The stereo bus. The master bus. The 2-bus. The Master Out. Outs 1&2. The ol’ Bottleneck. The Mississippi Mud Basin.
Whatever you call it (okay, nobody calls it the Mississippi Mud Basin), the stereo bus is paramount, as it constitutes the mix’s ultimate point of egress: every sound you manipulate, every instrument you pan, every reverb you instantiate, it all finds its way to the stereo bus, or else the listener doesn’t hear it.
It’s no surprise that engineers have their own special way of handling the stereo bus. I’d wager there are as many ways of treating the ol’ bottleneck as there are engineers. Still, we can separate the general approach into three overarching categories: no processing, minimal processing, and a veritable boatload of processing—a large chain that does much of the heavy lifting when it comes to the overall sound of your mix.
We’re going to break down how each method is accomplished, as well as the virtues and vices of each style.
1. No Processing
This is, by default, the most minimal way of treating your stereo bus. Instead of a compressor for glue, an EQ for sheen, a stereo-width processor for width, you forego all of it, and put exactly nothing on your bus—not a limiter, not a gain plug-in; nothing.
I can understand why this seems like a boring approach, but this method does have its advantages, some of which include:
Lack of Safety Net
This method forces you to work hard in the act of balancing and processing all the elements of your mix. There is no compressor to catch a wayward snare peaking above the mix. There is no equalizer to save you from the mud. As a result, you have to practice elegant signal flow and excellent gain-staging. The only thing keeping frequencies from masking each other—or saving your output from clipping—is you!
If you go this route, you’ll often find yourself using automation to accomplish a task that a compressor/limiter could handle with ease. Many are the times I’ve had to cycle a section and move a vocal down in 0.2 dB increments, so that the vocal would simultaneously cut through the mix and avoid hitting the digital ceiling with a sudden peak. Sure, I could’ve grabbed every element in the mix and pulled it down, but that leads us into this next bit:
An Excellent Confidence Booster
The no-processing approach has added benefits for your ego, and by extension, your efficiency. When you’ve hit a good level, and everything is flowing, pumping, and breathing musically, you’ll realize that you’re the one who made that happen, not the tool.
Why is this important? Because you never know when you’ll find yourself mixing in a situation where your tools are not available to you. When you don’t have your usual tools, you want to make sure you walk freely into the situation, without any crutches. This brings us to our next advantage:
An Excellent Shake-Up to a Boring Mix
Look, we all hit plateaus, and we all get stagnant. Sometimes a jolt to our workflow can re-energize our creativity. Sometimes we might have forgotten/overlooked basic tenets of mixing due to the crutches we’ve allowed ourselves for the sake of expediency.
If you’re someone who regularly puts plug-ins or hardware on the master bus, taking everything off and working dry is an undisputed jolt to the process, for all the reasons mentioned above. You’ll find creative roundabouts you’d never thought you’d try.
Here’s an example: on one mix in a strange studio, I didn’t like any of the software compressors. So, I rode the fader of a drum bus in time with the music to accentuate its swing, doing so to approximate what (I imagined) an SSL comp would have done. The effect wound up more musical than I could have ever imagined.
You Give the Mastering Engineer More to Work With
Ostensibly, stereo bus processing is the mastering engineer’s domain—at least, it used to be; everything has changed a bit in recent years, with a glut of good, cheap available processing.
But if you forego all that processing, you’re giving the mastering engineer more choices. Just because you can slap a Fairchild emulator on your mix doesn’t mean you should. That could get in the way of the mastering engineer’s decision to use different dynamic techniques to cohere your album. Similarly, you might have a silky shelf for the high-end, but it might not sound as good—or fit the album as well—as your mastering engineer’s equalizer/good taste.
Keep in mind that you could be one mixer among many for a project, and in these cases, it’s the mastering engineer’s job to tie all those mixes together. If you overcook your mix, you’re making their balancing act a lot more difficult to achieve.
Printing Stems Is Easy
If you’re working on a mix that has even a slight shot at getting licensed for a filmic medium, you’ll need to produce stems that feature the song in different ways. At the very least, you’ll probably need an instrumental track, devoid of a lead vocal, though I’ve been asked to produce vocal-down tracks as well.
Mastering engineers, production companies, and record labels sometimes ask for full-stem breakdowns, so you might need to deliver drums, instruments, vocals, and backup vocals on different tracks.
Here’s where stereo bus processing can get you into trouble: if your 2-bus processing chain is overly extensive, your stems won’t match the original mix, because that extensive processing was reacting to the mix as a whole. That compressor/dynamic EQ combination? It was triggered by the interplay between the drums, guitars, bass, and vocals. Take any of those elements away, and you’ll get a different result.
However, if you have no processing to speak of on the stereo bus, you’ve saved yourself a headache—the stems, summed together, will sound almost the same as the original (modulation effects that introduce random responses will of course play out differently, as will reverbs receiving different signals).
Okay, enough of the benefits. What are the drawbacks?
A Sacrifice in Efficiency
There’s a reason certain processors are described as adding “glue”—they help cohere the sound. It isn’t magic, as much as logic; if you feed a bunch of signals into one processor, it will react to all of those signals in concert. The interdependence of those signals will trigger one unified result from the processor, and a unified result will always sound more coherent and glued together.
Hardware processors/emulations can exaggerate this effect pleasantly, as the unique distortion introduced by their circuit-paths is distributed across every element of the mix in a relational and interdependent way.
You lose the ability to do this when you commit to a stereo bus devoid of processing, as all of the “glue” now comes from the track or sub-group level. In this framework, such interplay relies more on your dexterity than on your choices of tools. I’m not saying you can’t get to that wonderful interactive place—it just might take more work to achieve it.
A Lack of Safety Net
Yes, the advantage has now become a disadvantage: compression doesn’t just provide glue on the stereo bus—it can also tamp down elements that would otherwise leap out. EQ doesn’t just add polish to the master; it can also prevent buildup in harsh places. Take away these elements from your stereo bus, and you have to pay extra careful attention when mixing. You’re in greater danger of clipping.
Okay, let’s move on to our next overarching style of processing, which is…
2. Keeping It Minimal
The minimal approach is a popular path, and I must confess it’s my method of choice. It comprises using gear and/or processing, but doing so subtly, with a light touch, usually for “glue.”
If there’s an EQ, it might be giving a slight lift to the top. If there’s a compressor, it’ll be “kissing the needle” (I’ll explain that in just a bit). I might run a mix through a preamp at line level, or use my summing mixer, or both. I would consider using a summing amp, a preamp, and some light compression to be pushing the limits of minimal processing, but not by much. Using one or two processors conservatively will keep you in this minimal ballpark.
Now let’s explain the concept of “kissing the needle,” in case you aren’t acquainted with the technique:
Use the tips employed in sections three and four of this article to give your mix a musical pumping/breathing feel. Listen specifically to how the snare’s transient-response has changed, and how the whole groove has been impacted by the processing. I tend to employ slow attacks and faster releases here, but there are no rules other than getting what you like out of the process.
Once you do, bring the threshold back towards zero (or even above, if your compressor works that way) until you’re occasionally seeing one or maybe two decibels of gain reduction on the meters; now you’re “kissing the needle.” The needle (your meter) is only moving occasionally.
Of course, don’t rely on the meters alone. Use them to get in the ballpark of this technique and then tweak the rest by ear.
You might be wondering, if I go this route, when do I turn all the 2-bus processing on? At what point in the mix does it benefit me to start mixing through this chain?
Every engineer has their own approach. I can only vouch for mine:
When I set up the static mix, the compressors and EQ are bypassed, but any harmonic enhancers—such as a hardware preamp or a tape-emulation plug-in—are instantiated. Hardware EQs or emulations thereof are usually left on for their innate color, but not juiced in any way.
My initial panning and level-balances are done with no compression, partly to ensure proper gain staging. If the mix has drums, I move onto the next step, which is mixing the drums/bass elements while the rest of the music is dimmed down.
At some point, when I have the drums feeling pretty good, I switch on my compression and EQ. Then I can tweak the envelope of compression against the mixing I’ve already done in the drums, and use the EQ to enhance what I’ve done, rather than fight against it.
If I had to analyze why I do it this way, it’s because I don’t want to find myself fighting global settings in the tweaking of transient material. Although I have to admit, I love that moment when I turn on my API 2500 or Neve Master Bus Processor and the mix suddenly changes into something better.
If I don’t have drums in the mix, and if I’m going with the glue route, I’ll usually have the processors going the whole time, but won’t start “kissing the needle” until I’ve set basic level, panning, and EQ for all the elements in the mix.
Here are some of the benefits of the minimal approach:
A Quick Path to Glue
For my money, there is no faster way of getting that cohesive, glued feeling than putting the right compressor on the master bus and using it within reason. The process that focuses the mix in a unique way, and my speculations as to why have already been mentioned.
A Modicum of Global Control
In a mix, you might arrive at a place where everything sounds too harsh at 3 kHz—not just the vocal or the instruments, but everything in total.
Here you have choices: you can dive into your track count to figure out which elements you should tame, or you can put on an equalizer on the stereo bus and a nudge out the offending frequency. Neither technique is right or wrong, but one is undeniably faster and sometimes more conservative on CPU, depending on what you need to fix the harshness of the mix itself.
It’s Still Easy to Print Stems
If light compression is all you’re using on the master bus, it’s still relatively easy to print stems that, when played together, sound like the original. You basically feed a dummy copy of your uncompressed mix to the sidechain input of your bus compressor, and print each stem through that bus compressor.
For an in-depth look at how to do this, please check out my article on creative sidechaining techniques.
Also keep in mind that this will only work if your compressor is the last process in the 2-bus chain. Furthermore, if the compressor’s sidechain detector sums signal to mono, this technique will fall apart.
Now, onto the disadvantages:
So Many Options, So Little Time!
Indeed, you could drive yourself crazy looking for the right processing here, and that’s especially true in the beginning stages of implementing this technique. It’s also true if you have a ton of plug-ins to choose from!
If you don’t know the intricacies of your various bus compressors, the learning curve for how to A/B master-bus processing is a pursuit that, while not long in the tooth, could take more time than the session demands. Musicians practice their instruments; engineers should practice their craft, too: take a simple stereo track you’re familiar with and treat it through your various compressors beforehand. Note the results. Think of this like practice.
You’re at the Mercy of Your Gear’s Temperament
I know from personal experience that this can be an issue when processing the master bus. Recently my BAE 1073MP, which I often like to insert on the stereo bus at line level, malfunctioned during a mix; I’m sorry to say that none of my Neve-based emulators sound gave me the a suitable replacement, so I had to go back and remix.
Years ago, a similar situation occurred when my trusty API 2500 took itself out of line during a mix. The emulations might be great, but they weren’t giving me what I needed. I could hear the difference, and it bothered me, even as I forged ahead with compromises. I must say that these compromises sucked time and energy out of my lifetime and energy I could have put towards mixing or sleeping.
By the way, even if you’re primarily ITB, you’re not immune to this hiccup; now you’re at the mercy of software updates, which can have adverse effects on your mix bus processing.
Sure, all of this can mess up the track-level of your mixes too. But it’s worth mentioning here because stereo-bus processing can be quite noticeable, and thus, harder to replicate. In the mix, you can fake it a little more; on the stereo bus, it’s a little harder to do so.
Let’s move on to our final kind of bus-processing, which I’ve labeled…
3. The Veritable Boatload!
This method is the most extreme of the three, and involves layering a veritable boatload plug-ins and/or hardware pieces for their specific, innate sounds. It’s not uncommon to see a tape saturation plug-in going into one or two character EQs, all doing nothing (they’re on there for flavor), followed by a widener, a compressor, more EQs, possibly a clipper, and then a limiter.
You’ve probably seen examples of such chains on YouTube tutorials, because many people are proud of their chains (Consequently, if you follow up on these YouTube personalities, you’ll often see them updating their chains at a rapid clip).
Many engineers I know employ this type of heavy-duty processing in two specific scenarios:
1. They are mastering as they mix, as the project cannot afford a mastering engineer, or
2. They want to protect their mix from a mastering engineer they don’t trust—making it impossible for the mastering engineer to raise the level/affect dynamics.
In general, if you’re going this route, I’d advise that you spend a few days beforehand auditioning/getting familiar with all processing candidates in their instantiated states, because generally, you’re going to start mixing with this chain engaged, and go from there.
Personally, I would take a relatively acoustic or uncomplicated mix—something that really belies the processing—and start slapping on exciters, emulators, and more, noting how each one sounds, and deciding which you prefer for the music at hand. This will help you make decisions quicker when you’re actually mixing.
Audition literally every aspect of the chain, down to the limiter. If you’re using something like Ozone’s Maximizer, go through each algorithm and pick apart the differences between IRC I, II, III and IV, as well as the different substrata of each algorithm. As you will have processing engaged from the start, and will be mixing through your chain, knowing how each plug-in sounds will help you make better choices.
Here are your benefits:
The Prospect of a Finished Master
If you are very skilled, and more than a little lucky, you might end up with a finished product by the end of your mix session. The legendary Don Was once was quoted as saying he mixed through his mastering chain in just this manner, so it’s not unheard of.
Personally, I’d rather have someone else master my mix, or at the very least, divvy up the two processes between independent sessions, but other engineers do secure good results from this approach.
A Top-Down Approach
When you’re operating with a veritable boatload of processing on your stereo bus, mixing the individual elements becomes less about what you can add to each track, and more about what you can subtract to make the master-bus processing work.
As we mentioned before, sometimes re-jiggering your whole modus operandi can lead to a more exciting mix. If you’re used to building a mix from the ground up, this top-down approach can be a welcome change.
First you set up a static mix, and then you approach the master bus, tweaking the EQ, tape saturation, harmonic exciters, compressors, limiters, and whatever else you’re using. Then, you use subtractive equalization on a track by track basis with an ear on making the stereo bus sound better and more musical; you simultaneously employ compression where some tracks exhibit too much dynamic range, and add ambiance processing to taste.
The result can be a quicker road to the top of the mountain, with far fewer processors and plug-ins employed on the mix.
A Whole Lotta Mojo!
The fun thing about mixing into Pultec emulations, vintage compressors, and harmonic exciters is, in no small part, slapping these processors onto your mix in their instantiated state. There’s an X-factor they often add, one that can help your mix feel more alive from the beginning.
We’re mixing engineers: we love gear. Sometimes it feels great to embrace our sick love and let a whole lot of gear impart a whole lot of mojo.
As mentioned earlier, this approach can work if you’re the one-stop shop for the entire project. So using a uniform chain across every track can help imbue a sonic similarity between tracks across an album, EP, or single (with B-side).
Yes, these processes impart a whole lot of mojo, with your unique combination supplying a unique mojo. On a scale of multiple songs, these multiple processors, securing their unique combination of X-factors, can offer each tune a similar sonic stamp; this can result in a coherent experience when the listener enjoys the entire offering.
Keep in mind, however, that there are numerous downsides to this method, for instance…
All of These Upsides Are Also Downsides!
Every single one of the advantages mentioned above has its yang—its dark-side. For instance, that bit about achieving a final master right off the bat? More than likely, you will not be able to do this. The odds are not in your favor. It has nothing to do with you as a person except for the fact that you are a person:
The mastering stage is, among other things, a checks-and-balance operation for the project as a whole. Even technology like Ozone 8’s Master Assistant won’t do everything for you (but it will analyze your audio and suggest a logical starting point).
Furthermore, a top-down mixing approach sometimes means overlooking essential problems in a track until it’s way too late.
Here’s an example: say your heavily-processed static mix sounds so good that you forget to ask yourself, “Did I check the phase of this snare drum?” Then, after much mixing, you remember that perhaps you should check the phase of that snare drum. Once you do, it turns out that the mix sounds much better with the phase flipped, but now this glorious, vibrant snare is eating into vitality of your vocals, which already have their chain in place.
Because of the extensive processing on the stereo bus, you might find yourself in a situation where adjusting an EQ on that vocal isn’t getting you anywhere, as it’s fighting an EQ on the stereo bus; you go to adjust the EQ on the stereo bus, and then everything even more sounds off.
So you can see, rabbit holes abound, as do other issues:
All those mojo plug-ins? Stacked one on top of the other, they can have unwanted effects, such as harsh, high-frequency buildups. Likewise, your mix can become over-complicated and sound overcooked from the sheer number of processors. You could strain your CPU into a place where the session itself becomes unstable.
Also, forget about printing stems, because this complicated chain of exciters, EQs, compressors, multiband processors, and limiters will never add up on singled elements. System updates will definitely hamper you for a while.
So yes, there are a ton of downsides to this approach—but it is fun!
Just because there are three approaches offered here doesn’t mean you have to subscribe to any one of them all of the time! While I generally avoid the third one, I frequently volley between the first two depending on the project, the amount of time I have, and how I feel in a given day.
Remember, plateaus exist, techniques become stale, and we can always awaken ourselves with a jolt. It is my hope that these three tricks will help you coalesce your ideas, and force you out of creative ruts.