Ah yes. The inevitable compressor article... For whatever reason this is a topic that i think most young producers and novice engineers (age and experience alike) struggle with. I know I did. In fact compression and just dynamic processing in general (compressors and expanders), eluded me for quite some time. I understood the concepts but it took me a while to understand how to appropriately and deliberately use them. In my opinion what makes dynamic processing so hard to grasp is that it's not as glaringly obvious of an effect as say reverb, or delay. This makes it a difficult tool to dial in and understand if you have an untrained ear. Because of this, the majority of the time I was using compression I was using it incorrectly and was in turn taking away from the song instead of helping it. It took me a while to fully embrace and understand the magic of compressors, but once I finally did I started realizing how much my songs were benefiting. If you can tame your songs dynamics, you can maximize the musics impact and movement which not only helps emphasize and support the songwriting and arrangement, but also gives your more control of your mix. So without further adieu, lets jump into it. Here is Compressors 101. We will go over their parts, the basic concepts, a few ways you can implement them, and also how you should be approaching dynamic processing.
Just a heads up. This is dense.... alright now lets go!
So we want to understand compressors and how they work, but in order to do that we have to first understand why they are necessary and what their function is to begin with. In music there is a thing called dynamic range. Dynamic range is the difference between the quietest and loudest part in your song. The more dynamic range in a song means it has quiet moments and louder moments and potentially anything in between. The smaller the dynamic range, the closer your quieter and louder parts will be in volume. We can think of dynamics as variation of volume, and this exists both in terms of individual instruments and the parts they play, as well as the song as a whole. We quantify this volume and their changes with a unit called decibels, or dB for short. The dynamic range of the average human ear is 120 dB which is the difference between the loudest volume we can listen to before we damage our ears and lowest volume before we can't hear anything. While our ears have their own dynamic range, so do our various audio mediums, such as vinyl, CDs, mp3s, and Wav files. With audio formats, the dynamic range is the difference between the loudest dB level before distortion and the lowest dB level before it disappears to the background noise (the noise floor). With tape, this noise is a hiss and with digital media, its the amount of bits available for recording. It shouldn't be a surprise that digital mediums are going to have a higher dynamic range than older style mediums like vinyl and cassettes. With that said, just because you can reach these lower volumes without noise floor issues, doesn't mean this creates the best listening experience. We have to not only take in account the dynamic range of our mediums, but also what the listener is playing the song on (their iphone, laptop, hi-fi system, boombox, etc) and where they are (at home, in the car, etc). Both of these factors can actually increase the noise floor which in turn reduces the dynamic range even more. So as producers/engineers/musicians, in order to make music that can be enjoyed by the most amount of people we have to take this all into account.
Lets quickly break these numbers down. a 78 rpm vinyl has a dynamic range of 50 dB and this reduces dramatically with the age of the vinyl and how many times it gets played. A 16 bit wav files (your traditional cd quality file) has 96 dB dynamic range. If you take into account your average hi-fi set up with average room noise and ambience our 16 bit files dynamic range can reduce almost to 60. And just because you now have a 60 dB dynamic range, doesn't mean that quiet parts dancing in the lower part of that range will be easily heard. If you there is a lot of stuff happening in that lower range you will make the listening experience too strained. This means that your dynamic range for optimal listening is really more like 30 dB. This is a pretty wild thing. This is almost a 1/3 of the dynamic range of the recording medium itself. This is also why a majority of modern recorded music has a dynamic range of around 30 dB.
So if this made your head spin a little bit don't worry. What's important to take away from this is the understanding that we can only get so quiet with our music before we lose the ear of our listener and we can only go so high before we start distorting and reducing the quality that way (even though this is a stylistic choice used in a lot of soundcloud rap). On top of that, this range shrinks as we start taking in real world externalities like the ears of our listeners, the listening environment.
Dynamics are important, and while we don't want to go too low in volume we also don't want our music to stay at a constant volume the entire song. Dynamics are what creates ebb and flow in your song and having moments where energy clears out and moments where energy erupts is key for successful song writing (for most genres definitely not all). Not allowing your music to breathe in this way will dampen the impact it will have on the listener so we must figure out a way to control these dynamics so that we can still have our quiet parts heard and still allow for our high energy sections to stand out all within a small dynamic range.
So now that we understand a little more about dynamics and the limitations surrounding them, we can start learning about how to control them. Since the enjoyable listening range for dynamics is around 30 dB in modern music, it is our job as a producer/mixer to figure out how to fit our quietest and loudest moments into that window. While this may seem like a less than ideal amount to work with, proper dynamic control and processing makes it possible. There are a few different types of dynamic processing but for the purpose of this article we are mainly going to be focusing on compression. With that said, you can think of any tool that manipulates dynamics as a dynamic processor. There are gates, compressors expanders, dessers, limiters, and there purpose is to help you control the ups and downs of your song. In terms of compression, you can think of it quite simply as the act of making the quieter parts louder by reducing your louder parts. You can think of it working in one of two way:
- the compressor reducing the volume of the loudest parts making them closer in volume to your lower parts
- Making levels more consistent so that the lower volume parts are closer to the high volume parts
Either way you look at it, dynamic compression works by bringing the louder parts and quieter parts closer together in volume. This is compression in it's most fundimental form and it's the answer to the question of how we fit impactful music in a relatively small dynamic range. Below is a visual representation of dynamic compression at work.
This is a really basic representation so don't look into this too much, but this serves as a basic visual guide to what I mean by making the quieter parts louder. Notice how the bars on the left have a larger difference in height? This is because no dynamic compression has been applied. Now look at the bars on the right. See how those two bars have a more similar height? Well this is because dynamic compression has been applied to the signal. As a result we have tamed our loudest part by bringing it's volume closer to the quietest part in turn making our quietest part relatively louder. Think of it like this. A 6 foot tall basketball on an NBA court isn't really that tall. In fact he's probably one of the shortest players on the court. If we then move that same player to say a pick up game on a playground court, that player becomes relatively a lot taller because we no longer have the same abundance of tall players like in the NBA. He can now play power forward because everyone else is either shorter or maybe a few inches taller. This is the fundamental function of dynamic compression.
We have the concepts of dynamic range and dynamic compression under our belts, and now it's time to dive into how these things actually work. Like all effects in the audio world, compressors come in all shapes and sizes. In this article we are going to be looking at Ableton's stock compressor because it has pretty standard functionality and also has great visual guides which help drive home some of these points. So lets get into it.
It's easy to get ovewhelmed looking at a tool that you are a unfamiliar with. In my opinion, the best way to tackle something like this is to break it down and figure out a system for how to utlize it. Throughout this breakdown, I will be using this drum loop that i quickly threw together as an audio example. Feel free to download it and follow along, or use your own. Also feel free to use another piece of audio. Doesn't have to be drums since this is apliciable to all things.
Here is a preview:
The above picture is the waveform of the drum loop that I will be using. I'm going to throw Ableton's stock compressor plug in on this track and set all of it's paramters to zero so it's totally transparent and not changing the audio at all. When I hit playback this is what my compressor will look like:
Right now the compressor isn't doing anything but lets take a quick look at those two pairs of bouncing green bars. That's our drum signal hitting the compressor. Notice how one bar is titled "Thresh" and the other bar is titled "Out". Also take note of that gray arrow to the right of each bar. Now lets first examine the "Out" bar. This is our output signal. So whatever audio comes out after the compressor will be shown in dB via this output bar. We can think of the arrow on the right as our output gain. The lower it goes the lower the output gain, and the higher it goes the higher the output gain. Right now, if we were to raise or lower that arrow, it would in effect just be acting as a volume knob. If you look at the bottom of the Out bar you will also see 0.00 dB. This indicates that the compressors output is exactly at unity and it wont turn anything up or down. In other words, when the compressors settings are all set to zero, like the example above, the input and the output of the compressor will be the exact same.
This leads us to our "Thresh" bar. In the context of this compressor, this bar acting as our input meter and what is called our threshold (thus the "thresh") control. Since we know that the Thresh bar also acts as our input metering, and since we also know our compressor isn't active, it should make sense that our input and output meters are identical. Since we aren't affecting the signal with any type of compression, the same signal coming into the compressor is going out of the compressor with no difference in volume. Now just like the "out" bar, we also have an arrow to the right of our "Thresh" bar. This arrow is what makes this bar more than just an input meter. By moving this arrow up and down, you are setting the threshold of the compressor, meaning you are telling the compressor that when the volume of the incoming signal hits a certain point, that the compressor needs to start doing something. So go back up and look at that gif of the compressor one more time and notice the relationship of the input metering and the arrow. You can see that the arrow, or threshold setting of this compressor is all the way to the top of the bar. You can also see that the input signal is only going about halfway up this bar. Even if we maxed out all of the other paramters on this compressor, it still wouldn't do anything because the threshold level of when it is supposed to start working is too high for the incoming signal. Here's another weird metaphor for you.. think of the threshold as the hearing of an elderly man who is partially deaf but can speak every language on the planet. If you aren't speaking loud enough, if doesn't matter what language you are using, he still wont be able to hear you. If you can talk loud enough however, you will break his threshold of hearing, and can then have a converstaion with him in whatever language you want.
So lets put this into action. Lets drop this threshold arrow so it's now right below the top of the green bouncing bar.
We have now told the compressor that when the incoming signal hits a certain dB point, that it should start working. In this case, anything above -23.2 dB (0 dB is considered unity) should activate the compressor. Now remember all of those huge transients we saw in the waveform from earlier? Well those are the transients that are going to be breaking this threshold. Now look at the difference in the input and out metering... See a difference?... If you don't then good! You shouldn't see any difference because all of our parameters are still set to zero!
Lets start making this compressor do some work... We have our threshold set so that our louder transients (kick and snare) will start activating our compressor, and we have our output gain set to unity still, now lets dive into an elusive and hard to understand parameter for most, the ratio knob.
The ratio knob and the threshold knobs are the meat and potatoes of the compressor. The threshold tells the compressor when to activate, and the ratio tells the compressor how much it should attenuate the signal. You will usually see values like 1:1, 3:1, 5:1 and ∞:1 on the ratio parameter. If you are anything like me, then you math isn't your favorite thing in the world, and unfortunately this is where math starts to come into play (very basic math though), HOWEVER it doesn't need to be that complicated. For the sake of explanation I will go into it briefly, but I will also explain how you can look at a ratio knob without worry about math... So right now our compressor isn't doing anything even though our signal is going past the threshold mark. This is because our ratio knob is set to 1:1. So what does that mean exactly? Well it means that for every 1 dB above the threshold line, the compressor will out put 1 dB. So if we are 6 dB above the threshold while the ratio is set to 1:1, our compressor will output... 6 dB. Now if we had a 3:1 ratio that would mean for every 3 dB above our threshold our compressor would produce 1 dB of signal. If we went 6dB above our threshold with a ratio of 3:1 then our compressor would output 2 dB. Lets take this one step further and say we had a 5:1 ratio. Well if you were 5 dB above the threshold the compressor would output... yeah you guessed it 1 dB. Think of the ratio as "dB over your threshold : what your compressor will output".
If that was a little much don't worry. Hang tight for just a little longer and I will explain how to think about ratio without getting bogged down in mathematics and theory... Lets take our example and change the ratio to 5:1.
Now that we have increased our ratio there are two changes that happened to the compressor that we can see visually. For one, our output no longer matches our input. The other change is that we now have a bouncing yellow bar in the middle. These two visual changes correlate with one another. That middle bar is our GR meter. GR stands for gain reduction. Remember when we were talking about bringing down louder parts so that our quieter parts are louder relatively? Well thats happening here. So with our ratio now set to 5:1, for every 5 dB of signal above the threshold, only 1 db of signal will be outputted. The GR meter is a visual representation of that attenuation. If we were to raise either the ratio or lower the threshold, the yellow bar in the middle would start becoming longer because more gain reduction (attenuation) would be happening with the compressor. If we decreased the ratio or raised the threshold, we would then see less activity in the gain reduction area (GR) because less of the signal would be attenuated.
So lets make this a little more simple... The higher the ratio the more the compressor will attenuate the incoming signal. The lower the threshold, assuming your ratio is higher than 1:1, the more the compressor will attenuate the signal. The more the compressor is attenuating, or reducing (or compressing for the sake of simplicity), the more activity you will see in the GR meter section because it is a direct representation of that attenuation.
This leads people to two common questions.
1. If threshold and ratio both increase attenuation, then which one should I use to get the proper amount of gain reduction?
2. How do i know what ratio amount to use?
The key to both of these answers is thinking before you just start twisting knobs. The way I like to approach compression, and really all effects for that matter, is by deliberately thinking about what I'm trying to achieve with an effect. So to answer question number 1, lets say i was using the drum beat we have been using throughout this article, and I really wanted to attenuated the peaks of the snares and kicks because i didn't like how much they popped out. I would probably first adjust my threshold so the compressor activates only when the very peak of the snare and kick hit. I would then think about how much I want the volume of that peak to be attenuated. If i really wanted to squash that peak I would throw a high ratio on the compressor so every time the kick and snare hit, the compressor would squash it down. If i wanted the compressor to just keep it line with only some minor reduction then I would probably use a lower ratio. You might think that I could accomplish the same amount of gain reduction with a lower ratio and higher threshold, but if i lower the threshold too much, then other parts of my signal start activating the compressor. If i do that, then my subtle percussion underneath begins to get attenuated and maybe I don't want that. By being thoughtful with how you use an effect, you can create a process for yourself, and a process is what evolves you from aimlessly twisting knobs hoping you get something right (which definitely has it's place, trust!) to executing deliberately, and in turn saving time and getting consistent results. I'm also a big believer of using your ears too, so don't get too caught up in your meters. This is a good thing to reference when you are starting out, but don't restrict yourself by thinking you have to be at a certain GR number or ratio number. It's better to have a basic idea of where things should be and then once you get them there to tweak until your ear feels like it's right. This takes time, but BELIEVE ME... the more you do it the better you get, and in the audio world nothing beats time in the chair.
We have our ratio and threshold set, we are seeing gain reduction, so we know our compressor is working, now comes the fine tuning. While the threshold parameter tells the compressor when to kick on and the ratio tells it how much to reduce, the attack and release knobs are how you sculpt the compressor to work best with your incoming signal. In the example above we have our attack set to .01 milliseconds, and our release set to 1 millisecond. This means every time our signal passes our threshold, our compressor is going to kick on in .01 milliseconds of that threshold being broken and exactly 1 millisecond after that it will turn off. This is FAST. Significantly faster than the human brain can process, which is another reason why compressors are so useful. I believe back in the day the only form of natural compression was either in the gear people were recording with, or the engineering moving faders with his hand. I don't care how good of an engineer you are, you aren't operating at the speed... This means that our compressor is working extremely quickly which in some cases is exactly what we want. In other scenarios it isn't what we want so this is another parameter where it's good to put some thought in the desired effect.
What makes compressors so powerful is that with the attack and release parameters you can actually make your compressors emphasize your transients instead of just reduce them. This is something that I do quite frequently with my drums. You can accomplish this by increasing your attack time quite a bit and keeping your release time short. The result is your compressor turning on when the incoming signal breaches the threshold and waiting until the time value set on your attack parameter, and then releasing (how fast depends on what you set the release parameter to) after. When you do this, your compressor lets the initial attack of the incoming signal through and then clamps down on what comes after. This leaves you with some super punchy drums. If you then start increasing the ratio you will get even more of that transient because the compressor is clamping down even harder on everything that comes after that transient. This doesn't only work on drums, but on piano chords, vocals, etc. Attack and release can do more than just boost and reduce transients, they can also level out your signal overall. Something I do quite often on rap vocals is throw a compressor with a low ratio value and low threshold and keep the attack and release really slow so that the compressor is almost always reducing the signal. This evens out peaks and makes for a consistent level that works better with the music and sits tighter in the mix.
Just think about what you are trying to accomplish with your compressor, and then try going from there. I see a lot of people who just throw a compressor on anything and everything, twist a bunch of knobs, have no idea what they even changed and then keep it moving. More often than not, these people are doing this because they don't have a process. Instead they are just using a compressor because... well... that's just what you do.. right?
Create a process for yourself and you will be able to conquer any project. It may not sound exactly how you want it to at first, but the more you do it the better you will get and if you have a system for tackling things then your progress will be exponential.
The final parameter to go over is something that we already covered, and thats the output meter. Remember how that gray arrow to the right of the meter increases or decreases the out put volume? Well this is where perceived loudness comes in. As we attenuate and reduce the signal with compression, we usually want to bring up the output gain to match the gain reduction. Since we reduced the volume of our bigger transients in order in order to bring them closer to our quieter parts, we can make up the output gain and turn everything up which in turn makes these quieter parts louder. This creates more perceived loudness. Don't make yourself nuts and feel like you have to match the output perfectly with the gain reduction, but just keep in mind it's function. It's there for you to make up your reduced signal, thus the common name of "make up gain". If you have every heard a modern Jazz record and heard a vocalist and all of those mouth and breathe noises, well i can almost guarantee you that was done with some healthy compression and make up gain. They reduced the peaks and then boosted the remaining signal which brings up all those quiet nuances that you normally wouldn't hear.
The last thing I will mention is that not all compressors are created the same... Some compressors give you nasty artifacts if you push them too hard. Some compressors frankly just sound like shit. Some compressors only have a threshold and output setting. Some compressors are incredibly transparent even when you hit them hard. Some compressors are beautifully non transparent... You get the idea. The point is this is just a primer. Use this knowledge to start exploring what you can do with these powerful tools. I didn't even cover every parameter available on the STOCK ableton compressor, so the topic is dense, but hopefully this info is enough to get your brain working and thinking about how to use them.
Wow.... that was a lot. If you made it through that congratulations. If you are a little iffy on things I highly encourage you to re read and research these topics from various sources. Compression is a topic that you sometimes need to hear from various perspectives before it locks in. I will be doing a part two of this article going over slightly more advanced topics, but if you are feeling motivated you can get a head start by reading up on things like expanders, limiters, gates, and serial compression. We will be going over these things and more in part 2.
If you take away anything from this article it should be the concept of creating a process. I can't reiterate how important this is when learning anything. If you can be deliberate with your action it will force you to start understanding what you are doing quicker than if you were aimlessly trying things. NOW... there is always a time and place to try things, and a lot of amazing music, art and innovation can come from just trying to experimenting, but if you are starting out, get the basics down first. Understand how things work and then get crazy and weird with it. Your future self will thank you.
Hope you enjoyed this post. If you dig it let me know! If you hated it let me know! If you have any ideas or things you want me to cover in a future blog post, let me know!
Thanks for tuning in !