From time to time pop artists release not only the full mix of their single, but also instrumental and a-capella versions. As long as the increased scrutiny this places upon the production values doesn’t undermine the artist’s or producer’s reputation in some way, there’s a lot to be gained by this, as I see it. Fans who fancy a spot of karaoke, for instance, will likely buy/stream both the full-mix and instrumental versions, immediately doubling the revenue per song (ker-ching!), and an a-capella version provides an open invitation for up-and-coming electro-musos to promote the single organically (at zero cost to the artist, of course) by sharing their remixes on social media.

For us studio jockeys, though, these alternate versions provide a nice opportunity to more closely examine the vocal and backing-track sonics. In this case, it’s interesting how much of the vocal brightness is coming from the powerful frequency peak centred at 10kHz – notch that out and the remaining vocal tone feels surprisingly dull. With such a strong emphasis at that frequency, it’s little surprise that any vestiges of lip noise have been carefully edited out and that the breaths have clearly been supressed and dulled a great deal too, as all of these would have become pretty distracting otherwise. By comparison, the backing track is generally pretty restrained at the high end, which stands to reason if you want to keep the vocals way out front – it’s contrast with the backing track, not just sheer HF enhancement, that counts here in subjective terms.

If you’re after an impressive sense of stereo width in your mix, but don’t want to make a nonsense of the music under mono listening conditions, one sensible tactic is to over-widen some peripheral arrangement layers that don’t carry any primary musical content. And you can hear this approach in action here with the backing-vocal ‘hah-hah’ layer first heard at 1:54, where the level Sides component is 2-3dB more than that of the Middle component. That said, that vocal part still feels quite hooky to me, so I’d have been more inclined to choose a synth layer instead.

Another aspect of the vocal sound you can really nerd out on while listening to the a-capella version is the effects, which, of course, are normally very difficult to hear when the whole mix is going on. The first thing to notice is how the verse lead vocal’s long, expansive reverb treatment is partly a result of quarter-note tempo-sync’ed feedback delay. Both effects are also quite dull-sounding, again a sensible decision if you’re hoping to keep the singer super-upfront in the mix. Beyond this, though, the real takeaway from this mix for me in terms of effects is how often the effects are changed – a ‘restlessness’ that’s very much a hallmark of a lot of mainstream chart music. Check out, for example:

• the eighth-note delay spin on “love” at 0:55 (and again at 1:53)
• the 3/16th-note(?) delay with quarter-note gain-pumping on “love” at 1:00 (and again at 2:06)
• the brightening of the second verse’s quarter-note delay at 1:05 (relative to the duller tone of the first verse)
• the eighth-note triplet high-feedback delay on “good” at 1:35
• the boosted delay/reverb levels in the middle section from 2:10, and especially after the word “night” at 2:27
• the gain-pumped reverb after “me” at 2:30
• the long filtered delay after the final “love” at 3:24

…and that’s quite apart from all the different backing-vocal layerings, from multi-part harmonies to glassy-sounding layered double-tracks (eg. “wait for your love” at 0:55).

One aspect of the vocal mix does puzzle me slightly, though: why is the lead vocal about 3dB quieter in verse one than it is in verse two? I mean, producers have often mentioned doing the opposite, in other words making the first-verse vocal entry a little louder to grab listeners’ attentions from the get-go (I believe Geoff Emerick used to do this with the Beatles singles back in the day.), but the other way around is an interesting call. I initially wondered whether this might be a ruse to bring down the song’s overall average level in order to increase the subjective loudness of the main hooks on the streaming platforms, but my own experiments with a loudness normalisation plug-in suggest that boosting the verse-one vocal level by 3dB (to match the second verse) would have made a negligible difference in that respect. And I can’t imagine it’s some kind of oversight either, with Max Martin’s name on the tin, so what gives?

Well, my best guess at the moment is that the idea is to draw listeners in with the slightly lower vocal level, so that they either pay greater attention, or simply turn up their playback systems during the first verse, so that they when the first chorus hits it makes more of an impact – in other words, using long-term dynamics as a power play. This is something that many producers have kind of forgotten can be a thing, after so many decades of hyper-compressed radio/TV-broadcast dominance, so it’s an interesting to speculate that A-listers might be taking it more seriously these days now that loudness normalisation of streaming audio is pretty much the norm.

And, finally, one fun bit of serendipity! I was just in the middle of writing a post for my Cambridge-MT Patrons giving a bunch of examples of songs which draw inspiration for their songwriting, arrangement, or production directly from their lyrics, and had been drawing a complete blank about whether there were any other songs than Alanis Morrisette’s ‘All I Really Want’ that follow the word “silence” with an arrangement vacuum – until I heard the one in this song at 1:09 here! Talk about timing…