compression – Producer Society

What is Compression? In simple terms, compression squeezes the audio signal so the loudest sounds decrease in volume and the quietest sounds increase in volume. In other words, setting up a compressor is like having an ultra-precise and ultra-quick robot turn the volume up and down on the song's quiet and loudest sections. It's as if the nightclub DJ turned up the volume on a really quiet song, and then dialed it back when a loud track came on right after. A compressed signal brings balance and cohesiveness to an audio recording, and a by-product of this process is the capability for increased total volume on the track afterward. It helps to illustrate what I'm talking about, so I've created a few images and diagrams. The image below is an uncompressed file in Garageband. The quiet part of the recording, between 41 and 47, is much smaller than the louder part of the track, between 47 and 54. *The strength in signal usually corresponds to its size. Now, the next image below is a compressed file in Garageband. You'll notice that the part between 41 and 47 has accentuated a little, compared to the part between 47 and 54. The difference is no longer as large between the quietest and loudest part. That's essentially what the compressor does. It brings balance and consistency across the music. But don't get overly used to differentiating between sounds visually. Try and grow accustomed to how they sound rather than how they look. With that said, however, compressors that come with a visual interface are so much easier to use than the numbers-and-dials option like what comes as a stock plug-in in Garageband. An actual visual reference is quite helpful, and there's nothing wrong with using it. Explained in another way, compression functions by decreasing the loudest sounds/increasing the quietest sounds so they fit within a much smaller range. The result is more consistency, and thus the quieter musical sections can shine. For instance, if we increase the total volume of a track to bring up the quiet sections, an unintended consequence may be too much volume in the following loud parts, therefore, clipping and distortion. A compressor gets around this problem, and because of the decreased size of the overall signal, we can turn up the overall volume by a few dB (make-up gain). It's for this reason that compression is typically associated with the increased overall volume in music, however, what the compressor is actually doing is balancing the harshest and softest parts of the song; the louder volume is just a consequence. When it comes to vocals, for example, a compressor may bring up quieter words and even small breaths for the listener to hear, while at the same time, bring the volume down on easier-to-hear words. The subtle dynamic changes created by compression can be the difference between a great mix and a bad one. Because of that, there are a lot of terms associated with its function. Regarding the mastering process, people associate the compressor's effects with terms like, "pumping," "gluing," "punching," "gelling," and "sticking." Moving on to its parts........ The Basics of a Compressor in Garageband: Threshold, Ratio, Attack Time, Release Time, and Gain Compressors typically have the same parameters: Ratio, Attack Time, Release Time, Threshold, and Gain, although Garageband's stock compressor has only four, omitting 'Release Time.' The compressor in Garageband has four parameters in this order: Compressor Threshold, Ratio, Attack, and the Gain. Each parameter does something different. *It's worth noting that the Compressor settings are a bit different from how the rest of the internet specifies the numbers. When looking up compressor information online, you'll find that ratio settings are only expressed in two numbers: "5:1," or "20:1." Garageband's compressor ratio has three numbers, so the settings look more like "5:1:1." The middle number is for people who want to adjust the compression by 0.1 instead of entire numbers, ie, 4:1 to 5:1 versus 4:1:1 to 4:5:1. 4:5:1 is basically half-way to 5:1:1 Threshold The threshold is how loud the signal has to be for the compression to start. It's the point, in dB, where the compressor actually turns on. Explained in another way, if we were to imagine the audio signal pictured below as modeling clay, a compressor is like squeezing the highest and lowest points of the audio signal with one's fingers to round it out better. The threshold, essentially, is the point on the modeling clay where you start squeezing it down to balance it out with the rest. For instance, let's say that you've set your threshold to -5dB - a high threshold - any signal louder than -5dB will be compressed. As you can see from the image below, setting the threshold to -5dB will only compress a very small portion of the signal. A threshold of -30dB will cover almost half of the recording. If you only want to compress the loudest parts of the signal, the threshold will be a larger number (closer to 0 dB and above) To compress more of the signal, a smaller number is better, so -20 or even -40. It's fairly intuitive when you think about it. If the threshold is set to -40dB, much of the track's signal will be compressed. The lower the signal, the closer the threshold is to the center of the audio signal. Ratio Essentially, how much compression is actually being used; it's the strength of the compression. We measure the ratio using the comparison between the input and output signal. In other words, the ratio is by how much we're weakening the strongest signal outside the threshold. It's calculated with fractions. A 5:1 ratio means, for example, that whenever the signal exceeds the threshold by 5dB, the compressor will drop the signal down to 1dB outside of the threshold. 5:1:1 ratio is medium strength compression Another example: a ratio of 2:1:1, means that when the signal exceeds the signal threshold by 2 dB, the compressor will decrease it to 1 dB above the threshold. A ration of 3:1:1 is considered as moderate compression, 5:1:1 is medium compression, 8:1:1 is strong compression, and when it gets up to 20:1:1, it technically means it's "limiting" rather than just compressing, because it's stopping the signal from exceeding the threshold completely. *Fun fact: A limiter is basically just a compressor with its ratio set super high, 10:1:1 and above. Attack The attack is how fast the compressor kicks in or grabs on to the signal. It is how long it takes for the compressor to fully compress the sound once it has crossed the threshold level. A fast attack time is typically between 20 and 800us (microseconds) and slower times are in between 10 to 100 milliseconds. It really depends on what you're using the compressor for. Most parts of a drum kit utilize a fast attack time, like the kick and snare, and some kinds of singing, like rapping. Engineers need a fast attack time for compressing things like snare and kick drums because those instruments are 1) loud, and 2) the soundwaves are essentially like short bursts. Whenever the drummer hits the snare, it's quite loud and fast, so the compressor needs to be fast enough to quickly latch on to that sound and bring the volume down. Instruments and singing tend to be the areas where more dynamics are wanted. Piano playing, singing, and the guitar, for instance, are all cases where a slower attack speed is sufficient. Using the example of singing, a fast attack time and the high threshold will cause the compressor to grab on to the signal quickly, decreasing the highest point of the signal and increasing the lowest part of the signal, creating somewhat of a "flattening" effect. A really fast attack time can actually cause distortion by disrupting the signal flow. Let's say a sound wave is slow moving, like a bass sound. If the attack is too fast, it means that the compressor thwarts the signal's vibration right away, and so if the signal's slow, it'll literally stop its flow, thus creating distortion. Gain Gain, in other words, is the strength of the overall signal. Due to the compressor bringing the low's up and highs down, the compressor is actually decreasing the total power of the signal. Thus, engineers make up for this by increasing the output gain, which is essentially turning up the total volume on the signal, after we've pushed the highs down and the lows up. More advanced compressors offer meters showing precisely how much the compressor has decreased the signal in dB, therefore, allowing the engineer to "make-up" for this loss of volume by increasing according to the proportionate amount of output gain. In other words, the 'gain' portion of the compressor is a tool for compensation. The more you've compressed the signal, the more gain needed to match previous volumes. Compressing Guitar, Vocals, and Drums It's very abstract and hard to put into words, but a fast attack time or over compression can literally squeeze the life out of the music, whether its vocals, guitar, drums, or bass. The part of the signal that normally is compressed due to a high threshold, strong ratio, and fast attack time is the same portion where the personality of the recording lies. Explained in another way, eliminating the higher dB parts of the music will eliminate its individuality, making the drums or female vocals sound like everybody else's. When a person sings, it's worth hearing those very small breaths and the shuffling in the studio, but we don't want to hear too much of it. Overcompressed vocals, however, will accentuate the breaths too much and then bring down the singing. Most people don't have this effect in mind when they start using a compressor. Compression isn't a panacea to one's recording and mixing problems. Chances are, if the first take isn't good, slapping on a compressor won't fix that problem. You may have to just go back over and re-record whichever part is problematic. Overcompressing a bad part will only make it sound flat, dull, and squashed. As I mentioned earlier in the article, how much compression one uses depends on for what instrument or audio recording you need it for, and in what genre it is. In popular music, there is quite a bit compression due to the fact dynamic music isn't the norm in the mainstream music scene. Hit songs typically aren't very dynamic, because they're mostly played in big clubs, bars, and on people's car stereos. Jazz music, on the other hand, is a genre where the listener takes subtle differences and dynamics more seriously. Regarding how to use compression, experimentation is a necessity, and you'll have to use the pre-sets offered in Garageband before fully knowing what you're doing. The presets are a solid jumping-off point. Guitar If mixing a lead guitar part, enough compression is just maintaining a relatively consistent volume across all of the notes. A compressed guitar part allows for the much quieter notes to be heard, while the loudest ones are brought down just a little bit. Increasing the volume of a guitar part may help the desired notes get louder, but it'll also bring up the volume of the loudest parts as well. Typically at the beginning of the effects chain, compressors are supposed to hold a signal within a set volume range, so it's important not to use it following a reverb or a delay (unless you like the way it sounds). Distortion and overdrive are already very high-gain, so using a compressor on this sort of sound won't sound great. We'll usually use a compressor on clean tones, to accentuate the softly-picked notes and decrease the loudness of the loudest ones. Looking at a non-compressed clean guitar signal, versus a compressed one, accentuates the way the two differ. A compressed lead guitar part will increase the sound of one's fingers hitting and scraping against the strings. It may also accentuate small mistakes in the audio, but one could argue that those small mistakes are part of what makes it unique music. Drums Drums are more sophisticated than guitar parts because mixing engineers will compress and EQ each drum part differently. The snare will be increased to sound "punchier," as will the kick, while the cymbals won't face as much compression. In hip-hop/rap music, for instance, EQing the snare correctly to one's choosing is a crucial part of the mixing process. So is mixing the bassline and boutique 808's. Both the kick and snare will usually have some form of compression on them, whereas cymbals will almost never be compressed, simply because you need the symbol to ring out in the higher frequencies. Vocals Regarding a vocal track, a little bit of compression can do quite a bit for singing. However, it's very easy to abuse. For instance, as I mentioned earlier, too much compression will likely increase the volume of less desirable parts of the singing, like the sound of the person taking a breath. Or the shoes shuffling in the background. Usually, vocals are considered to be one of the more important parts of the mix, and for that reason, they're mixed so as to sit at the "front and center" of the mix. Backing vocals, on the other hand, are usually compressed a lot more. How Do I Actually Use The Compressor? Before using it, ask yourself what you need it for and whether it's even necessary. (It isn't a necessity to always use compression, believe it or not. Sometimes, you may like the way things sound as they are). Personally, I think the best way to go about using the compressor as a beginner to audio engineering, is to simply use the presets that come stock in Garageband. If you're unhappy with the way those sound, you can adjust them exactly as the parameters are laid out: Threshold, Ratio, Attack, and then Gain, and in that particular order. Remember that the threshold adjusts what part of the signal is being compressed, the ratio is the strength of that compression, the attack is the speed at which the compressor works, and the gain is an overall volume knob on the final channel. Before even bothering to touch the knobs, ask yourself for what purpose you're using the plug-in in the first place? The last time I used it, it was for a lead clean guitar part. I wanted to hear more of the subtleties, like my fingers moving around on the strings, and those little soft and quiet sounds. Those subtleties that I just described, those are called "transients," a common term used by audio engineers. Transients, essentially, are short, abrupt, and subtle sounds that one either wants or doesn't want to shine through the mix. The 'ringing out' sound heard after smashing a cymbal on the drum kit is a "transient" sound. It's abrupt; it doesn't last a long time. Knowing that was my purpose for the plug-in, I turned the compressor's threshold up high, around -25dB Threshold, Ratio of 4:1:1, attack to around 10ms, and the Output gain to +2 dB. The aforementioned settings are similar to the pre-set, "Classic Guitar," within the Compressor's smart controls. After choosing a setting from the pre-sets, just go through each parameter one by one and figure out how to change it. Remember, the Threshold is how big the dynamic range is; the ratio is how much compression we're using; the attack is the speed of the compressor, while the gain is the total volume of the signal. I find they work great, and they're quick and easy to use. You can customize each parameter one-by-one to your choosing. Usually, the main thing I use compression for is to increase the overall volume at the end of the process. However, in recent months, I've become a little more sophisticated in my use of compression. On my last song, "Romania," I only used a multi-pressor at the end; I never even bothered to compress any of the tracks. I didn't want to mitigate the dynamism of my kick, snare, shaker, and hi-hat combo. I thought it sounded great the way that it was, so I only used a multi-pressor instead: the setting, "Final Hip Hop Compressor." Compression Today Because of the way in which a compressor can increase the volume of an audio recording, a lot of beginning engineers think of a compressor as something that just makes the recording louder. And due to the relationship between the compressor and volume, in hip-hop production at least, music producers misuse the compressor as just a volume tool. To be fair, however, pop music in the modern era uses a ton of compression and the majority of songs aren't meant to be the most dynamic. Compression is used in nearly everything in pop music because engineers are trying to get their music as loud as possible while still sounding good in what's been called the "loudness wars." A compressor is a tool for dynamics. It literally squeezes the signal. It's not just about volume. And how one uses the compressor depends on the instrument or recording. Using the compressor on a guitar solo will be for a different purpose than for the kick or the snare. Putting aside its uses, the compressor can be abused horribly. It has the ability to literally squeeze the life and dynamism out of one's music, effectively eliminating the music's magic. That might seem like pedantic nonsense, but it's the truth. An overcompressed song doesn't sound that great because it no longer has the natural vibe and flow of music. Conclusion Anyway, just think about what you're doing whenever you reach for the compression. It's a very useful tool, but if you don't feel it's necessary then it's probably not. Compression isn't so confusing, and it gets easier to use the more you play around with. Also, head on over to my recommended gear page for more products that will help you create great music.