This production demonstrates one of the things I think hip-hop does best: taking a bunch of disparate sonic elements that many traditional music makers would disregard completely, or would never dream of using in combination, and fearlessly juxtaposing them into a unique tapestry that convinces almost by sheer force of character. Just the wayward tuning of the opening whistle, for instance, would have spurred many producers to retake the performance or fire up Melodyne, but not here, where the phrase’s unflinching repetition instead pulls the ear magnetically into the breathy texture of the whistle itself, and immediately adds a sense of authenticity to the production.
But what really caught my ear in this particular case was unapologetic dissonance of much of the bass part. This first becomes apparent at 1:12, where the line not only contains a prominent internal F-B tritone leap, but the main E and F pitches clash strongly against the sustained F# of the “it’s getting sticky” vocal hook and a clear F# pitch in the upper percussion loop. Then at 2:23, that same bass part is joined by brass layers which introduce a D#-E pitch contour that doubles the bass in parallel major sevenths. Although the bass seems to fall back into a more traditional consonant role following the arrival of the main brass riff (sampled from Young Buck’s ‘Get Buck’) at 2:36, the jazzy chord extensions that appear at 3:13 once more clash repeatedly, pitting Emaj7 and Bbb5 chords against its F# note and Dmaj7 against its E note.
What this approach does is undermine any expectation of traditional harmonic function (ie. that dissonances create a tension that requires a consonant resolution), allowing the coloristic and textural qualities of dissonance to be celebrated in their own right. The Second Viennese School made a stab at this in the classical world back in the 1920s, but unfortunately managed to exclude the bulk of the listening public in the process, something I’ve always felt was a enormous setback. What I love about the kind of hip-hop production Tyler The Creator has given us here is that it feels like it’s essentially pursuing the same underlying goal, but in a much more constructive and inclusive way, straining against the confines of functional harmony, but still maintaining broad appeal amongst musicians and non-musicians alike. With Viennese Schools, as with many things, perhaps three is the magic number…
There’s a fun little piece of production sleight of hand going on in this track. Notice how, towards the end of the prechorus the drums start to develop a roomier sound and the cymbals begin to open out in a classic “here comes the chorus!” build up. Have a listen: play_arrow | get_app Crucially, though, I’ve deliberately stopped that audio file right before the chorus downbeat. So just close your eyes for a moment and imagine what you think the chorus entry should sound like, based on that preamble. Got that sound in your mind? Great!
And now listen to what it actually sounds like: play_arrow | get_app I don’t know about you, but I’d have imagined some kind of indie-rock overdriven guitar texture, as a contrast to the cleaner, tighter band texture of the verses. But in reality, it’s only the drum sound that really opens out at all here, wheras the guitars remain pretty clean and restrained, with only a hint of overdrive and very little upper-spectrum energy at all to speak of. Effectively, this production is doing its best to present the illusion of rock chorus dynamics, except without the guitars!
But why the smoke and mirrors? Well, it strikes me that The Bieb’s breathy vocal tone wouldn’t fare very well against the masking from a bunch of distorted guitars. For all the tats, he’s fundamentally a pop vocalist, not a rock vocalist, so until he changes the way he sings, something else has to give.
A lot has been said about Young’s vocal performance here, and it is genuinely impressive in its loose, conversational style, navigating the fertile ground between pure song and melodic rap delivery. (And I love the little ping-pong delay on the last word of the first chorus at 1:28 too!) But there is plenty to be said about the backing arrangement and performances as well, which have a lot more heavy lifting to do here than in a lot of songs, given the lack of harmonic interest in the ever-present two-bar harmonic loop.
As with Olivia Dean’s 'Man I Need', which I’ve also critiqued recently, the keyboards provide some of the section definition, with the softly brassy pad first heard at 0:19 helping underline the verse’s brooding introspective character and a glassily modulated upper-spectrum synth layer (nicely showcased during the outro at 4:21) opening up the chorus texture – with the exception of the first half of Chorus 1, where a subtle stereo tambourine component effectively serves this function in its place. But it’s the live drums, bass, and guitar performances that really make this production shine, for me, introducing constant variations (and, crucially, of the tasteful human variety, rather than any kind of algorithmic pseudo-randomness) that meaningfully maintain the song’s sense of freshness, despite its almost five-minute running time.
The drummer, for instance, subtly announces the arrival of the first verse with the subtlest of pushed-snare fills at 0:18; reinforces the power of the vocal rhythm for “is that not allowed” midway through each chorus at 1:10, 2:36, and 3:45; and naturally lifts the sense of pace of the first chorus with a busier kick-drum line, and the second chorus by moving the tambourine from eighth-notes to sixteenths. The bass player delivers a tasteful upper-mordent fill heading into the second half of the first verse (0:35), then develops it into a double fill four bars later (0:44); and introduces more sustained lines at 0:16, 0:51, 1:43, 2:18, and 3:27 (none exactly identical, mind) to build towards the start of each of the verses and choruses. And the guitarist pitches subtle fills into gaps in the second-verse vocal line at 1:47, 1:57, and 2:09; gradually lengthens his sustain to build up towards the end of each verse, as well as then seemingly pushing the drive a bit harder to help the chorus sections deliver more power and edge; and contributes a range of spacey, effects-laden atmospheric overdubs such as the one-shots at 0:10, 1:44, and 2:06, the textural layering during the second and third choruses, and of course the soloing at 3:03-3:29.
This is the magic of using live musicians. Whether Young would have had a hit without them in this case, who can say? But I’m pretty confident that they have a lot to do with how big a hit she had, because they make this song more replayable than its core musical content has any right to be.
There have been a lot of productions which have used a sudden modulation towards the end of a song as a cheap means of squeezing out a bit of extra performance intensity, so the key-shifted final chorus repetition we get here has a good deal of precedent – but that’s about as far as things go in terms of it making any sense! For a start it shifts down a half-step, rather than up, which quite literally undermines the idea of giving the final chorus a ’lift'.
But on top of this, there’s not even the slightest musical warning or logic for the key change. The song clearly finishes with the exposed “in the back seat” vocal hook at 2:15. There’s nothing about that ending to suggest that there’s anything left to say, or any kind of build-up to indicate that there’s another iteration on its way – compared with the well-prepared modulations in, say, Abba’s 'Money, Money, Money' or Celine Dion’s 'My Heart Will Go On'. There’s not even the pure shock value of the willfully unprepared modulations in songs like Michael Jackson’s ‘Man In The Mirror’ or Backstreet Boys’ ‘I Want It That Way’.
And to add a further whiff of flippant disdain to the proceedings, it sounds a lot to me as if many of the parts have been simply pitchshifted down, rather than being rerecorded in the new key, undesirably compromising the sense of ‘air’ in the sound as well overlaying a distinct flavour of formant-shift on the vocal parts. If you ask me, what we’re hearing here is a song which was originally recorded and produced to completion at a duration of 2:15, but that was then somehow rejected as too short by someone in a position of power (perhaps some suit in the record company or someone behind the scenes in distribution), leading to a certain amount of last-minute scrabbling around and eventually this threadbare bodge-job that now everyone involved is trying to style out as if it were some kind of bold artistic statement.
Come off it! You’re having a laugh…
In this MOBO Award-sweeping song, Dean continues the increasingly popular trend of two-chorus song structures, an approach that offers more space for chorus repetitions within a shorter overall run time. After all, why bother with a middle section if your song hardly lasts long enough for anyone to get sufficiently bored of the hook that a mid-flow injection of new material is required? That said, I do think it’s a bit of a poor show that the prechorus lead vocals appear to have just been copied between the section’s two iterations. It just seems like a shame that the producers wasted the opportunity for such a fine singer to increase their performance intensity as the song progressed. And what’s the thinking with that treading-water outro section at 2:45? Wouldn’t it have been stronger just to leave the singer’s final “man, man, man” tapering out at 2:43?
The mix features some nice depth and timbre contrasts, with the punchiness of the dry drums set against the warm, diffuse width of the synth-pad texture and the generally fairly reverberant vocals. It’s a configuration that has definite 80s overtones, especially in the chorusey bass synth, LinnDrum-adjacent drum part, and the brightness of the reverb on the vocal and tambourine. But the aspect of this song that I’d say is most rewarding from the perspective of students of production is the keyboard arrangement. I’ve often talked about how, in most mainstream commercial styles, it doesn’t really make sense to talk about a vocal part or a guitar part, because in both cases ‘arrangement’ is usually a more applicable word than ‘part’. (For more on this topic, check out my previous critiques of Noah Kahan’s 'Stick Season' and Billie Eilish’s 'Everything I Wanted'.) Well, this is a great example of the same kind of situation in the realm of keyboards.
Changes in the keyboard arrangement are being used to define the different song sections, for a start. The opening verse (at 0:08), for intance, is underscored with warm-sounding stereo pads that also seem to have some kind of gentle piano layer mixed in (presumably the one you can hear briefly during the drop at 1:45), but when we hit the chorus, a brighter, highly compressed piano component (which you can hear more clearly during the arrangement drop at 1:14) joins in to give the hook a lift. Then when we get to the “talk to me, talk to me” post-chorus section, the upper-spectrum sonics thicken further with the addition of some kind of rich stereo string synth.
But there are small-scale nuances too. Notice, for instance, the slow increase in the synth’s low-end warmth over the course of the first eight bars of Verse 1 (0:08-0:24) up until the entry of the bass synth in bar nine, which I think is quite effective in giving a sense of build-up through those crucial first few moments of the song. And how about the fact that the intro’s chords are noticeably brighter than those of the first verse, which makes a lot of sense in terms of balancing the need not only to announce the song’s arrival with sufficient presence, but also to keep the vocal’s upper spectrum dominant over that of the synth for maximum lyric transmission.
A good deal of the arrangment interest here, however, is in the fills, of which there are plenty! There are the fairly regular melodic bass fills every two bars in the verse from 0:23, for example, and syncopated piano/synth stabs between phrases of the prechorus (first heard at 0:45), but there are also lots of situations where the sudden absence of some/all of the keyboards itself constitutes a fill: across the opening of both post-choruses (1:12-1:17 & 2:25-2:30); at the end of both post-choruses (1:30 & 2:43); in the seventh bar of Verse 2 (1:45); and on the last beat before Chorus 2 (2:04). And before I leave the subject of arrangement, let me also point out the sensitively managed off-beat guitar ‘skank’ (with a lovely tight slapback feedback-echo) that first enters halfway through Verse 1 at 0:22, but which then tastefully drops out for bars 1, 2, 5, and 6 of the prechorus, supporting that section’s pleasing call-and-response quality.
This is a production which features that classic hip-hop staple, a kick-drum that’s distorted just enough to ‘fold’ slightly on itself, as if your speaker’s woofer’s bottoming out, unable to cope with the driver excursions demanded of it. But where a lot of modern productions seem to achieve this kind of sound with simple digital clipping, a glance at the waveforms suggests that we’re hearing a combination of analogue and digital processes here, and I have to say I do like the warmth of the result – the clipping doesn’t have that same hard and rather impersonal ’edge’ that you might hear on, say, a digitally flat-topped dance-pop record.
Other than that, this track also provides quite an extreme depth contrast between the punchy, upfront drum samples and the rest of the track. Yes, the snare does appear to have a dusting of some kind of very short ambience on it, and is plenty wide into the bargain, but on the whole the drums still feel clearly foregrounded in the mix. In a sense, this chimes with the 70s sensibilities of The Enchantments’ ‘Silly Love Song’, an eight-bar snippet of which forms the bedrock of the song. (Apparently, though, it’s a re-recorded facsimile, rather than a direct sample – a common means of side-stepping mechanical copyright restrictions.) There was a trend in the 70s for designing studios with extremely dead acoustics, and then adding masses of plate and chamber reverb to reinstate a lush, expansive sonic wash, and this often gave drums a characteristically tight, upfront sound – despite those reverbs, given that neither plate nor chamber reverbs usually have much of the kind of early-reflection information that most contributes to distancing sounds naturally.
But what’s also interesting here is how recessed the lead vocal is in the depth perspective, by comparison with the drums. Partly, this is a function of the fairly heavy reverb, which forms a clearly identifiable part of the vocal’s sonic character, not just a subtle blending effect. Furthermore, it sounds rather like a room reverb to me, gemerating a real sense of a physical space around the voice, and perceptually distancing it more than the delays, plates, and chambers that most other mainstream productions tend to use these days. The fact that this reverb has significant upper-spectrum energy only emphasises this effect – again, long-tail effects for chart-ballad lead vocals are often quite dull to position them behind the singer, thereby maintaining an illusion of closeness for the singer themselves despite the effect.
And, speaking of high end, the vocal timbre is quite dull-sounding too, which also pushes the vocal backwards in the image by contrast with the upper-spectrum presence of the drums. Just to give some context here, try comparing it with Huntrix 'Golden' and Sabrina Carpenter’s 'Manchild' or (perhaps more appropriately!) with Bad Bunny’s 'DtMF' or Tyler The Creator’s ‘Sticky’. For anyone labouring under the idea that there’s any such thing as a standard ‘professional vocal sound’, hopefully just those four multi-Grammy-nominated productions should lay that misconception to rest.
PS. Just as an aside, it’s ironic that Victoria Justice, who played the lead ‘pop star in the making’ role in Disney’s successful teen sitcom Victorious has now in real life seen two of her co-stars launched into the chart stratosphere instead – the other being a little-known niche artist called Arianna Grande. I have to say, though, it would be churlish to begrudge Leon Thomas his success, as he’s clearly an extremely skilled singer, especially on a technical level. His recent NPR Tiny Desk appearance was particularly impressive in terms of the accuracy of his R&B riff-blizzard. That said, while listening to The Enchantments’ original production, I did find myself reminiscing wistfully of a time when R&B songs actually had some development in their arrangements…
PPS. Also, although that NPR show features a much more conventional vocal-up-front depth perspective than the song’s studio production, I do somehow find it quite amusing that Thomas and most of his band feel like they’re still perceptually distancing themselves by virtue of their eyewear – singer, drummer, guitarist, and bass player are all rocking the indoor shades, by contrast with the keyboard player and backing vocalists who aren’t. Am I the only person who feels my gaze inadvertently drawn more towards the players offering the direct eye contact?
The mix sonics here have clearly been optimised for the most mass of mass-markets, in that the priority is clearly maximum audience reach rather than necessarily the most pleasant listening experience. So not much mix real estate is going to headroom-stealing low end, for example. There’s no EDM weight to the kick-drum here, despite abundant disco influences elsewhere in the production, with its main low-end energy being centred around 50-60Hz and pretty restrained in the balance to boot. Similarly, the chorus-section’s bass seems to have been fairly aggressively high-pass filtered below 60Hz, given that there’s precious little fundamental frequency to any of the notes except the highest (the C that underpins the final bar of the repeating harmonic progression) – something that results in a rather musically uneven line as that note suddenly booms out every four bars!
The extreme high-frequency density of the mix tonality is another clear indicator of the production’s commercial targeting, ensuring the clearest transmission of the hooks despite environmental interference (think car noise, hoovering, workplace/mall ambience, that kind of thing) and off-axis or acoustically shadowed listening positions. That does, however, make it a pretty abrasive listen on wider-range loudspeaker systems, and also pretty fatiguing for some headphone users too, which means it’s the kind of track where it’s hard to turn up the volume very far without wincing.
The stereo image fits the chart-pop mould too, with masses of width in the chorus’s textural guitar/synth layers and effects to impress earbud listeners, while all the core music content (ie. kick, snare, bass, vocals) remains fundamentally central to avoid any significant transmission loss under mono listening conditions. In particular, it’s worth noting how narrow the layered vocal texture’s image is here. This is quite unusual in pop productions, which often use wide-panned vocal double-tracks as image-widening fairy-dust – as in another of this year’s big hits, Huntrix’s 'Golden', for instance. Single-sided listening is also well catered for, with stereo elements generally being fairly evenly spread across the panorama and most panned mono sources being balanced using opposition panning (such as the double-tracked guitar riff at 1:13-1:20) – a notable exception being a few incidental backing-vocal responses (eg. “hey men!” at 2:21 and “always come running” at 2:45).
However, what really caught my ear in this mix was the beautifully paced long-term mix dynamics, and in particularly the way the production makes each chorus section a big deal in its own right, but also manages to build up the intensity of that section as it recurs throughout the song. For comparison, here’s the start of each chorus: Chorus 1: play_arrow | get_app Chorus 2: play_arrow | get_app Chorus 3: play_arrow | get_app On the most surface level, the fact that the second and third choruses begin with a big cymbal hit differentiates them, but you can also easily hear that the first chorus lacks some of the vocal layers too. However, it’s easier to appreciate the other textural changes if you compare midway through each chorus, away from those distracting cymbal hits: Chorus 1: play_arrow | get_app Chorus 2: play_arrow | get_app Chorus 3: play_arrow | get_app This reveals how much the chordal ‘filling’ of the sound is being inflated with each successive chorus, most audibly with the introduction of background distorted guitars in chorus two and sparkling sitar layers in chorus three. One of the most common shortcomings of project-studio mixes is the lack of this kind of long-term mix development, so I’d encourage anyone serious about commercial music production to start using locator points in their own mixes to do the same kind of listening comparison between different choruses that I’ve just shown you here, because it really helps focus your ear on exactly how your long-term mix dynamics are working. It’s such a powerful listening trick, and I personally use it on at least 90 percent of the mixes I do.
One final little tidbit: I can’t help feeling that the vocal line at 0:15-0:26 sounds a bit weird, on account of there being no breaths at all. That’s a long time for anyone to sing without breathing, and while it’s definitely within the bounds of possibility that Carpenter could sing it that way, she certainly didn’t do so in her live performances at the Grammies or on SNL (punctuated by three and two breaths respectively, and all in different places in the lyric too), so I’m dubious about any serious artistic intent there. Fundamentally, hearing this kind of thing on record always feels subliminally a little claustrophobic and disturbing, and I’d personally avoid it if I were you – unless that’s the effect you’re actually after!
In a sense this feels very much like an old-school Akai MPC production in the way it’s based around a core eight-bar ‘construction kit’ (which you can hear in full at 3:00-3:19) comprising a triggered drum-sample pattern (primarily a combination of distorted kick drum and what sounds like a noisy cymbal-tail) and a handful of heavily processed and edited four-bar sample loops featuring bass, electric guitar, organ, and trumpet sounds – although the last two are so heavily messed with that it’s hard to be 100% sure of their provenance! Now, on the face of it, this approach might seem rather limited in its ability to generate textural variety or arrangement build-up, but this production deftly illustrates how to get masses of mileage out of such comparatively modest means.
For a start, if you look at the maths, there are 63 possible ways you can combine our six construction-kit instruments (kick, cymbal, bass, guitar, organ, and trumpet), and I counted nine different combinations on display in this case. However, what the producers have also done in this case is create two different variants for each of the bass, guitar, and organ loops. So the four-bar bass riff (first played at 0:10) is edited down to a single-note alternate line (eg. the following four-bar section at 0:20); the guitar loop at the start of the song features a rising line, whereas its variant (eg. at 0:10) just repeats a single high note; and the two organ loops appear to be just different collections of retriggered and manipulated snippets. The mixing then builds on this concept by creating further variations with processing, for example the duller filtered version of the bass riff heard at 1:00, 1:40, 2:30, and 3:50, or the more distant ghostly version of the guitar’s high-note variant that accompanies the later appearances of that filtered bass.
Extra arrangement variety also comes from loops that are muted just before the end of a given four-bar section to generate a kind of ‘drop fill’. So, for example, the last four-bar section of Pusha T’s verse (1:20-1:29) drops out the kick, cymbal, bass, and organ, leaving only the trumpet and second-variant guitar loops. I counted at least nine occasions where this technique was employed – there are three just in Malice’s verse (1:40-2:20), for instance. And all that’s before we even consider the five different voices that perform over the track!
Speaking of the voices, even though the lead rappers sound pretty dry in this mix, you can still hear them clearly in the stereo Sides signal, indicating that they’re being stereo-widened in some way. Given that it doesn’t sound like reverb, and that the waveforms of the Middle and Sides components during the vocal fill at 1:07 look pretty much phase-aligned, I’m guessing we’re hearing some kind of EQ-based widening technique, and (based on the tone of the sides signal) probably one that’s only operating above about 200Hz.
But that’s not the only interesting thing about this mix’s Sides signal, because you’ll also notice that the sub-80Hz low end of the kick drum and bass are also present, which is a little unusual because a lot of mixing engineers prefer to keep those frequencies effectively in mono to maximise their power on playback. That said, even though stereo localisation of bass isn’t great down at those frequencies, phase decorrelation can still add more of a sense of ’envelopment’, so there’s a trade-off to be considered here.
And finally, the Sides signal also features plenty of upper-spectrum distortion components from the kick drum, and although they’re triggered by the kick drum, what’s fascinating to me is the way they actually seem to act almost like a layer of upper-register percussion, filling a role that’s otherwise left pretty much vacant by the generally sparse rhythm programming.
One of the fundamental truths of audio mixing is that human tonal perception is adaptive – in other words, we tend to adjust our perception of any mix tonality we’re hearing to match our preconceptions of how it should sound. This is why referencing is such an important part of the mixing process, because doing A/B comparisons between commercial releases and our own mix work goes a long way to defeating this natural perceptual adaptation, thereby revealing the disappointing truth that our my mix has, for example, waaay too much 4kHz, but that I’ve just not noticed because my ears have adapted to that!
And this aspect of the human listening experience presents mainstream mix engineers with a quandary. What if the track before mine in a playlist is, for example, unusually bright and subby, and therefore makes my mix sound boxy by comparison, simply because listeners to the previous track have gotten used to its skewed frequency profile? After all, I wouldn’t want to make my mix any brighter and subbier than it is, in case it follows a more normal-sounding mix in another playlist, which would then make it sound hollow, woofy and tizzy.
Well, one solution to the problem can be heard in this Lady Gaga song, which starts with the main hook-section texture (always a smart commercial move in itself), but with some kind of band-pass filter over the whole mix, removing effectively all the frequency energy below 200Hz and above 4kHz. Pretty much no matter what precedes this song in a playlist, this song’s intro will always sound boxy (but clearly on purpose), and the seven seconds that the mix tonality stays like this provides ample time for the listeners to forget the tonal profile of the preceding song, but not enough time for their auditory system to make such a middly new tonality into the new normal. So when the band-pass filtering is suddenly removed at 0:08, the unfiltered texture now sounds like it has plenty of exciting-sounding low-end and high-end energy by contrast with the intro section, even if the previous song’s spectrum was actually smilier. Think of it as a kind of audio palate cleanser!
Another nice little production technique is well exemplified here: abruptly muting all the background parts to make some foreground feature suddenly pop out at the listener. The specific moment to listen for is the fill at 0:14, where pretty much the whole texture is muted for the tom fill, and that fill is then itself muted to highlight the following beat’s vocal/synth stab. And the reason why this is such a cool example is that at 0:22 you get to hear exactly the same fill+stab combo, but without the backing mutes, and comparing those two moments really highlights how much more attention-grabbing those musical features become when they’re not having to compete with other backing parts.
This is a song that really capitalises on the musical power of harmonic rhythm – ie. the rhythm with which the chords change. A surprisingly large number of songs maintain a consistent harmonic rhythm throughout, which naturally means that this characteric is common to a lot of the AI-generated music that’s now flooding the internet, so it’s worthwhile investigating alternative possibilities here if you’re going to set yourself apart from the advancing legions of robo-tunesmiths.
In the first instance, the simple fact that the primary chorus texture (first heard at 0:51) just speeds up the harmonic rhythm from its slow preceding two-bars-per-chord pace immediately adds a sense of extra forward impetus. But we don’t just get an increase in the pace to one chord per bar (as plenty of other songs do), but instead we get a much more ear-catching alternation of dotted-half-note and quarter-note chords, made all the more propulsive by the powerfully rising scalar bass line.
Not content to rest on their laurels, though, the producers here double down fabulously on this tactic for the final chorus section by combining it with another tried-and-trusted technique: the ‘drop chorus’. In other words, from about 2:40, the arrangement feels like it’s building up towards another big full-texture downbeat for the second half of the chorus at 2:43, but then instead we get a surprisingly sparse bar of just gospel choir to start with (the ‘drop’ of the ‘drop chorus’), before the band rejoins the melee a bar later. But rather than beginning bar two with the expected second bar of the harmonic progression, the band instead squishes together the bar-one and bar-two chords at an even brisker quarter-note note rate, thereby compounding the arrangment impact of the drop-chorus’s delayed full-band entry with the extra musical excitement of the faster harmonic rhythm.
I’d use the word ‘synergy’ now, if it didn’t make me sound quite so much like a tech CEO.