A well-deserved winner for Best American Roots Performance at this year’s Grammies, this song demonstrates tremendous vocal maturity and control. On the one hand, there are spine-tingling moments of technical mastery, like the long, developing sustains of “honesty” (2:31), “need” (2:43) and “rain” (3:31), or the seamless transitions between head and chest voices on “so what if the hardwood stains” (0:22) or “we won’t drown in the tears” (2:06). But on the other hand there’s a breadth of long-term range and register, from the sultry lows of the first verse to the forceful highs of the middle section at 3:01-3:22.
But this kind of performance will frequently bring technical challenges with it at mixdown. If your singer only uses a fairly narrow expressive range, then mixing is typically a lot easier, because plug-in settings that suit one part of the track remain just as suitable elsewhere. Where vocal mixing can get really tough, on the other hand, is where your vocalist covers a large register, a wide range of performance intensities, and switches between various head voice and chest voice timbres – as Sara Bareilles does here.
For example, it’s so easy for the low mids to get bloated and the true midrange to take a dive on head-voice notes like “[hones-]ty” at 1:01, but there’s not a hint of that here – switch onto a mono midrange speaker, and the vocal balancing sounds as dependable as on full-range monitors. Likewise, it’s super-common for an airy vocal tone to become over-strident and even shrill when the singer moves to louder, higher-register notes. But when Bareilles really opens up after 3:01 the tone remains smooth and full, yet it’s not as if the harder edge that you’d subjectively expect has been completely stripped away – it’s just in its place and beautifully balanced against the rest of the vocal tone, so you can turn this track up all the way without a hint of abrasiveness.
Engineer Mike Piersante has made it look easy, but this is a vocal-mixing tour de force. The overall tonality is, it has to be said, quite warm, so that’s something to bear in mind, but even if you lift the highs to be more in keeping with pop sensibilities, the vocal tone loses nothing of its perfect poise.
I did have one small technical niggle, though: the occasional bursts of distortion during the song’s middle section. The most audible of these are probably on the first and fourth eighth notes of the bar at 3:01-3:03, on “bone” at 3:08, and on “sow” at 3:14, but there are a few lesser instances besides those – you can hear them most clearly if you solo the stereo Sides signal like this: play_arrow | get_app For the most part, they sound like mastering-stage clipping distortion and, sure enough, if you zoom in on the waveform, you can see the tell-tale flat-tops. Now, you might justifiably say that it’s nothing compared with the kind of flat-topping routinely meted out to a lot of chart productions, and that’s true in absolute terms. However, the thing with mastering-stage clipping is that its distortion side-effects are much less audible and distracting in noisy, saturated, drum-led productions than they are in such an acoustic and restrained production sound as this. Loudness processing isn’t something that can be applied with a one-size-fits-all attitude, because each different production will best tolerate a different balance of processing artefacts. Rampant clipping distortion on Skrillex’s 'Bangarang' is all grist to the mill, whereas even the smallest sprinkles of it here feel obviously out of place.