Learn Music and Audio Production | iZotope Tips and Tutorials
8 Common Mistakes in Mid/Side Mixing
Once upon a time, people spoke of Mid/Side mixing in hushed tones—this was a process left largely to mastering engineers, who had specialized gear that could transform common L/R stereo signals into Mid/Side. Today, most of us operate from a digital baseline, and in our digital playground, Mid/Side options have become de rigueur. Mid/Side is understandably attractive to recordists, producers, mixing engineers, and mastering engineers alike, as it provides even more detailed control over your audio.
But take heed: Mid/Side techniques can ruin your mix if applied poorly, which is unfortunately very easy to do. So let’s dive into some common Mid/Side mistakes, and what we can do to avoid them.
Note: This article covers the mistakes of Mid/Side mixing and mastering techniques; we won’t be covering Mid/Side recording here.
1. Not understanding what Mid/Side actually is
We think of Mid/Side as “mid/sides”—a way to separate the apparent center of the audio signal from the apparent sides (i.e. the extreme stereophonic elements). Snare and vocal? Mids. Drum overheads and double-tracked guitars? Sides.
But this is not exactly correct. The better term for “mid/side” is arguably “sum/difference,” and here’s what I mean:
What we call the "mid" is actually the sum of the left and right channels. This ends up emphasizing things which are the same on both sides, and de-emphasizing things which are different. Since similar material is emphasized on both the left and right channels, the resulting audio will be perceived as mono. The “sides” are also a mono signal: it comprises all the information from the left side minus the information from the right side (left + phase-inverted right).
Full disclosure: there’s more that goes into the routing of a Mid/Side matrix that I can describe here (you can read more about Mid/Side matrices in this article). But highlighting what Mid/Side actually is helps us illuminate a common Mid/Side mixing mistake:
Mid/Side does not give you perfect separation between the center and the sides. There will always be overlap. Failing to understand how this overlap can manifest in your mix can result in severe and odious effects.
Consider this: encode stereo audio into an Mid/Side matrix, and any audio that exists in the far left corner of your mix will be found in both the M and the S signals. How could this be?
Well, the left channel would have some amplitude (let’s say x), and the right channel would have none (so 0), because we’re talking about audio panned hard-left. Using our formulas from earlier, we can see that Mid = L + R = x + 0 = x. However, Side = L - R = x - 0 = x. So audio on the far left appears identically in both the M and S signals.
What if we’re dealing with audio panned to the far right, meaning that L = 0 and R = y? Well, Mid = L + R = 0 + y = y. But Side = L - R = 0 - y = -y. So audio on the far right will be found with the same amplitude in the M and S signals, but will be phase-flipped in the S signal.
In practice, this means that even instruments that are panned hard to one side will actually be affected in both the M and S signals through Mid/Side processing.
Say you’re mastering a tune with the guitar panned to the far left. Let’s go further and say you’re using a dynamic EQ to tamp down the shrillness of a vocalist. Well, if you’re doing this in Mid/Side, you’re also going to affect that far-left guitar, which could be problematic—it could mess with the instrument’s clarity, and cause a weird, spatially-distorted effect to boot.
So, if you’re using Mid/Side techniques on any sort of global scale, make sure you keep in mind how audio will be split amongst the Mid and Side signals.
Learn more about using Mid/Side for mastering in the video below:
2. Improperly using Mid/Side compression
Even 10 years ago, a plug-in with Mid/Side capabilities felt like a real gem. These days, the process is freely available on many digital processors, including Ozone 10. Even if you’re mixing with stock plug-ins, free Mid/Side matrices abound to turn any two instances of a compressor into a Mid/Side processor.
But proceed with caution: I would advise against using Mid/Side compression, unless you’re trying to get a specific creative effect. For naturalistic widening, the technique should not be your front line, and the reason for this relates to the last tip:
If something exists in the far left channel, like an electric guitar, it will exist in the M and the S when put into a Mid/Side matrix. So when you apply Mid/Side compression to something like an entire stereo mix, the same element will receive different compression treatments, and you’ll wind up with a singular sound being pulled and pushed in different directions by two different compressors.
So if you have a rock mix with guitars in the left and right channels, and you want to apply Mid/Side compression to compress the center channel more than the sides, you could very well wind up with a left-panned guitar both rising and following in level at odd intervals; this could destroy the balance of the mix as a whole.
A better use of Mid/Side compression, I believe, centers around trying to get desired effects that don’t occur in the natural world. Take an EDM mix with a bunch of stereophonic synths: you could slap a multiband compressor on them, put it in Mid/Side mode, and sidechain the S channel to the kick drum. Then, when you get that pumping which is a hallmark of the genre, the instruments won’t just crescendo in level—they’ll also increase in stereo spread in relief to the kick. That’s a cool effect for Mid/Side compression.
You can also use this with compression’s misunderstood cousin, expansion, to boost the side signals of an EDM mix when something interesting happens, like an intricate high-hat pattern. This too is a creative application of Mid/Side dynamics processing that is a credible tool, if done for the right reasons.
3. Not making use of a simple EQ in Mid/Side mode
It feels so tempting to reach for a stereo-image-enhancing plug-in, maybe even the free Ozone Imager—their results are so sweet, so seductive. Unfortunately, they can leave one feeling discombobulated when heard for more than thirty seconds at a time, and often contribute to ear fatigue.
That’s why it can be a mistake to jump to a broad-strokes imager first when trying to correct or enhance a stereo mix. They make for great sound design tools, but for subtle work, I would turn to a Mid/Side EQ. Mid/Side EQ can be a wonderful tool for enhancing the stereophonic quality of instruments because they allow you to work surgically.
Many times in a master, for instance, I’ve come across a section of a tune where I wish the background vocals—layered on the sides of the mix during a chorus—could be more present. Luckily, it is the easiest thing in the world to make that happen with Mid/Side equalization. The results are often relatively inconspicuous as well:
Simply identify the frequency ranges of the vocals (say they’re anywhere between 500 Hz and 2.5 kHz), and automate a slight, wide boost of their center-frequency when they come in on the side channel. Don’t go over, say, 1.5 dB at the maximum. With little intrusion, you can raise awareness of those vocals.
Mid/Side EQ can also be used in a mixing context for carving space, or creating effects, and we’ll get further into that in the following sections.
4. Overdoing boosts to the S channel
Sometimes we overdo Mid/Side processing—especially on the sides. Lifting the sides seems great at first, but you must maintain a subtle hand, otherwise you distort the image of the stereo field in a way that does more harm than good.
This can be especially brutal in the mastering process: if you boost the side channel too much, you can destroy the balance between the elements of the mix—a balance the mixing engineer has fought hard to achieve.
It’s tantamount to taking a photograph, putting it in Photoshop, and doing everything you can to make the background elements as sharp as those in the foreground. The effect might be cool at first, but look at it for a while, and everything feels a bit off; we’ve lost our sense of perspective, or it feels too unnatural. Everything is pushed way upfront and seems overly emphasized. In the sonic world, this can happen when we place too much emphasis on the S channel.
You can also accrue this spatially-distorted effect within individual mix elements. It might be tempting to increase the width of the drum bus with a Mid/Side technique, but then the kit might feel less real. To achieve the right blend and balance of drums, make sure the phase relationships between elements are exactly what you want them to be, so the space is clearly delineated.
5. Not making intentional panning decisions before applying Mid/Side techniques
If you’re going for Mid/Side techniques in your mix before making use of good, old-fashioned panning, you’ve lost the war before you’ve fought any battle. This approach is bass-ackwards, as the kids would say. Width in a mix comes not from stereo trickery of any kind, but from good, careful, and intentional use of panning—especially with mono elements.
Say you’ve got a stereo piano, a stereo Hammond organ, a stereo drum kit, a stereo bit of background vocals, four guitars, stereo synthesizers, and a horn section. You could put all these stereo tracks up the middle and use Mid/Side techniques to make them all fit, but where would the focus be? Where would the feeling of impact come from?
In a dense mix like this, you’re far better off eschewing one side of the stereo piano, organ, synth, and perhaps drum tracks or collapsing them into mono. As the guitars and horns are mono tracks, you can pan them into interesting corners of the mix, along with your mono pianos, organs, synths, etc.
Making an intentionally-plotted chart of all these elements throughout the corners of the mix will give you a mix that ultimately feels wider than a whole bunch of stereo trickery across stereo stems.
6. Destroying transient relationships or phase relationships
Panning isn’t the only thing that determines width, believe it or not: how you handle the transients has a big influence on the spatiality of the mix. If you have overly-compressed elements in the hard left and hard right positions—and if you EQ those elements so that their midrange presence is muted—the mix won’t feel as wide as if you left it snappy.
It goes back to our earliest days on this planet: when humans had to be alert for animal predators, our ears became attuned to the sounds of twigs being crunched, paw prints on mud, anything that sounded like an impending attack. With all this evolutionary buildup, our ear is drawn to transients, particularly in the midrange. You can use this to your advantage when mixing for a wide feel.
But you know what can ruin this effect? You guessed it: Mid/Side processing on stereo instruments. Overcook Mid/Side on a stereo piano or a drum bus, and you can smear transients to an unnatural degree, thus creating obfuscation rather than spatial clarity.
Why is this so? The answer lies back in Mistake #1.
Mid/Side techniques can also very well destroy phase relationships between instruments, or the character of an instrument itself. This is one of the biggest issues with Mid/Side techniques, especially when deployed across a stereo stem.
Take a piano stem: if we were to engage Mid/Side EQ trickery on a piano, boosting the lows 8 dB on the sides and cutting by a similar margin in the mids, we could result in a track that, when played back in mono, wouldn’t be very audible, because of the mangling we’ve done to the stereo waveform.
Be careful when engaging in such feats during the mix. When I’m working on a stereo synth, I tend to have a correlator up to look at every once in a while, and if things go too far into the negative, I check everything in mono and make sure I can still hear the instrument. This can help you avoid the mistake.
7. Failing to hear when you’ve overcooked a mix with Mid/Side stretching
It took me a few years of stereo bus work to unlock this realization—but once I did, things tightened up immensely. An over-compressed mix and an over-widened mix can induce a feeling similar to ear fatigue, especially on loudspeakers, hi-fis, monitors, and car speakers.
I came to this conclusion after trying to suss out apparent compression issues I was hearing across a stereo mix: I’d take the mix to the car or listen on my monitors with fresh ears, and I’d feel that closed-off, constrained feeling—the hallmark of overcompression. I’d check my main bus compressor, limiter, and any multiband dynamic processor to see what was going on, and find myself dumbfounded: only -0.5 dB of gain reduction? Musical time-constants? Why doesn’t the feeling go away when I bypass the compressors? What could be going on?
Going through all my processors for the culprit, I hit bypass on a Mid/Side band within my EQ, and then wham!—that “closed off” feeling went away. So, I opened up previous mixes where I had felt similarly, and recreated what I was hearing with exaggerated Mid/Side bands. I began to hear the differences.
Why can overdoing Mid/Side equalization feel somewhat akin to overcooking compression? I suspect it has to do with items covered in tips 1, 6, and 7: Mid/Side processing is not a true delineation between the sides and the center. As such, it can drastically mess with frequency balances, phase relationships, and, sometimes, transient relationships.
So if you find yourself taking the mix to the car, listening to it, and finding it’s overcooked in that special ear-fatiguing way, consider that it may be over-widened too—especially if you know you’ve added Mid/Side processing to begin with. Consider it a useful alarm bell.
8. Widening just 'cuz
As with many things in mixing, sometimes we like to engage in a technique just because it’s there—that is, we have no intention behind the trick. We have all, at one point or another, craved a wider mix. But have you ever stopped to ask yourself the question of “why”? If you did, maybe you’d realize that one particular tune might benefit from a narrower approach—or better still, a narrower approach in one section and a wider one in another.
We might increase the width of an element or a mix for a variety of reasons. It might be genre-appropriate to do so. It might be emotionally satisfying to give the listener an enveloped soundstage. Within the context of mastering an album mixed by multiple people, it might be appropriate to widen one song to make it stand up to the rest. Maybe the record was mixed badly and needs widening to aid in avoiding a cold and sterile feeling (so that it surrounds you, rather than sitting limply in front of you like badly cooked pasta). Whatever the reason, it always helps to ask yourself why you’re widening something in the first place, rather than just going about the process willy nilly.
This certainly doesn’t comprise all of the Mid/Side mistakes you might encounter, but I believe it’s enough to get your mind going. With these ten common scenarios addressed, you’ll be more aware of the strange terrain that encompasses the Mid/Side technique. Keep in mind the first two tips above all, for I have seen many a mix ruined in those first few quandaries. Arm yourself with the knowledge of what Mid/Side technique really does, and you’ll proceed winningly—with caution to be sure, but you’ll proceed nonetheless.