Decreasing Dynamics - An Honest Guide for The Musician (1/2)

Decreasing Dynamics - An Honest Guide for The Musician (1/2)

Introduction

Hello, everyone, I’m Alex. If you’ve seen my previous two-part post, I discussed ‘transparent’ vs ‘character-driven’ limiters. You may notice I did not provide the actual definition of a limiter. If I did that, the two-part post would have been even longer. However, this next two-part post will be a deep dive for musicians, composers, and performers who want to understand dynamics processing in depth.

Understandably, this can be an intimidating subject for musicians and even misleading for mixers at first, but I can’t stress how important it is to tear apart the different parameters in compressors, limiters, clippers, etc. Other posts may touch upon their functions briefly in just a few sentences, and most may be comfortable with that. However, many struggle with compressors especially, because they do not understand the attack, release and threshold on a compressor. Let’s tackle dynamics processing once and for all, and get you better mixes and masters!

What is dynamics processing?

“Dynamics processing” is any process that affects the dynamics of your music. Common processes that you may see mentioned are compression, limiting, expansion, and noise gates. There are other processes, such as clipping, which aren’t mentioned as much as they are much more in a ‘grey area’, but I think it goes in hand with the other processes. Noise gates and expanders will be discussed in separate future posts.

Dynamics processing is a major part of mixing, and is one of three major parts of mixing: Filters (equalisation), dynamics processing and “effects” (reverb, delay, flangers, chorusing, etc).

General Purpose of Dynamics Processing

There are two main purposes of dynamics processing. The first purpose is to decrease the dynamic processing of a track, and the second is to increase it. In this first post, I will discuss specific dynamic processes for when you may want to decrease the dynamics of a track in a compressor. In the second post, I will do the same for limiting and clipping. Without further ado, let’s dive in.

Compression

As the name suggests, a compressor typically ‘compresses’, or ‘squashes’ the dynamics of a track. The opposite of a compressor is an EXPANDER, which I will discuss separately in a future post.

INPUT GAIN - The gain you are putting into the track. It is the equivalent of putting a gain or a trim plugin before a compressor.

THRESHOLD - You will often hear that audio that goes above a threshold is where the dynamics processing takes place. If the gain reduction meter isn’t moving, or you look at the needle and it isn’t twitching, the threshold is set too high and your audio has not yet been compressed. I will later explain why the ‘audio going above the threshold gets compressed’ segment isn’t the most accurate explanation of the threshold. After I have provided an explanation on other parameters, I will delve more into its true function on a compressor.

RATIO - The ratio determines how much of the signal is compressed. I will provide you some examples of how a compressor ratio works:

Ratio - Example 1:

If you were to set your threshold at -14dB and you had set your ratio at 2:1, this means that for EVERY 2dB a signal increases, you will only hear 1dB above the threshold. Without compression, it could have potentially have been another 2dB of gain without compression. You have 1dB of gain reduction:

(2:1 = 2dB minus 1dB is 1dB)

For all of this to happen in its full effect, the signal has to reach -12dB, with the threshold and ratio specified in this example.

If you implement this ratio, your compressed signal will be -13dB. Without this ratio, it would be -12dB.

Ratio - Example 2:

A threshold set at -5dB, and a ratio of 4:1, means that for EVERY 4dB a signal gets past, you will only hear 1dB over the threshold over what could have potentially been 4dB of gain without compression. You have 3dB of gain reduction:

(4:1 = 4dB minus 1dB is 3dB)

If you implement this ratio, your compressed signal will be -4dB. Without this ratio, it would be -1dB.

Ratio - Example 3:

A threshold set at -3dB, and a ratio of 8:1 means that for EVERY 8dB a signal gets past, you will only hear 1dB over the threshold over what could have potentially been 8dB of gain without compression. You have 7dB of gain reduction:

(8:1 = 8dB minus 1dB is 7dB)

If you implement this ratio, your compressed signal will be -2dB. Without this ratio, it would be 5dB.

ATTACK - Attack is NOT how long it takes for the compressor to start compressing. This is a simplified explanation on what you hear from YouTube, books, magazines and lecturers.

A compressor will start working IMMEDIATELY after the threshold, regardless of how fast or slow your attack is.

Attack is the RATE a compressor will apply two-thirds of the targeted gain reduction. But since each compressor is designed differently by its manufacturer, it may not always be two-thirds for every brand, however it is MUCH closer to what actually happens.

Crucial ATTACK Theory Almost Nobody Talks About:

If your ratio is 8:1, and you have a slow attack of 1000ms, or one second, your compressor will start compressing immediately, but the ratio will take a long time to increase. If your target is 1000ms, and it reaches 250ms, the ratio may be 2:1. By the time it reaches 500ms, the ratio may be 4:1. When it reaches 750ms, the rate of compression may be 6:1, and when it reaches 1000ms, it may be 8:1. Remember, most compressors work at a rate of two-thirds of what you asked them to do, so the actual ratios may be even lower. The last third can barely be determined by you, if at all. Some compressors out there, however, will listen to the attack that you set to them, and apply the full target of gain reduction, or at least pretty damn closely.

Earlier, I mentioned that I would dive in more about what the threshold actually does. In the following image below, made by Music Sequencing, you can see that the second peak has been drastically reduced, while the rest of the peaks have barely changed. Notice how these first three peaks you see are all above the threshold. How is this possible? All of the peaks should be drastically affected, right? Well, the RATE set by the ATTACK seems to be set fairly fast, but the RATE of the RELEASE may be fast too. This may explain why there is a small difference between the two waveforms. On the right waveform, you can see that the second peak has already had the extreme 8:1 ratio implemented. The third wave onwards has already recovered. (In this drawing by Music Sequencing, a horizontal dash above the threshold represents a decibel of gain.)

An extremely absurd analogy may be to think of the attack as an ‘aggressor’, ‘bullying’ the top peak to be much closer in level in relation to the other peaks. But the RATE can determine how much it will ‘enforce’ the ‘squashing’ of the peaks.

Below is a link to a Facebook post,containing an analogy by Gregory Scott. He explains the concept of attack, which may be particularly useful to those you that drive on the road:

There is another video by Alex from Music Sequencing which illustrates what actually happens to the waveforms, for even more explanation. Adding to what Gregory Scott said in the screenshot, this analogy is exactly another reason why you may hear others say that a compressor starts ‘compressing immediately’.

RELEASE - The release is the opposite of attack. Release is the RATE a compressor applies two-thirds (usually) of the targeted gain being restored. Release is NOT how long it takes for the original signal to recover. Consider what has been said about attack and reverse the process into release.

OUTPUT GAIN (makeup gain) - It’s important to level match what you did with the compressor in relation to the original signal. It is the equivalent of putting on a gain or a trim plugin right after a compressor. If you were to bounce out the results, you may notice a small to a large difference in your waveform. If you’ve set your attack and release long, set a small ratio and you level matched, you may notice that only some of the transients have been altered. This is a subtle example of compression. The differences can often be difficult to hear on one sound, but if you do this to many sounds, the results of compression become more obvious. However, to fully understand compression, it may be best to over-exaggerate the amount so that your ears eventually get used to less amounts of compression. Easier said than done.

When level matching, you should avoid having the same threshold as your makeup gain. For example, a common mistake is you set your threshold at -6dB and your makeup gain at 6dB. The sound will most likely be louder than before, and so you are not necessarily level matching. This is especially true if you set a longer ATTACK and RELEASE, because the envelope will not be consistent. Longer ATTACK and RELEASE is especially effective if you want the sound to be compressed, but also have parts of the sound that aren’t being heavily cut down on. In some cases, the signal may also be very hot or distorted, and so a low THRESHOLD and high RATIO you set will make your makeup gain uneven compared to your input.

Thank you for reading this first part of decreasing the dynamic range, concerning compression. In the next part, I will be discussing this further in the context of limiters and clippers, as well as summarising when you should use these three processors.