You may have heard many varying descriptions of the difference between track compression and bus compression, usually including vaguely-defined, mysterious terms like “punch” and “glue” which don’t really help us understand anything.
Well, i have a kinda dorky yet effective way to think of the difference between track and bus (aka buss) compression for you.
First of all, what the heck is bus compression? It’s when a group of multiple tracks are combined and sent to a single compressor, which is working on their summed signals as a whole rather than working on each separately. With track compression, on the other hand, each audio part has its own individual compressor, which is reacting to and affecting only that track (unless you get into side-chaining, in which case compressors can fully or partially react to signals other than the ones they are causing their gain dynamics adjusting effect upon… but that’s another topic).
Think of bus compression… as an actual bus. You know, like a Greyhound bus, a London bus, a school bus, or even as a hippie bus painted fuchsia parked in your driveway for far too long. Each track in the group can be thought of as a distinct (and hopefully personable) passenger on that bus.
Everything on the bus moves together. When the bus turns, accelerates, or stops abruptly, all the passengers jostle and shift.
Think of the compression action as the bumps and swerves in the road. All passengers in that bus are feeling every bump and swerve. If one extra-rambunctious passenger hopping around vigorously results in the bus lurching spasmodically, all the passengers on the bus are going to feel it.
Unruly passengers might cause the bus to behave in erratic ways! If a particular track is affecting all the other tracks adversely, consider using some track compression on passengers prior to boarding the bus, to placate them in advance and keep their behavior under control.
Automation can be though of as the bus driver and can be used to keep the bus on the road during extra-bumpy moments, and make sure you get to your destination without any passengers getting sick. Like for example, if a huge bass swell causes everything to dip uncomfortably, perhaps automate the ratio at that particular moment so that the dynamic squashing is less intense.
What about bus limiting? Technically, any master limiter you use is serving a bus function. Limiters tend to be more transparent than compressors, as shaving off small amounts of thin peaking transients can be relatively inaudible to the human ear. If your master limiter is jumping around a bit too much, you might want to consider trying some track limiting on the specific elements causing the most significant occurrences, to curtail them before they hit the 2buss.
For workflow optimization, it might make sense to “work backwards”. Start with bus compression and bus limiting, and mix into that. Then assess which tracks might benefit from their own additional “pre-bus” dynamics based on how unbalancing their influence on the bus is. If you compress and limit every single individual track as a routine habit, you might be applying dynamic alterations to the raw signal when you don’t really need to be doing so.
What about the master channel (aka “2buss)? I tend to think of effects on the master channel as “pre-mastering”, since technically you could avoid them entirely, render your mixdown, then apply them to that and render it, and achieve identical results. However, mixing into effects can lead to a more natural, “analog” sort of working environment in a way, as the effects will change based on how hard you push into them. I advocate, however, while working on a multi-track mix, that you focus primarily on adjusting your group busses, channels, and auxes, and try to avoid the temptation to juice everything up with master channel processing. Doing so can be useful to audition how your track might sound when mastered, but it can also be like painting yourself into a corner. Remember the school bus metaphor? You can’t get off that bus, so make absolutely sure that you trust the driver, or you might end up somewhere you don’t want to be, or even in a wreck. The mastering engineer can do whatever they want to the stereo mix… but they can’t (in other words: don’t want to) mess with the individual tracks, so you should focus on getting those as cherry as possible during the mixing stage.
If you’re struggling with getting a mix to sit right when you compare it to released material, maybe try hiring a mixing engineer. There’s no shame in being the composer, producer and/or arranger of a piece of music and passing along the next step to someone else to inject their unique energy. No matter how good you are on your own, each human has their own nuanced energetic character of distinct personality, and chances are it will benefit the end result to have some forms of external viewpoints involved in a project. This is also a reason to not fall too heavily to depending on the usage of the tempting “AI” processing plugins now becoming increasingly popular—if everyone uses them ubiquitously, all music will end up sounding the same, which seems rather desolate and bland of a sonic landscape to explore, rather than certain songs and artists popping out by sounding fresh and unique.
If you do find yourself loving the results of a master channel effect during mixing, go with it… but be sure to LU-match the before-and-after to verify that it’s actually benefiting the material and you’re not getting suckered by the “louder is better” illusion (whose favorite dwelling place happens to be the 2buss).
If you have bus compression on groups of track types, some of those tracks being individually compressed, and also with compression on the master, it’s like… a bus full of busses full of people on bikes. Whoah. That’s kinda surreal.
But what about Stem Mastering? That’s like post-mixing, which could be thought of as the inverse of pre-mastering. That part of the process is more akin to mixing, technically, but it’s sometimes done at the mastering stage if the artist is a bit unsure of the final mix, or possibly due to format issues, like mastering for surround sound. It’s basically a step created that is halfway between mixing and final mastering. Generally, every single channel is not provided individually because that would be too much minutiae to deal with, but rather as sensible groups so there’s only a handful of stems to work with (all rhythm guitars summed to a stereo track, for example, etc).
Lots of cool stuff on the horizon. Stay real, y’all.