Compressing the stereo bus can give your music coherence, smoothness and, above all, punch — but there are plenty of pitfalls for the unwary. We lead you through the minefield that is mix compression.
Mix-bus compression can be a confusing topic for inexperienced mixers and engineers. After all, some great mixers swear by using mix-bus compression, while others prefer to keep the mix bus pure. There are also many misconceptions about how and why mix-bus compression is used, and if used incorrectly, it can diminish a lot of hard work. In this article, we're going to explain what consequences compressing the mix bus will have for your music, what its sonic advantages may be, and how best to set up a compressor in this context. The main question to be answered is: Is mix compression right for your mixing style?
To set the record straight, there is a huge sonic difference between mixing into a stereo compressor from the outset versus slapping a stereo compressor on the mix bus just before you print a mix. For our purposes, using 'mix-bus compression' means mixing into, or through, a stereo compressor that is inserted onto the main mix bus before the signal passes to the master recorder and monitor speakers. The whole idea behind this technique is that you are mixing through the compressor from the beginning of the mix process; you are carving your mix, dynamically speaking, through the compressor, and monitoring the compressor's output.
The point is that the same set of fader settings will actually produce a different mix balance depending on whether you apply compression, and how much. Applying compression after the mix is complete thus risks changing the balance you have carefully set up, so unless you mixed into a compressor from the start, compressing the mix is best left to the mastering engineer. Experienced mastering engineers know that compression and limiting after the mix is complete can result in negative mix balance issues, so they do so with the utmost care.
Used appropriately, mix-bus compression can be another tool for adjusting the overall balance of your mix, giving you 'glue' to help meld sounds together and make the overall mix balance sound more cohesive. Obviously, compressing any signal way too much will cause negative effects, but mixing through a compressor does allow you to achieve positive compression effects with balance in mind.
The most common questions regarding mix-bus compression seem to be: are there sonic benefits? Does the experienced mastering engineer want me to use it? What sort of compressor should I use? What are common compressor settings when starting a mix? How will my use of mix-bus compression affect the mastering process? How does mix-bus compression affect the overall balance, compared with not compressing the mix bus?
There are many sonic and dynamic differences between different types of compressors that are suitable for mix-bus compression. In general, there are two main reasons for these differences: the type of circuit used to achieve gain reduction in the compressor itself, and the range of time constants and other controls that the design affords the user.
Four different types of analogue compressor are widely used for mix-bus compression: optical, variable–mu tube, VCA (voltage–controlled amplifier) and FET (field–effect transistor). Digital processors, either in hardware or software plug–in format, can also be used for mix-bus compression.
Optical compressors are very popular in modern music production, and are generally felt to sound both transparent and musical, partly because the photo–optical detector circuit that controls the amount of gain reduction can be designed with a minimal number of components. Popular optical compressors for this application include the Avalon AD2044, Joemeek SC2, Tube–Tech CL2A and Millennia Media TCL2.
Many modern compressors contain vacuum tubes within their signal chains, but a true 'variable–mu' tube compressor uses a valve to actually achieve the compression, by re–biasing the tube to control the amount of gain reduction — in hi–fi language, 'mu' means gain. A good variable–mu compressor has the characteristic sound quality often associated with well–designed tube equipment — that is, slight but pleasing harmonic distortion and a warmth and smoothness that is hard to find with other compressor designs. Some famous and terrific tube compressors suitable for the mix bus include the Fairchild 660 and 670, Manley Variable Mu and Pendulum ES8.
VCA designs also have a 'colourful' sound quality, but the colour involved is quite unlike that obtained from tube designs. A voltage–controlled amplifier, as the name suggests, applies more or less gain depending on the level of a control voltage; when that control voltage is derived from the audio input signal itself, it can be used as a compressor. VCAs have been at the heart of some of the classic solid–state mix-bus compressors, including the Neve 33609, SSL G–series compressor, API 2500 and Focusrite Red 3.
FET compressors use a particular type of transistor to control the amount of gain reduction, similarly to the way a triode tube works, and are not as widely used for mix bus applications as optical or VCA models, thanks to a distinctive bright and focused sound quality, which can become less pleasing with the amount of gain reduction that is applied. The attack times achievable are generally much faster in FET models than any of the other three analogue compressor types. For this reason FET models are generally used and thought of as limiters. There are, however, classic and modern units available that can produce great results when used on the mix bus, including the Universal Audio 2–1176 and Cranesong STC8.
The sonic attitude of all compressor designs also depends on factors other than the gain–reduction circuit itself. For example, an optical compressor can have a tube input and/or output stage, and can be transformer–balanced or not. These factors create wide sonic variations, and when choosing a mix-bus compressor it is wise to think of the whole design statement rather than just focusing on the 'type' of compressor in question. When you are considering investing in a mix-bus compressor, trust your ears rather than the technical specifications.
There are several other considerations to make when investing in one of the many compressors suitable for the mix buss. It is necessary that the compressor be able to react in exactly the same way to peaks in the left and right channels, so it's no use having a dual–mono compressor that does not provide a stereo link option. Having said that, a dual–mono compressor that does have a stereo link option can provide good bang for your buck, because you willbe ableto use it for tracking as well as mix-bus compression.
A programme–dependent or 'auto' release setting can be very useful on a mix compressor. Busy mixes can present complex and unpredictable dynamic variations, and the release time that gives the most natural or smooth–sounding response may not be the same throughout.
Another important factor is the availability, or otherwise, of a side–chain filter. This allows you to equalise the signal going to the detector circuit, but not the audio path itself, which can be very useful if you have a prominent bass or kick drum in your mix. In this circumstance, filtering some low frequencies out of the side–chain can make the compressor respond more smoothly to the dynamics of the mix as a whole, instead of 'pumping' every time the kick drum is hit! Some compressors go further and allow an external stereo equaliser to be connected through the side–chain circuit.
Digital processors, especially plug–ins, are perhaps more commonly used as internal mix processors rather than for mix-bus compression, but if you mix exclusively 'in the box' they are an option worth looking at. Popular stereo compressor plug–ins often emulate actual hardware units, including many of those named earlier in this article, and prominent hardware manufacturers such as Neve, API and SSL are now choosing to license official software versions of their popular processors.
Processing digitally opens up some interesting possibilities that the analogue world does not accommodate, such as perfect automation and very fine–tuneable feature sets, but there are still many people who believe that the sound of plug–in compressors does not yet match that of analogue hardware, even though there are many plug–ins that claim to add analogue magic through the sophisticated use of digital algorithms. In my opinion, some of the best analogue emulations are made by Universal Audio, whose Precision Mix Buss Compressor for the UAD1 system sounds better than anything else I have ever used in the digital domain.
Finally, there are also multi–band compressors, which divide the incoming signal up into two or more frequency bands and compress them individually. These can be an option for the mix bus, but you may be trying to put the cart before the horse. If you are encountering problems during the mixing stage that are requiring the use of a multi–band compressor over the mix bus, it's almost always better to address them by changing the balance or equalisation of individual tracks within the mix.
If you're familiar with the sonic effects of different compressor settings on individual sources, you'll find that these are not so different when you use the same settings to compress your mix. Many musical events, whether they be snare hits, guitar chords or whatever, have this in common: the impact, where the loudest peak is reached, comes at the start, while the musical resonance occurs afterwards. A compressor allows you to change the balance between them, either by enhancing the initial impact relative to the decay, or compressing the earlier stages of the sound and letting the harmonic resonance breathe.
This principle applies to complete mixes as well as individual tracks. For instance, a fast attack setting and a medium release time will tend to suppress transients and bring up the relative level of ambience and sustained notes. However, mix-bus compression is very dependent on the tempo and the musical content of the song. When compressing individual sources within a mix, you often don't need to worry about how attack and release settings relate to the timing of the song as a whole, but when compressing the mix bus, you need to keep the whole groove of the song in mind. The wrong settings, such as a very slow release on an uptempo rock track, can choke your mix, while a release time that is too fast can cause uncomfortable 'pumping' effects that undermine the 'feel' of the music.
A good way to find initial compressor settings for your song is to monitor the gain reduction with a VU meter. Because the ballistics of a VU meter are relatively slow, the swing of the needle will 'smooth out' small, rapid movements and focus on the 'big picture'. Think of the VU meter as a visual representation of the groove of the song: once you determine the fundamental aspects of the groove, you'll be able to adjust the release setting so that the VU meter pulses in a way that matches the timing of the song.
For instance, if the song is driven by strummed acoustic guitars, you can set the release time so that each guitar strum triggers gain reduction, before the meter then returns to unity gain in time for the next strum, but not at a quicker rate than the groove dictates. If the VU meter is bouncing faster than the tempo of the song, this means that the compressor is trying to get to unity gain too quickly and the release time may be too fast. If the needle stays to the left all or most of the time, the release time may be too slow.
Timing the release setting to the song in this way takes some practice, but will soon become second nature, at least on material that is dominated by a strong rhythmic instrument such as acoustic guitar or drums. With more complex or less predictable material, however, you may find that your compressor's programme–dependent ('auto') release setting is the best bet. A compressor over the mix bus will respond to any content that exceeds the threshold level, and if your mix contains a number of different elements that combine to do so in different ways, a single, fixed release setting may well not be sympathetic to all of the differently 'shaped' peaks that result.
Compared to release time, attack time is a somewhat more straightforward setting to comprehend. The attack time only deals with how fast the compressor reacts to sounds that cross the threshold. Once sounds cross the threshold, they will be compressed depending on the ratio of compression. If the attack time is very fast it will grab more of the attack portion of whatever sound has triggered compression; if it is slow, it will let faster transients go by unscathed, but will allow the compressor to grab slower peaks. If your mix contains nice dynamic instruments with lots of transient information, such as drums, piano and guitars, having too fast an attack time can cause your mix to sound dark and flat; the drums may even sound as though they are behind the speakers.
It is worth noting that even with the fastest attack settings, a conventional compressor will not be fast enough to respond to transients in time to prevent brief overloads in digital recordings. To grab these transients, you would need a limiter with an attack time that was close to zero, or a lookahead limiter. These do exist and can help raise the overall level of digital recordings, but their use is best left until the mastering stage.
Returning to the example of a song driven by strummed acoustic guitar, using an ultra–fast attack time will tend to diminish the rhythmic impact of the instrument, so a slower attack time may be more suited.
Typically, a stereo compressor setup with a low ratio such as 1.5:1 or 2:1, and a medium attack and medium release, is a great starting point for mix-bus compression. Begin by setting the threshold to give a minimal amount of gain reduction, perhaps -1 to -3 dB. When you're experimenting with bus compression for your mixes, set the make–up gain so that the level of the processed signal matches the level of the unprocessed signal as closely as possible. This way, when you are in the middle of a mix you can bypass the compressor and hear what the compression is doing to your mix. Although having the right compressor settings for mix-bus compression is important, you don't want to be adjusting the settings in the later parts of the mix. The best thing to do is to get a feel for what the mix is going to be about early on, set the bus compressor then, and leave it be. Any changes that you make to the compressor settings later will change the internal mix balances.
The best way to grasp the concept of the sound of mix-bus compression is to mix the same song twice: once with compression and once without. This test should give you a great indication of how mix-bus compression affects the overall cohesive balance of the loudest and most important parts of your mix.
When you mix without compression, the relative level of each element within the mix is entirely determined by the position of the faders. When you add a compressor over the mix bus, you're adding another layer of dynamic management, one over which you have less control. In some circumstances, it can feel as though your fader moves are 'fighting' the actions of the compressor.
Typically, the compressor will tend to react to the instruments that are already the loudest parts of your mix, such as lead vocals and drums, because these are the first and loudest signals to cross the compressor's threshold. Different compressor settings may tend to bring out different elements in the mix but, in general, you will find that the harder you hit the compressor, the less difference your fader movements seem to make to the predominant parts of the mix. Consider a lead vocal that was compressed during tracking and is also being compressed in real time during mixdown. If you are mixing this lead vocal, among other sounds, through a bus compressor, the lead vocal itself will become even more compressed if it is the loudest part of the mix. So in absolute terms, mix compression will tend to make the lead vocal quieter, and you may need to compensate a bit more with positive fader values to achieve the loudness that you want from the lead vocal — but also consider that the louder you push the faders, the more you will compress the output! And, of course, the gain reduction that is triggered by the lead vocal crossing the mix compressor's threshold is applied to all elements of the mix at once, so the lead vocal's apparent level relative to the other instruments may actually be increased by mix compression.
Finally, a word of warning: a common mistake made by inexperienced mixers is adding a stereo compressor or limiter to the mix after the mix is complete, while bouncing files, for the sole purpose of making their mix sound louder. This action is usually done out of fear because they are concerned their clients will not understand why their mix is so low in volume when compared to other commercial recordings. This is fine for a 'client mix', but doing this to the final mix pass and sending it to the mastering engineer creates a problem, because it can't be undone. Mastering engineers will most likely add more compression during mastering (if need be), but keep in mind that this is their forté.
Mix-bus compression may or may not be for you, depending on how you work with dynamics when mixing. The only way to find out is to experiment and see which option produces a better mix. Whether or not you choose mix-bus compression as a staple of your mixing arsenal, the mastering engineer should not have to go to forceful extremes to make your music sound stellar!.
To get some other perspectives on the issue of mix compression, I talked to mix engineer Kevin Doyle, whose credits include Anne Murray, Alannah Myles, Emmy Rossum, Glenn Gould, the Passion Of The Christ soundtrack and Sinead O'Connor, and mastering engineer Peter J Moore of The E Room Mastering in Toronto, whose credits include Bruce Cockburn, Cowboy Junkies, Diana Krall, Finger Eleven and Oscar Peterson. Moore routinely masters Doyle's mixes.
WH: Kevin, do you mix with mix-bus compression?
Kevin: "Not usually. I am aware of why some mixers prefer to use mix-bus compression, but I choose to control all of the dynamics internally. For example, when mixing acoustic drums and percussion I create my mix by compressing individual tracks to achieve a cohesive balance for the sound I am going for. Then I will eliminate the transient peaks that are not affecting the overall mix balances by limiting them with a peak limiter. This may only be a minimal amount of gain reduction, like 2dB, with a fast release setting, so that the limiter does not affect the drum mix but effectively reduces the peaks. I have a lot of experience with the mastering process and with what to expect from my mixes, so I make sure my mixes are going to respond in a positive way to the mastering compression and limiting. As I am completing a mix I may insert a compressor and/or limiter onto the mix bus just to reference what is happening to my mix balances; this allows me to hear an example of what my mix will sound like when it is processed in mastering. For client mixes, and for my own reference, I will bounce using a compressor and limiter just to hear the effects, because I need to hear what compression and limiting is going to do to the final mix. Effectively, my goal is to send Peter a mix that will allow him to do his mastering work without him running into snags and changing my mix balances. In a sense, bring the music up to a new level of quality without changing my mix visions — this happens when both the mixer and mastering engineer are working together."
WH: Peter, can you describe for me your views about mixers using mix-bus compression before the mastering stage?
Peter: "I would recommend that mixers get into the habit of using proper mix-bus compression. This will allow me to dynamically process less and still achieve the desired quality and average output level. Think of compression as a really good paint job on a car. It's better to have many thin coats of paint, rather than one thick coat. Coat one, the primer coat, is compression applied to the tracks in the recording and mixing stage. Coat two is mix-bus compression. The third coat is the mastering compression that I will apply. If you send me a mix that's had only one coat, I will have to apply more compression in mastering to achieve the desired finish, because the average RMS level is still too low. Compressing properly in three stages prior to mastering will increase the average VU level dramatically. Three coats will provide a better finish with more lustre and will last a lifetime!"
WH: Do you ever receive pre–master mixes that have been so compressed at the mix stage that in your opinion they are somewhat wrecked?
Peter: "All the time! Mind you, this is not proper mix-bus compression that is applied; what happens is that the mixer will apply an L2 or other brick–wall limiter to the mix bus for the sole purpose of making the mix louder. What I get is an over–compressed/limited mix that I cannot undo. The mixer did not mix into a compressor; a limiter was slapped on at the end strictly for extra volume. The mastering stage at this point is more of a salvage session rather than what true mastering should be. Mastering should be hair, make–up and wardrobe; it shouldn't be plastic surgery! There is a lot that can be done to salvage bad mixes, but keep in mind I'm good but I'm not God! It seems the biggest fear these days of young producers and engineers is supplying their clients with a low–level reference mix before the mastering stage. If they are mixing for a client who is not educated about the mastering process, what they should be doing is supplying the client a 'client mix' that has been compressed a bit to compensate for volume and dynamics issues. Obviously, I want the original mix files, the ones that the mixer heard while mixing."
WH: For mix-bus compression, how do you recommend a stereo compressor be set up? How much gain reduction is typical?
Peter: "Start with a low ratio like 1:1.5 or 1:2, with a static medium attack time, say 50ms, and a medium release time of around 250ms. Using the auto release setting on a bus compressor is a great starting point, and highly recommended if you are not used to hearing what the release time does to the music. Auto release will get you 98 percent of the way. Try experimenting with extreme attack and release settings, like fast attack and slow release, and vice versa, to learn how different settings are affecting the music. Then, when you get a better understanding of what you are hearing, you can fine–tune the attack and release settings to better suit the groove of the song. The amount of gain reduction is going to determine the average RMS output level for the track, so 2dB of gain reduction will allow 2dB of extra average level before compressing in the mastering stage. Typically, conservative levels are the norm: even 1dB of gain reduction will make a difference."
WH: Should you leave the bus compressor static or make continual changes to the settings while you are mixing?
Peter: "You may need to make slight adjustments while you are starting your mix out, but it is typically better to leave the compressor static once you have decided on your settings. What you want to do is mix the other track levels into the compressor and be changing the individual channel compressors as need be."
WH: Should multi–band compressors be used for mix-bus compression?
Peter: "No, multi–band compressors are not recommended. Definitely they can be used on individual track channels if you need to, but whatever problem you are encountering, you should be able to fix it in the mix with EQ and track compression. It is typical to use multi–band compressors in the mastering stage, but even then they are more of a surgical tool for mix problems after the fact, like when I have to perform liposuction on a mix!"
WH: How does one go about selecting a particular mix-bus compressor? Do some perform better with certain styles of music?
Peter: "Definitely, certain compressors react better than others depending on the music style. My favourite mix-bus compressor is the Focusrite Red 3. With this compressor I can find a balance for most mixes. If it's jazz and it needs to be warm and cosy, I can find it. If it's rock and it has to be tighter I can also find it. The Manley Stereo Variable–Mu compressor is another great one for jazz and softer music, because it's lush and expansive. The Joemeek SC2 works wonders for rock. I have many different stereo compressors, but if I had to choose one for everything I would go with the Focusrite Red 3."