In part 1, I explored methods for enhancing clarity and space in mixes by thinking conceptually in three dimensions: horizontal (left to right), depth (foreground to background, as governed by volume of individual elements), and vertical (frequency). I also addressed reverb and how it can be used to enhance space or overused to fill up too much space. (Read part 1 in the December 2008 issue, available at emusician.com.)
In this installment, I'll address additional methods to open up your mixes and create truly 3-D results. I'll show you how to avoid common problems such as overuse of compression, overly dense arrangements, and misapplied equalization. Such “space invaders” are problematic enough on their own, but when more than one is present in a mix, complex interactions can occur that may smother and overwhelm what were once well-recorded tracks.
Squeeze with Care
When used judiciously, compression is a powerful tool for holding expressive tracks like vocals and instrumental solos in one dynamic place rather than letting their focus slide between the foreground and background. But as with reverb, too much of a good thing can be harmful, and you can end up with an overcompressed production that doesn't breathe dynamically.
Compression becomes a space eater particularly when it is overused on percussive and low-end elements, especially electric and acoustic bass. During consultations for remixing (in the old-fashioned sense of the word) and mastering, one of the things I most often recommend is to remove or reduce compression on percussive and bass instruments.
In the depth dimension, percussive sounds generally exhibit sharp transients that are heard initially in the foreground or midground and then fade rapidly into the background. Functionally this kind of movement is very important because it punctuates sustained sounds and reinforces a feeling of space and propulsion in the depth dimension. Overcompression dilutes attack and punch, and obscures rhythmic relationships by inverting the natural dynamics of a percussive event (see the sidebar “Compression: Don't Overdo It”).
For instance, when low drums have a noticeable decay or a dominant note (or both), compression can intensify tonal components, eating up space and adding muddiness in all three dimensions. Increasing sustain in the low range can create dissonance, beating, or competition with bass instruments, especially when combined with reverb.
Likewise, remember that bass — whether acoustic or electric — fulfills a percussive and timekeeping function in many musical styles. Moderate compression helps keep the bass present and consistent. But overcompressing it sabotages its rhythmic role, boosts muddiness and tonal competition, and can create a variety of problems during mastering.
As with reverb, the origin of space-sucking compression abuse is often a lack of familiarity with parameter adjustments, or a blind reliance on default settings. In particular, failing to alter a compressor's attack time from a default zero setting will make almost any kind of track sound overcompressed and lacking in transients.
Mixes that sound flat can often be revitalized by examining and increasing attack times of compressed tracks to 10 ms or more. This adjustment not only passes initial transients — which are vital to vocal intelligibility and percussive attack — but also lessens gain reduction generally to restore depth and dynamics (see Web Clips 1a and 1b).
Another way to breathe dynamic life and space back into your productions is to utilize stereo-mix-bus compression as an alternative to compressing many individual tracks. To preserve spaciousness and dynamics, my mixes use light compression on individual tracks only where necessary (usually kick drum and vocals, sometimes bass, guitar, or instrumental solos), combined with moderate stereo-bus compression. A third compression level of digital peak limiting is added during mastering. The cumulative effect of three small compression stages — rather than piling on the compression all at once in mixing — works well to preserve the liveliness and dynamic depth of background and midground tracks.
FIG. 1: This screen shot from the Sonnox Oxford Dynamics plug-in shows a good starting-point setting for a mix-bus compressor. To avoid squashing transients, the slow attack is key.
To avoid adding coloration or extra noise, I recommend using a sonically transparent, high-quality hardware compressor or plug-in for the demanding task of mix-bus compression. As a starting point, recommended parameter settings include a low ratio (2.5:1 to 4:1), a slow attack (10 ms or more), a fast release (200 ms or less, depending on tempo and musical style), and a soft knee. Set the threshold so that maximum gain reduction is only 2 to 4 dB, then adjust to taste from there. Adjust makeup gain so that the highest peak gain levels reach -2 dBfs or lower, to allow some headroom in mastering (see Fig. 1).
The main point is that compression causes problems when it limits or eliminates the depth dimension, crowds out space in the horizontal (imaging) dimension, inverts the natural dynamics of percussion and bass, and increases low-end competition and mud. Too much compression also puts all frequencies in your face, forcing the ear to prematurely tune out the subtleties of space and dimensionality as listening fatigue takes over. Overreliance on compression also increases the general noise floor of your mix. Even at low levels, noise obscures the ear's perception of background details that contribute so much to the unique spatial qualities of a mix.
Quality, Not Quantity
A recording often sounds dense and lacking in space simply because it is cluttered with too many tracks. One of the big drawbacks of the high track counts available in DAW recording is that it's easy to keep adding instruments. And seemingly, come mix time, it becomes even more difficult to take anything away to make space.
One of the most extreme cases of overtracking I encountered was a song brought to me with four similar bass tracks running simultaneously! “Beefing up the chorus” is another practice that, when taken to the extreme of adding four or five keyboards and guitars at once, can suddenly double the apparent volume of a song, overwhelm any subtlety, and diminish the potency of subsequent verses.
Whenever possible, I get clients to consider the old adage “Less is more” in relation to their arrangements. In 3-D mix-speak, this translates to stripping away unnecessary tracks and then emphasizing depth and imaging relationships to open up space and dynamics within a wall of sound. Which tracks are going to be most crucial for your music is really case dependent and defies easy generalization. But here are a few things to watch out for if you feel your mixes have become an overstuffed closet in need of a sonic spring-cleaning.
Easy on the Icing
It's one thing to have a lush, lavishly orchestrated track as the centerpiece of your next CD project, but it's another thing entirely if every song you've recorded has multiple keyboard and guitar parts, strings, horns, five or six background vocals, and enough percussion instruments to start a school music program. I call this condition “all icing, no cake.” When this happens to you, it's time to take a step back and admit that you've become a track addict.
One useful exercise for regaining perspective on track addiction is to strip things down and keep your arrangements real. Think for a moment what your ideal real-world band would be: perhaps a folky 2-guitar-and-percussion unit, a conventional modern-rock band, or a larger R&B combo with keyboards, horns, and a couple of background singers.
With this real-world band in mind, go back to your mixes and mute any tracks that couldn't be performed by the players you envision. This may be a big shock at first, but give yourself a chance to get used to this part of the game. Next, devise an arbitrary rule that will govern the number of tracks you allow yourself to unmute and add back to your production. This could be as basic as using only one new instrument per song, and one doubling track.
Revisit your levels and panning and then give the song a listen to see if you can live with this new austerity. (You'll probably add a few tracks back in that you just couldn't live without.) Hopefully this track trimming will clean up your arrangements, improve vocal clarity, and restore a sense of space simply by clearing out the clutter.
Another part of a mix that's often ripe for trimming is doubled tracks. Doubling has its place for vocals and chordal instruments, but too much can easily backfire and turn a mix from lush to impenetrable. As with the previous exercise, start by muting all tracks that double other tracks. Doubling can mean a duplicate performance, a similar instrument, or an identical rhythmic feel on a different instrument. Add doubles back into the mix one by one, starting with the tracks that are most essential.
As you do this, consider the content that you are adding and how it fits into the three-dimensional mixing space. Experiment with foreground-to-background relationships and depth by reducing the level of any doubling tracks at least 3 dB below the existing track. Explore the horizontal dimension by panning the double away from the existing track, or change the frequency of the double to differentiate it within the vertical dimension (or do both). Ideally you will end up using less doubling and reapplying these tracks in interesting and subtle ways to enhance the mixing space (see Web Clips 2a and 2b).
Another way to approach dense recordings is to build them up and break them down by way of the arrangement. Varying the dynamics can really help a song breathe. For instance, after a dramatic bridge or chorus, stripping a verse's support down to one or two rhythm instruments is a tried-and-true way to reengage the listener. Beginning a song with simplified orchestration is another conventional way to build up to an impressive chorus, while establishing a sense of space right from the start.
From an EQ perspective, don't underestimate the spatial importance of treble in your arrangement, especially when using sparse instrumentation. Most instruments — even bass and low brass — have some high-frequency content. But, for example, arranging cello and acoustic guitar behind a male voice may produce a dull or predominantly bottom-heavy production that needs tambourine or hi-hat to establish a feeling of air and help fill the vertical dimension.
Reverb needs to be considered as an interactive arrangement element as well. Once you have lightened your track load and banished space invaders from the arrangement, double-check to make sure that the amounts, lengths, and character of your reverb choices are still appropriate to the production. Bear in mind that changes in doubling, panning, and compression will also impact reverb perception. And whenever possible, resist the urge to fill up that newfound space with reverb.
All Things Being Equalized
Just as blazing sunlight can make it hard to see visual details, and strong scents can overpower subtle aromas, excessive EQ can mask perception of audio nuance. When frequency buildup becomes extreme — particularly in the upper midrange, where the ear is most sensitive — subtle space-enhancing details in a mix are the first to fall victim to frequency masking and hearing fatigue.
On numerous occasions, I have encountered mixes, especially rock recordings, where it is obvious that the guitar is too bright in the upper midrange around 1 to 3 kHz. This timbre can make it hard to hear the “crack” of the snare drum, which then ends up having to get boosted as well. As a further consequence, you might then feel you need to add a brighter edge to the vocals.
As more additive EQ (boosts as opposed to cuts) gets piled on, the final result is often grating, tinny, or downright unlistenable at moderate volume. Such an excessive buildup — in addition to hastening listening fatigue — may smear or obscure subtler aspects of a production. The interaction of compression or reverb in this range further compounds the problem.
Sometimes it is possible in mastering to solve or at least soften this type of masking effect with a judicious frequency cut, or de-essing tuned to the 2 kHz area. In the cases where this approach actually works, the difference in space and openness in the mix is immediately obvious. Occasionally the transformation borders on the miraculous, causing band members to wonder what kind of sophisticated mastering black magic has been used on their tracks.
Of course, a frequency cut to the most active part of the midrange affects all instruments to some degree and could cause presence problems for some tracks. Similar issues can plague mixes where the low end or treble has been pushed out of balance. In cases where the mastering fix previously described doesn't do the trick, it's time to go back and remix the track, focusing more on subtractive rather than additive EQ.
I'll start with the low end. Once the mix level of a basic track has been established, bass boosting can be done with compression, as we have already discussed, or with EQ. A common practice that quickly eats up both space and headroom is boosting with a low-shelving EQ, which increases everything below the specified corner frequency. When applied indiscriminately, this practice typically raises the gain of a track without significantly increasing usable tone, and also boosts 60 Hz hum, ambient rumble, and other muddying artifacts.
FIG. 2: This chart shows key frequency ranges for additive and subtractive EQ that affect various elements in your mix.
To enhance your mix's spaciousness and clarity, keep any low-end EQ boosting moderate and targeted between 150 and 400 Hz (see Fig. 2). This range is most effective to add power and warmth to tonal and chordal instruments, including vocals. The tone and punch of drums can be improved in the same manner. Combining this kind of EQ boost with a careful, low-shelving cut is an excellent way to clear up mud in a mix, enhance bass tone in real-world playback environments, and achieve hotter levels in mastering. In order to make effective judgments about the low end when mixing or tracking, it is also very important to use a subwoofer.
In mastering I find that 320 to 450 Hz is a region where I often use subtractive EQ to attenuate the buildup of room tone and leakage on multiple-miked recordings, effectively reducing murkiness. Improved clarity and space usually results from decreasing the boxy-sounding effects of standing waves in small or acoustically flawed rooms (see Web Clips 3a and 3b).
Subtle attenuation throughout the midrange is another easy way to alleviate a feeling of crowding in a mix, and can make a track seem lighter or airier. Of course, when taken to extremes, subtractive EQ between 320 and 1,000 Hz risks the empty, hollow tone of the smile curve.
As a rule, whenever you add gain to a track with equalization, it is good practice to try to use subtractive EQ to cut the gain of some other frequencies. Make it your goal to keep the level of a track roughly the same, while increasing its desired tone as well as carving out more room in the mixing space. In addition, don't be single-minded about highlighting one frequency area to the detriment of overall timbral balance.
Into the Pan
In part 1, I addressed stationary panning of basic tracks. Now, here are a few advanced tricks to bring attention to the 3-D mixing space with creative panning and movement of sounds within the horizontal panning dimension.
Panning reverb to the side opposite from a dry track is a hip way to create vivid, dimensional space and draw a listener into the mix. This trick is one of my favorites to use on guitar solos, percussion, effects, and vocal accents. It is also possible to employ long, splashy reverbs this way, because the reverb return will be mono or narrow stereo and therefore will not dominate the entire stereo spectrum.
Manual or automated panning is a good way to energize the horizontal dimension in a mix. A Leslie cabinet or stereo Leslie simulation can help introduce some lateral motion into your recordings.
Another way to add space horizontally is to employ asymmetrical effects. Start by duplicating the desired track(s), then pan the original track to one side. Pan the copied track to the opposite side and try out some effects (modulated effects like flange, chorus, and tremolo work great), distortion, or radical EQ or compression on it.
Stereo keyboards, Leslie effects, stereo acoustic guitars, and the like sound full and can be relied on to add a lush sense of space to sparse productions. But as great as most stereo sources sound, adding too many can choke the horizontal spectrum of a mix. For this reason, I usually observe a limit of two stereo instruments (not including drums) in a mix, and pan these sources opposite each other with minimal overlap. Generally, with the occasional exception of synth pads and organ parts that are mixed low, any additional stereo tracks in the arrangement are panned in mono or very narrow stereo.
Conceptualizing the space between the speakers as a three-dimensional realm — rather than as a line between two points — opens up limitless possibilities for the mixer's craft. And hopefully, thinking of mixes in terms of balance, symmetry, and space will help you get over some typical production hurdles.
The tips and methods described here are certainly not intended to be hard-and-fast rules. Nor does this information need to be embraced totally or exclusively. Consider these 3-D mixing concepts as a jumping-off point for your own creativity — a way to make audio dimensional, and to transform the mixing space into a sonic sculpture or an artist's canvas.
Myles Boisen (mylesboisen.com) consumes significant quantities of space and time at Guerrilla Recording and the Headless Buddha Mastering Lab in Oakland, California. Thanks to Jonathan Segel, David Blatty, Kevin Cunningham and the Goat Family, Freddi Price, Wink Paine, and Rube Waddell.
Compression: Don't Overdo It
Loud is good, and most of us have to use some compression to keep our final mixes competitive in this increasingly noisy world. But louder is not always better. Level-headed mixing and mastering professionals, joined by other voices of sanity in the music business, have been warning for years that the loudness war is simply a losing battle. To illustrate that point, this year even hard-rock fans are blogging that the new Metallica CD has suffered sonically from mixes that were brickwall limited before they reached the mastering lab.
So what are the implications for compression in terms of space and 3-D mixing? When a digital mix becomes squashed with compression, the foreground tracks can't get any louder once a high proportion of peaks have reached digital zero (0 dBfs). The background elements — vocal nuances, chordal decays, room sound, and so on — come to the foreground, hence the phrase “in your face.” And this sounds exciting, for a little while. But without some quiet dynamics, there is no longer any meaningful reference for what loud is.
In 3-D terms, if there is no background, your mix becomes a flat plane with no depth and no empty space in the horizontal (imaging) dimension. Similarly, spaciousness is squeezed out and listening fatigue takes over as a result of overcompression crowding the vertical (frequency) dimension.
Tips to Avoid “Space Invaders”
- Avoid muddying your mix with too much compression. Make sure that all compressor attack times are at least 10 ms. Combine judicious use of track compression with mild mix-bus compression.
- Keep arrangements relatively sparse. Decide what a real-world band would be for each song, and try not to add extra parts beyond that.
- Avoid excessive doubling of parts.
- Consider varying dynamics through the song by breaking down and building up the mix (through muting and unmuting) as the song progresses.
- Try panning a reverb return to the opposite side of the track it's affecting.
- Consider adding motion to your mix by automating pans and/or using rotating-speaker effects on selected elements.
- Avoid clutter by limiting the amount of wide stereo tracks in your mix to two. Pan other stereo sources more tightly or make them mono.
- Don't overdo additive EQ. Try to make an equal EQ cut to compensate for level added by boosting.
David Blatty Links