I can remember a very peculiar question that somebody asked me once. They said: "Rob, I read that applying compression to a voice over with a ratio of 20:1 will make it instantly sound more powerful and authoritative?"
I can still remember the disbelief that I felt at that moment as well. I have no idea where they read that—but it's most definitely not true.
Compression is one of those tools that most people like to set and forget. When you are more concerned with content than quality, it can easily happen. Especially when you run a podcast, record voice overs or make your own videos. You're not an audio engineer.
You don't, however, need to spend hours on research and practice to have a basic understanding of compression. Actually, I can do you one better—you don't need to spend much time on research and practice at all.
Perhaps you've heard of Pareto's Principle.
Pareto was an 18th century Italian philosopher who made an interesting and unusual discovery. He found that 80% of his garden peas were produced by only 20% of the pods in his garden.
After some thought he applied this 80/20 concept to other areas and discovered a bit of a trend.
He realised that 80% of the results generally come from 20% of the efforts. This means you can get 80% of the way towards an objective in a fifth of the time if you can only figure out what that 20% is.
You can apply this principle to compressing a voice recording. The key to mastering compression is, of course, practice. I can get you 80% of the way there with a few guidelines.
I'll show you how to improve your voice recordings with a little bit of compression.
The ratio determines how much the audio is reduced in volume by the compressor. The first number tells us by what factor the level will be reduced. So if your ratio is set to 3:1, the output will be 3 times quieter.
The main thing that you need to remember when applying compression to voice is to keep it subtle.
Unless you're a radio DJ or recording a voice over for an advert, you don't want the compression to be noticeable. It should subtly improve the sound of the voice without showing off.
With this in mind, you want to keep the ratio low in most circumstances.
You could go above 5:1 if you were recording a voice for an advert, or wanted to use compression as an obvious effect.
But for a natural sound it's important to keep the ratio below 5:1. I find that even 3:1 is more than enough in most situations.
For this reason, my guideline for ratio would be somewhere between 2:1 and 4:1, with 2.5:1 being a great starting point.
Although all the parameters of a compressor are equally important, getting the threshold right is key.
The threshold determines the level at which the compressor kicks in. If the threshold is too high, the audio won't be compressed at all. If it's too low, the voice will sound overly compressed and completely squashed.
The right setting for the threshold depends entirely on how loud the audio is. You should try to set the threshold 6-10dB below the average volume of the audio.
In the world of digital recording, the sweet spot for the level of your recording is around -18dBFS (with the peaks around -6dBFS). If you are in this sweet spot, a threshold somewhere between -24dB and -28dB would be appropriate.
But as a general rule of thumb, start with -24dB and tweak from there.
The attack time dictates how quickly the compressor kicks in and starts reducing the level of the audio. A medium attack time for an instrument could be anywhere in the region of 10-60ms. But voice is different.
If the attack time is too slow, it starts to make the voice sound unnatural. The compressor will affect the end of the words, but not the beginning—because it's too slow. You want the compressor to affect the whole word, not just the ending.
On the other hand, if the attack time is too fast the voice will sound over-compressed and squashed.
A good guideline for attack time on voice is 1-5ms. If you go too far outside of this zone you might have issues. Within that range, I find that 2ms is a great starting point.
Release time is how long it takes the compressor to dis-engage and bring the audio up to its actual level. As with the attack time, setting the release too fast or too slow will result in an unnatural sound.
A release time that is too slow often leads to a pumping effect where the compressor doesn't actually release until the following word. A release time that is too fast will cause the voice to sound wild and uncontrolled.
Try to keep it between 10-15ms for release time. A good starting point is 12ms.
After all this compression and gain reduction, the audio coming out of the compressor is of course a bit quieter. To make up for this, you can adjust the output gain of the compressor—sometimes it is referred to as makeup gain or just gain.
Use the level meters on the compressor to match the output level to the input level. Start with 5dB and adjust from there.
That's all the main settings covered. Let me give you some general advice for applying compression to a voice recording.
- If the compressor has a knee setting, turn this up to make the compression more subtle
- Avoid de-esser's. This is a form of compression that reduces that harshness of sibilant S sounds in the voice. These are great for sung vocals, but sound unnatural on spoken word recordings
- In most situations you should apply the EQ cuts before the compressor and the EQ boosts after the compressor
- If you want to heavily compress a voice recording without making it sound too unnatural, use two or three compressors in series. This will sound better than using drastic settings on a single compressor
- If you can't hear the difference between different attack and release times, bring the threshold down really low. This will fully engage the compressor and you will be able to clearly hear any small changes that you make.
In this video you can watch me apply compression to a voice recording step-by-step. I'll also show you how changes to these parameters affect the sound of the voice.
The guidelines in this tutorial will give you a great starting point. You can't go too wrong if you stay within these suggested boundaries. But don't forget that rules are meant to be bent—and even broken. The most important thing is to trust your ears. If it sounds good, it is good.
Before you use compression consider how you want the recording to sound. After you've applied your compressor and tweaked the settings, consider if this suits the character of the voice or if it sounds unnatural.
Perhaps you don't want the voice to sound natural, so you can use compression as an effect to create drama and power instead.
Leave a comment below and tell me about your experiences with compression.