FIG. 1: Place markers with descriptive titles or comments at key points in the song to document your mixdown ideas while tracking.
Of all the processes involved in music production, mixing is arguably the most complex. Changing only one aspect of a mix can throw the entire balance out of whack. For example, raising the level of the lead vocal throughout a song might cause the instrument tracks to sound relatively weak. Similarly, making the electric guitars louder can create excitement but can also blanket the drums to the point where the groove goes out the window. And EQ'ing one track to make it sound huge might make all the other tracks sound tiny in comparison. You fix one thing, only to create a different problem. You fix the problem you just created, only to cause another. And so on.
Considering how all the elements of a mix interact so heavily with one another and how subjective the end goal of a great mix is to begin with, it's easy to chase one's tail when mixing. But while there's no miracle cure to prevent you from endlessly obsessing over a mix, there are several helpful techniques you can use to finish a mix faster and be confident that you nailed it. In this article, I'll discuss some of the techniques I routinely use to keep a mix on track and make it arrive at the station on time.
A Method to My Madness
It pays to develop and stick to productive methodologies — including those borrowed from someone else — when mixing. Having a routine way of doing things keeps you from wandering aimlessly from one task to another and losing focus. While every mix is different and you need flexibility for dealing with unique circumstances that may arise, following a more or less set routine while mixing will move the process along faster and make it less likely that you'll overlook important details.
For instance, I always begin a mix by setting initial levels, EQ, and effects for the drums and bass. Doing so helps me establish my headroom for the mix early on. That's because percussive and low-frequency elements inherently use up the most headroom (that is, they register the highest levels on peak-reading meters) when compared with other elements of the mix having the same perceived volume (see Web Clip 1).
I don't slave over every detail of the drums and bass at this early stage of the game. I only get their sounds in the ballpark, knowing that I'll need to tweak them further once the other tracks are brought into the mix.
After I've got the drums and bass sounding pretty good, I bring up the faders for the other instruments and vocals in turn to set the Scene (or Snapshot settings) for the start of the song. Then I deliberately move through the song in sections from start to finish to automate any needed mix moves.
One engineer I know feels that working on bass and drums in isolation is a waste of time because everything will change once they are placed in context with the other tracks. He throws all the faders up at once and immediately gets to work on the whole enchilada. I've tried that approach, but it's never worked for me — I've ended up with mixes that weren't punchy enough for my taste. To my surprise, however, my all-faders-up friend produces mixes that sound plenty punchy.
My point is that there are many different ways to mix a record, and no one way is the right way. What's important is that you figure out which game plan works best for you and stick to it. Your mixdown sessions will proceed quicker, your decisions will be more deliberate, and you'll get better results.
FIG. 2: The Avant Electronics Avantone MixCubes provide a great reference for confirming that midrange-band elements of the mix are in proper relation to one another.
Always Something There to Remind Me
On projects for which I also serve as tracking engineer, mixdown begins for me very early on. While I'm tracking, I set markers at key points in the song as mixdown ideas come to me. I give the markers descriptive titles or — in MOTU Digital Performer — attach comments to them that will remind me during actual mixdown what I wanted to try at those points in the song (see Fig. 1). Whether it takes a week or a year to get around to mixing that particular song, descriptive marker titles or comments like “Boost gtr hook next 2 bars,” “Mute clashing kybd, beat 4,” and “Double the ld voc here?” will get me back on track in a flash and resurrect my inspired ideas.
Whenever possible while tracking, I'll also start fashioning the best rough mix I can. This places me that much further along in the process of learning what the song needs when it comes to actual mixdown. And it can also inspire creative overdubs whose spark might never have ignited when listening and playing to a ho-hum rough mix.
Check Your Ego at the Door
The mix engineer who is also the artist on the record often creates their own roadblock to a great mix. The most common mistake they make is that they mix for themselves instead of for the song. For example, the engineer who is also the guitarist might tend to make the guitar too loud throughout the song. And if the artist is an insecure singer, they might bury their vocal instead of placing it more prominently in the mix, where it will be most effective. As you mix, continually ask yourself if your parts are being mixed the way that best presents the song.
Of course, dealing with multiple egos can hamstring a mixdown session with endless debate. The longest mixdown sessions are usually those attended by the entire band or by all the performers on the record. On such occasions, it seems that everyone wants their part to be louder, leading to a circular levels war:
Guitarist: “I can't hear the guitar. Turn it up.”
Bass player: “Now the bass sounds weak. Make it louder.”
Drummer: “That made the drums sound wimpy compared to the bass. Bring them up.”
Guitarist: “The guitar is too low again. Turn it up!”
To avoid getting bogged down in such a situation, either mix alone or with no more than two other people in the room. If you have a voice in selecting who else will attend the mixdown session, choose those who have the best ears and the smallest egos. Three people is actually a good group size, because there will always be a tiebreaker when polar-opposite opinions arise.
FIG. 3a: The Waves Q10 EQ plug-in is used to apply a shelving cut to low frequencies on an acoustic guitar track in order to make room for bass and kick in the mix.
That's a Switch
The use of alternate sets of reference monitors will preclude most EQ'ing mistakes that might otherwise prompt you to remix the same song again and again (assuming your room has been acoustically treated to have the flattest response possible). Switch between monitors whose different strengths aid in the accomplishment of various tasks. Don't rely on one set of monitors to do everything.
For instance, it makes no sense to apply a bottom-octave EQ boost or cut to kick and bass tracks while listening to speakers whose frequency response rolls off below, say, 60 Hz. If you do, you might end up with a mix that exhibits a boomy bottom end when played on full-range monitors. In this case, your bass-lean monitors have convinced you there was less bottom end on your mix than there actually was, prompting you to add more (or not carve away excess). Remixing to fix the bass imbalance will always be a hit-and-miss proposition until you work on monitors that let you accurately hear the effects of your EQ tweaks down low.
Few reference monitors exist that reveal all EQ and level adjustments optimally. For this reason, I frequently switch between alternate sets of monitors while mixing.
For example, when EQ'ing drums I'll listen to my D.A.S. Monitor-8 speakers — which offer excellent high-frequency extension — together with my Tannoy PS-88 subwoofer (a discontinued model). With this combination of monitors, I can hear at once all of the high frequencies in the cymbals' range and all of the lows the kick and toms might produce. I'll switch to my midrangy Yamaha NS-10M studio monitors (also discontinued) when applying EQ to vocals, guitars, and fiddle. The Avant Electronics Avantone MixCubes (see Fig. 2) are my preferred reference for confirming that the relative levels of lead and background vocals and electric guitars are in good balance with one another. Furthermore, I'll alternately listen to the entire mix on all three systems (as well as the Monitor-8s without the sub) before “printing” it to confirm that it sounds great on a wide variety of speakers.
All Together Now
I never work on EQ settings for soloed tracks for very long. Generally speaking, it's important to apply EQ to a track while listening to it in context; that is, along with all the other tracks that should be playing during that part of the song. A guitar part, for instance, might have plenty of presence when soloed but sound murky once it's competing with the bass and piano. Nobody who buys the record will ever hear the guitar soloed throughout the song, so why EQ it to sound right in isolation?
To be sure, it's sometimes helpful to use the solo button to assess those aspects of the untreated track that you're having trouble hearing in context. You can even make coarse EQ adjustments to the track while soloed, to get its sound in the ballpark. But quickly bring in other tracks to hear what the combined effect will sound like, before you take things too far. Otherwise, you'll probably end up redoing all of those EQ settings.
FIG. 3b: The low-shelving band is bypassed on the acoustic guitar track, restoring its bass frequencies, when drums and bass temporarily drop out of the mix. FIG. 4: A brickwall limiter adds distortion to a track when it squares off the top of its waveform on peaks.
While a given track needs to sound right in the context of the entire mix, sometimes it helps to break down a few instruments into a subgroup to shape a composite sound. For example, where the guitar, piano, and fiddle all play the same melodic hook (perhaps in different octaves), you might want to solo all three instruments in order to work on the EQ for each in turn. Listening thus will also help you set relative levels within the subgroup. Once you've attained a dynamite submix, pop the channel group back into the full mix and make any additional tweaks that are needed in that context. You'll probably find this technique to be a lot less confusing and more efficient than immediately trying to EQ a single track with all the other tracks also playing.
Till Death Do Us Part
Once you've got the EQ for a track sounding great in one part of the song, don't feel like you're married to it for the rest of the song. A great mix often requires that EQ settings be changed — sometimes drastically — on one or more tracks in different sections of the song.
For instance, an acoustic guitar that starts out a song alone might require that no bass frequencies be rolled off, in order for the mix to sound warm and full right from the get-go. But once the bass and drums come in, that bottom end on the acoustic guitar track may no longer be needed. In fact, it could be stepping on the bass and making it sound less tight and clear. In this case, roll off the bottom end as needed on the acoustic guitar to give the bass guitar its own frequency range to project itself in (see Fig. 3a). Then, if the acoustic guitar takes the spotlight again in a breakdown later in the song, restore its bottom end as needed so that it doesn't sound too thin in isolation (see Fig. 3b).
Some tracks might need level changes instead of EQ adjustments in certain sections of a mix. For example, the cymbals or hi-hat may need boosting or attenuation in some spots, even while the traps' faders stay static.
Do whatever it takes to make the mix sound great from moment to moment, even if that means making dozens of EQ and fader moves by the time the song ends. While this will take time to execute, you're more likely to be satisfied with the end result. And you won't waste time tweaking and retweaking elusive static settings that you're never quite happy with — another recipe for endless mixdown sessions.
Listen. Don't Listen!
How many times have you declared a mix was finished, only to discover a few days later that a key part was buried? Especially in productions having large track counts, it's easy to overlook an important detail here and there. To help prevent such oversights from occurring, I use a technique I call “selective listening.”
First, I'll selectively listen to only the lead and background vocals while the mix plays back. Yes, I'm hearing the other tracks play, too, but I'm concentrating on how every lyric of the lead vocal sits in the mix and how the lead and background vocals work together. Are any lyrics unclear? Does the vocal dip in level or pop out too much at any time? Does the lead vocal get too edgy or muddy at any point in the song? Do the background vocals overwhelm the lead vocal at any point or fail to cut through the mix? I'll concentrate on these aspects of the mix while listening to each set of my reference monitors in turn, making sure no monitors reveal any problems.
Once I've vetted the lead and background vocals' treatments throughout the mix, I'll do another set of selective-listening passes while concentrating on only the instrument tracks (again, using alternate sets of monitors on each listening pass). Are the bass and drums punchy enough? Did I goose all the key parts I wanted to? Do the cymbals get too loud and shrill during the choruses, or does the hi-hat and side stick get too soft on the verses? Did I succeed in attaining a nice composite blend of tracks on shared melodic phrases? How are the overall spectral balance and dynamics? Allowing myself to ignore the previously vetted vocal tracks lets me hear everything else all the more clearly and catch any problems or oversights that might otherwise have gotten past me. The result is that I need to remix a lot less often.
FIG. 4: A brickwall limiter adds distortion to a track when it squares off the top of its waveform on peaks.
Silence Is Golden
Sometimes, no matter how much you tweak the mix, it just sounds bad in one or two sections. The mix is probably not the problem here. The culprit is the arrangement, and no amount of EQ, effects, or riding faders will help until you start ditching or replacing the problematic parts.
When you suspect the arrangement might be killing your mojo during a particular section of the song, mute individual tracks one at a time to see if ditching any parts improves the mix. Often, muting one or two tracks for just a few bars will clear out clutter and bring back the air in a mix, allowing more musically important parts their time in the sun.
You may find upon muting some tracks that the mix has improved but the song is left with a big hole in the arrangement. If time permits, don't be afraid to create and record an entirely new part to replace what's been ditched. Just make sure it's substantially different from what you got rid of, or you'll be back where you started. You just might come up with another hook that transforms the song.
Pull the Plug
If you've done everything you can to polish the mix and it still sounds bad, maybe you've done too much. With hundreds of DAW plug-ins available for processing, sometimes the temptation is to pile on the sauce to the point where the main dish loses its original flavor and appeal.
Used wisely, the right plug-in can wonderfully enhance the sound of a track. But many plug-ins also degrade sound quality in subtle ways that might not be immediately obvious (most often by reducing depth and nuance). Make sure the improvements and trade-offs sum up to a net gain.
If you've been tweaking a track's plug-in settings for a while and are still not satisfied with the results you're getting, try bypassing one or more plug-ins to see if the track sounds better with less (or no) processing applied. Maybe an overly dynamic vocal track will sound rounder and sweeter if you ride its fader up and down where needed, rather than beating it into submission with a high-ratio compressor. And a hazy guitar track that's having trouble cutting through the mix may not need an EQ boost in the upper midrange after all; there just might be too much reverb — or the wrong type of reverb — on the track. Sometimes less is more.
Hit Your Peak
Considering that it might take an entire day or longer to pull off a truly great mix having many tracks, it's no surprise that ear fatigue can often derail the process. Monitoring at low levels and switching monitors often to change your perspective can help keep you fresh. Equally important is to carefully watch your meters, especially those for your main stereo bus. Keeping tabs on your mix's crest factor on your stereo-bus meters will often point out problems with the big picture that tired ears have missed.
Crest factor is essentially the difference between a mix's average and peak levels. A mix that has a large crest factor exhibits big swings in level when percussive elements such as kick and snare drums strike (see Web Clip 2). Speaking generally about popular genres of music (and not classical music), a mix with a large crest factor will sound more rhythmic and, if the bottom end is in proper perspective, punchier. A mix with too large of a crest factor, however, will sound less full and not as loud compared with others.
A mix with a small crest factor typically causes stereo-bus meters to barely fluctuate at the top of their range (see Web Clip 3). Kick and snare hits may not even move the meters because those instruments are, in fact, somewhat buried in the mix. Such a mix is usually very loud, fatiguing, and deficient in rhythm and punch.
Natural sounds that have both an attack and sustain portion, such as a plucked acoustic guitar string, might produce a peak that is 10 dB or more above its average sustain level. A brickwall limiter can reduce the guitar's peaks so that they swing perhaps only 3 or 4 dB above the sustain level. But the trade-off is that a square wave is created on the peaks (see Fig. 4). That generates distortion.
The same principles apply to a mix's crest factor. A maximizer can be used to reduce peaks and raise average levels to make a mix sound louder, but that also generates distortion. The distortion may be subtle enough that it's not easily identified as such, yet it most certainly will fatigue your ears if it occurs repeatedly throughout the song. On the other hand, even a rock mix with a large crest factor will — all other things being equal — sound smoother, sweeter, and warmer.
Even without using a maximizer or brickwall limiter, a mix can end up with too small of a crest factor. All it takes is having tracks with sustained sounds cranked and drums and percussion buried in the mix. Such a mix will also sound fatiguing.
If you find yourself chasing endlessly after the proverbial warm mix and all the modeled vintage-EQ and tube-emulation plug-ins in your arsenal can't quell the edge, check out your mix's crest factor. If your stereo-bus meters' levels are swinging only 3 or 4 dB throughout the song, the guitars and keyboards might need to be lowered and the drums (and possibly the bass, whose optimal relationship to the kick drum — once it's dialed in — should be maintained) might need to come up in the mix.
Spend time analyzing the crest factor for mixes you admire and love listening to. Then aim to achieve a similar crest factor while mixing your own similar tunes. With practice and the guidance of your stereo-bus meters, you'll be able to move quickly beyond an edgy mix and arrive in warmer waters.
In the Can
The techniques I've detailed in this article should not only help you create a better mix, but also help you do it faster. Just as important, using these tools should increase your ability to recognize when you've made the mix as good as it can be. When it comes to an open-ended mixdown session, knowing when you've crossed the finish line is half the challenge.
EM contributing editor Michael Cooper is the owner of Michael Cooper Recording in Sisters, Oregon. Hear his mixes at www.myspace.com/michaelcooperrecording.