Half Annual Sale: All Plugins $29.99* | Shop Now »

Recording Vocals at Home #5: Dynamics & Compression

Aug 16, 2021

Controlling the dynamics of your vocal recordings is critical. Learn the most straightforward and effective solutions to compressing your voice, making it warm, present, nuanced and human.

By Craig Anderton

Recording Vocals at Home #5: Dynamics & Compression

This is the fifth chapter of the “Recording Vocals at Home” series. Here you can find the full series.

Dynamics control is an important part of mixing vocals, and this begins at the recording stage.

The most fundamental dynamics control is good mic technique—getting closer for more intimate sections, moving further away when singing more forcefully, and discovering the right angle to sing into a mic (to minimize problems like pops and mouth clicks).

Vocalists with a great command of mic technique are relatively rare, so you may need to use electronic dynamics control, like compression. Yet even vocalists with excellent mic technique can benefit from these tools to make an outstanding performance even better.

Before proceeding, let’s give a shoutout to the Motels’ Martha Davis, who generously provided some raw vocals for the audio examples. As with the previous installment, she could do only one take, had to use cheap gear, and couldn’t do anything to improve the vocal. As you’ll hear, the voice that powered a string of ‘80s hits still sounds great—even when trying to make enough mistakes so I can write about how to fix them!

1. Controlling Dynamics: Initial Problem-Solving

Inconsistent vocal levels, whether from poor mic technique, running out of breath at the end of phrases, or finding it more difficult to hit some notes than others, reduce a vocal’s effectiveness. Compressors and limiters can solve these issues. However, a couple of initial, DSP-based fixes within your DAW can prepare your vocal for more effective results with dynamics processors.

Manual Envelope Editing

Some DAWs can create envelopes that alter a clip’s signal levels. An envelope starts as a straight line. Clicking on this line creates nodes. Dragging a node higher or lower changes the level at that point, while dragging left or right alters the node’s position on the timeline. Enough nodes allow for adjusting levels precisely (Fig. 1). Audio examples 1 and 2 play audio before and after applying a level envelope for a more consistent sound.

Figure 1: The level variations in the left waveform have been made more even and brought to an overall higher level by using an envelope within a clip to compensate for gain changes.

Figure 1: The level variations in the left waveform have been made more even and brought to an overall higher level by using an envelope within a clip to compensate for gain changes.

Phrase-by-Phrase Leveling

This also doesn’t replace compression, but like the previous example, it can condition audio so that a compressor performs at its best. The vocal has been split between phrases, and the levels raised to have the same general level (Fig. 2). Of course, you don’t want to remove a part’s natural dynamics, but this technique can create a more consistent vocal performance. Audio examples 3 and 4 play the audio in Fig. 2, before and after levelling.

Figure 2: The red lines on the right indicate where the vocal was split. Note how the levels are more even. Now, if you add compression, it can have a subtler touch because the dynamics are more uniform

Figure 2: The red lines on the right indicate where the vocal was split. Note how the levels are more even. Now, if you add compression, it can have a subtler touch because the dynamics are more uniform.

2. Tame Low-level Artifacts and Breath Inhales

Headphone leakage, mouth sounds, mic handling noise and other low-level noises can be annoying when the vocals aren’t present to mask these sounds. Breath inhales may not seem as objectionable, but if you add compression to vocals (as described later), it will amplify them.

These issues can be dealt with manually—like cutting sections between vocal phrases to delete low-level audio. However, a noise gate can remove or reduce low-level audio automatically. It mutes a track’s audio when the input level consists solely of low-level noise. When the vocal resumes, the audio returns. Several Waves plugins incorporate gating. For example, the Renaissance Vox (described later) includes a gate whose response is optimized for voice.

Renaissance Vox

Renaissance Vox

Breath inhales are a natural part of singing, so don’t remove these entirely. For example, a deep inhale cues the listener that the subsequent vocal section is probably going to be more intense. However, remember that subsequent compression and EQ can bring up breath levels, so you’ll want to keep them under control.

Waves DeBreath (Fig. 3) separates a vocal track into only voice and only breath. You can vary the proportion as needed, as well as process the breath. DeBreath can even add “room tone” (i.e., the sound in a vocal’s background) if breaths are removed instead of just reduced to avoid any interruption in the overall flow.

Figure 3: DeBreath is a sophisticated way to control inhales and breaths

Figure 3: DeBreath is a sophisticated way to control inhales and breaths.

Audio examples 5 and 6 are the before and after clips from Martha’s hit song “Only the Lonely.” The after version eliminates the breath completely, which I wouldn’t do—but this gets the point across that it’s possible to take out every breath, if that’s what you want.

3. How Dynamics Processors Work

Dynamics processors are like invisible hands that alter gain automatically, as needed. Although, in theory, you could do this manually, an electronic device can control gain more precisely and rapidly. This is basically what Waves Vocal Rider does: it adjusts only level to compensate for level changes. It has a very light touch on vocals because it doesn’t narrow a signal’s dynamic range.

In other words, to reduce peak levels, it doesn’t shoehorn a signal with a wide dynamic range into a narrower dynamic range, as compression would. It simply turns down the level temporarily to reduce the peaks.

Figure 4: Waves Vocal Rider isn’t just good for vocals, but for narration and podcasting as well

Figure 4: Waves Vocal Rider isn’t just good for vocals, but for narration and podcasting as well.

A limiter is like a motor’s governor. If the input signal is below a certain level (called the threshold), there are no changes. Above the threshold, the limiter reduces (limits) the gain as much as needed to keep the signal from going over the threshold.

A compressor provides additional ways to control the dynamics after the input signal exceeds a threshold. Instead of just clamping down on the level, traditional compressor designs let you set the rate at which it turns down the level. For example, it could turn down the level progressively more as the input signal goes higher above the threshold.

Fig. 5 shows the waveforms for unprocessed, limited and compressed sounds. Audio examples 7, 8, and 9 play the unprocessed, limited and compressed sounds, respectively. Aside from listening to the overall effect, pay attention to the inhales, and how the compressor brings up their levels compared to the unprocessed and limited examples.

Figure 5: The unprocessed signal on top is the original waveform. The middle, limited waveform has reduced the peaks but left the rest of the audio intact. The lower, compressed waveform has lowered the peaks and brought up the softer sections

Figure 5: The unprocessed signal on top is the original waveform. The middle, limited waveform has reduced the peaks but left the rest of the audio intact. The lower, compressed waveform has lowered the peaks and brought up the softer sections.

4. Compression in Mixing

Now that the vocals have been prepped, start thinking about the mix. A common question is whether to edit EQ or dynamics processors first. There’s no “right” answer. I usually fix glaring EQ issues first, however, adding dynamics processing may offset some of the effect that EQ has on the audio, so I’ll re-tweak the EQ if needed.

The other common question is whether EQ should go before dynamics processors or after. Again, it depends. Because dynamics undoes some of the effects of EQ (i.e., a boost will trigger compression before frequencies that aren’t boosted), you might want to insert EQ last. However, inserting EQ before dynamics can “push” some frequencies into compression sooner than others, or reduce frequencies you don’t want compression to emphasize. Listen, and decide which you prefer. With vocals, I usually insert EQ before dynamics.

Watch this video to learn more about compression before or after EQ.

Is “More” Always Better?

Compression is a popular effect, and there are many available compressor plugins. However, they are often designed to be more general-purpose and feature an array of options. In this piece, we’ll focus on using the CLA-2A and Renaissance Vox, which are two straightforward compression plugins that are particularly well-suited to vocals and get you great results quickly. These plugins may seem simple, but there’s a lot going on under the hood and they are go-to vocal processors for many world-class engineers.

Editing a signal processor’s parameters from scratch to create the sound you want can be daunting. For those starting out, presets provide starting points. Choose a preset that comes closest to the desired sound, and then you may need to modify only a few parameters to obtain the sound you want.

However, here’s a word of warning: if you search the internet for “best compressor settings for vocals,” you’ll find a zillion “pro tips,” and all of them claim to have the answer for the best settings. Which might make you wonder why all those “pro tips” don’t agree on the same settings...

The best settings will depend on the singer, the musical genre, the intended effect, the dynamics processor itself and many other factors. The settings for a smooth pop ballad will likely be different from those for aggressive rock vocals or rap. Sometimes, dynamics control is used not to provide a specific effect, but to compensate for a deficiency in the singer’s delivery.

The only way to come up with the “best” setting for your music is to learn what the various parameters in dynamics processors do and then vary them to hear how they affect vocals. Over time, you’ll learn what settings you need to obtain specific sounds.

5. Choosing a Compressor

Some of the iconic compressors commonly used for vocals are based on vintage hardware and are easy to adjust. The CLA-2A (Fig. 6) is a good example of a “go-to” compressor for vocals.

Figure 6: The CLA-2A is easy to adjust, yet over the decades, it has proven its effectiveness with vocals

Figure 6: The CLA-2A is easy to adjust, yet over the decades, it has proven its effectiveness with vocals.

Adjusting the controls is straightforward:

  1. Listen to the vocal going through the compressor.
  2. Turn up Peak Reduction until you obtain the desired compression effect. Keep it natural at first—you don’t want to squash a sound so that it “pumps” or “breathes” during the vocals.
  3. Adding compression lowers the output because the peaks aren’t as high. Turn up Gain to compensate for this loss in level. Gain doesn’t affect the compression’s character, only the output level.

That’s all there is to it! Bypass the CLA-2A to compare the compressed and unprocessed sounds and adjust the signal level in your DAW’s mixer, so the peak levels are about the same, whether bypassed or not. The compressed sound will probably sound louder and “punch” more. If it instead sounds thinner, there may be too much compression. Reduce the amount of Peak Reduction, then change the Gain control to compensate.

Adjusting the controls is straightforward:

  • Switching Compress to Limit clamps down harder on the signal. It gives a more drastic compression sound to vocals.
  • Turning the HiFreq control counterclockwise reduces compression at lower frequencies, which de-emphasizes the highs somewhat. You’ll probably want to leave it in the Flat (fully clockwise) position.
  • Analog injects subtle noise and hum into the compressor to model the original hardware more accurately. I prefer leaving it off.
  • The VU Display for the meter selects the input signal level, output signal level, or GR—the amount of Gain Reduction that’s occurring at any given moment to create the compression effect. If the Gain Reduction meter spends most of its time toward the meter’s left side, you may be using more compression than is really needed, which can give an unnatural sound.

Renaissance Vox (Fig. 7) is another popular compressor and is optimized specifically for processing vocals. This also makes adjustments easy because it doesn’t need lots of parameters to accommodate different types of audio sources.

Figure 7: Renaissance Vox’s design is optimized for voice, so finding suitable settings is easy

Figure 7: Renaissance Vox’s design is optimized for voice, so finding suitable settings is easy.

Adjusting the controls isn’t much different from the CLA-2A. Moving the center Comp control downward increases the amount of compression, while Gain sets—you guessed it—the output level. The red meter in the Comp control indicates the amount of gain reduction. The orange meter on the left shows the input level, and the one on the right, output level.

To learn about more compressors and find out which one is right for you, check out this article.

6. Analog EQ Flavors

  • To train your ears to recognize subtle amounts of compression, observe the gain reduction metering. Start with a conservative gain reduction setting, where the meter doesn’t go much lower than -6 dB. Live with the setting for a while, and you may decide that’s really all the compression you need.
  • Some engineers smash vocals with huge amounts of compression. That’s not my style, but it’s just as valid.
  • Vocals are a natural target for compression. Compression brings up low-level sounds, so vocals often seem more humanized because you hear mouth noises, breaths and other characteristic sounds that add expressiveness. It’s not uncommon for vocals to use more compression than other instruments.
  • To clamp down on peaks while leaving the rest of the vocal dynamics more or less intact, choose a compressor’s Limit option (if available).
  • Make sure you can hear every word distinctly. Too much compression can “smear” the difference between louder and softer words or even parts of words.
  • Reality checks are important. Toggle the bypass switch frequently to compare the compressed and uncompressed sounds. Match their peak levels closely for the most realistic comparison. Even a little compression may give the desired effect.
  • For transparent compression where you may not even be able to tell any dynamics control is in use, insert two compressors in series, each with very little compression (e.g., only a couple dB of gain reduction). This can smooth out the effect of the compression.

This is the fifth chapter of the “Recording Vocals at Home” series. Here you can read chapter 1 on mic choice and technique.

Musician/author Craig Anderton is an internationally recognized authority on music and technology. He has played on, produced, or mastered over 20 major label recordings and hundreds of tracks, authored 45 books, toured extensively during the 60s, played Carnegie Hall, worked as a studio musician in the 70s, written over a thousand articles, lectured on technology and the arts (in 10 countries, 38 U.S. states, and three languages), and done sound design and consulting work for numerous music industry companies. He is the current President of the MIDI Association. www.craiganderton.org.

Want to get more tips straight to your inbox? Subscribe to our newsletter here.

Loading....