed with.
39. Do listen for artifacts. The more manipulation you do with pitch correction plug-ins, the more likely you are to introduce sonic artifacts. They can sound like robotic glitches, grainy texture, distortion, or drop-outs. If you do hear such things, try changing the parameters on your tuning plug-in or do two passes of more gentle tuning for those parts.
40. Do put the tuning plug-in (such as Nectar) before your mix processing. You don’t want to have the pitch correction impacted by other processors! Picture this: You already have an EQ, delay, and reverb directly on the vocal track, so you insert a pitch correction plug-in after the reverb. Bad move! The pitch plug-in will try to pitch shift the delays and reverb. That’s chaos, I tell you. However, if you place the pitch plug-in before the other plug-ins, the pitch-corrected vocal will be fed into the other processors, which will behave normally.
The DON’T list:
41. Don’t tune before comping. If you’ve amassed ten takes of a lead vocal, why tune all ten takes if you’re only going to use some of each one? Comp the takes to create the best version, then tune that. You’ll spend far less time tuning and will retain your sanity for much longer.
42. Don’t always tune everything. Not only is tuning every vocal line frittering away valuable time, it’s increasing the likelihood of unwanted sonic artifacts caused by the tuning process. Pitch correcting every word will undoubtedly result in sorrow at some stage of your life.
43. Don’t tune by sight. Most pitch correction plug-ins have a visual interface that displays the source pitch. Helpful information? Yes. Too much information? Sometimes. When people see that the source note is sharp or flat from the perfect pitch, they tend to conclude that it needs to be corrected. A certain amount of pitch variation is natural and often necessary for a vocal to sound right with the accompanying music.
Deciding how to deal with the vocal levels is simple in theory, but complicated in practice. Some considerations may be obvious or only become obvious when stated.
The DO list:
44. Do use your meters. Most audio software programs and hardware units have meters designed to keep you informed of your signal levels. Utilize them to ensure that your levels aren’t too low or high.
45. Do use compression (with care). If properly used, a compressor can tame jolting jumps in level and output a more stable vocal signal. It’s typically much easier and faster to use a compressor than it is to manually adjust the levels of all the various words and phrases. Consider using one or two compressors to control a vocal.
46. Do use automation. Although compressors help stabilize level changes incurred by a vocalist’s performance, they don’t adjust the vocal level according to the music, arrangement, or emotion of the song. A compressor won’t know when the music gets quiet or loud and when the vocal needs to be accordingly quiet or loud. Using fader automation, you can raise or lower the vocal level to match the desired emotion and impact in relation to the music that you hear.
The DON’T list:
47. Don’t clip. Clipping can occur at the microphone, preamp, outboard gear, converters, plug-ins, faders, buses, and aux sends. Be careful with your gain settings and fader positions, check meters at various points in your signal path, and always keep an ear open for unintended distortion.
48. Don’t over-compress. Too much compression can cause distortion, strange attack and sustain relationships, and pumping. Over-compression might make the vocal meters extremely stable, but that doesn’t mean it will sound stable with the music.
49. Don’t bury the lead vocal. Musicians often follow the unconscious tendency to mix their instrument the loudest. Guitarists want the guitars loud, drummers want the drums loud, and zitherists want the zither loud. If you’re a musician, don’t let your preferred instrument bury the vocals. How’re the average listeners supposed to sing along if they can’t hear the vocal clearly?
EQ and effects
Your use of EQ and effects can make your vocals sweet and blend beautifully with the music or stick out like two sore thumbs.
The DO list:
50. Do use different EQ on different tracks in a stack. As mentioned previously, tonal variations help avoid frequency build-up in vocal stacks. In the likely event that you end up mixing tracks that someone else created, you may encounter vocal stacks that were recorded without varying the mic, position, or preamp. Use different EQs and different EQ settings on each track of the stack to induce differences in frequencies and harmonics.
51. Do use time-based effects to create depth differences. When layers of vocals are recorded with the same microphone in the same room at the same distance, the vocals will probably lack depth differentiation. Use effects such as delay, reverb, and chorus to add unique space to vocals. Bear in mind that dry vocals tend to sound closer and vocals with reverb tend to sound further back.
52. Do explore non-traditional effects. Desperate times call for desperate measures; abnormal effects such as vocoders, ring modulators, bit crushers, and multi-voice pitch shifters can empower you with enough unique textures to make vocal parts stand out. Such effects are best used for accent rather than being applied to an entire track.
The DON’T list:
53. Don’t always use the same processing on every vocal. In the spirit of saving time, some people throw the same processing chain on each vocal and recall the same settings they always use. Though there are times when that works well, the danger of doing so is a lack of uniqueness—ignoring the exclusive needs of the song. Sure, 65% of the time, your go-to vocal chain works all the time, but that doesn’t mean it’s right for every song.
54. Don’t blindly rely on presets. Yes, it is much faster to load a preset than to configure a plug-in from scratch. Yes, presets can help you learn about proper use of a plug-in’s parameters. No, presets are not a substitute for subjective and creative adjustments.
55. Don’t make time-based effects overly bright. Time-based effects such as delay, reverb, and chorus easily become distracting when they have a generous portion of high frequencies. Usually, you want your effects to be less present than the signal that triggers them. For example, don’t make your vocal reverb brighter than the dry vocal.
That’s a lot to keep track of; do this, don’t do that, sometimes yes, but sometimes no. Also, you cannot forget to adjust your methods to suit your client. Be flexible, communicate, and practice, practice, practice!