How to pick the correct key for an arrangement is a core part of our basic craft as arrangers, and there are standard ways of going about it. To begin with, there are the ‘typical’ ranges for each part, and you’d start off by seeing which key would leave the melody lying in the classic range for a lead.
There isn’t always a single obvious answer from this initial process, though. Very rarely, a tune sits well within the standard range, giving room to move it up or down; more often it is going to spill over at either top or bottom. The decision at this point is inflected by multiple considerations that may not all point to the same answer. Clearly there’s the question of who is going to sing it: preferred voice range is something I’ll always ask for a commission, and I prefer it if I know the sounds of the actual voices too. It’s always good to keep your lead happy, after all.
Then there’s how the harmonies lie in different keys. Again there are the practical considerations of the actual voices you are working with: the lie of the bass line is clearly dictated by the range of the ensemble. But there’s also the sense of how the voicings are going to work out, and the baritone is the key player here. Baritone lines rarely stretch a singer’s vocal capacity – because they fill above and below the lead, they are nearly always working within a smaller range – but they have a huge impact on the feel of the harmony. A tune can lie quite happily in either of two keys a tone apart, but one of those keys may produce a much better baritone line than the other – one that gives colour and texture, not merely fill. It has been a reasonably common experience for me over the years to pick a key according to the dictates of the melody, then change my mind when I discover how the voicings are shaping up.
And then there is the whole thing about the expressive character of different keys. My hunch is that popular music traditions don’t have the same depth and consistency of classical traditions in terms of the kind of music that appears in different kinds of key. But to cater for those people who do have a strong sense of key and character, I’d tend to put gentler music into flat keys and perkier music in sharp keys, assuming that everything else was going to work out vocally.
But the things that has really been rising to the top for me in recent times is the question of the original key. This is less an issue for music whose primary form is written – the performance history of 20th-century popular music shows a strong tendency to transpose to suit the singer – but it does make more of a difference for music where the definitive version is a recording.
This is an issue for both listener and for the singers. Now I’m sure most people would argue that the average listener wouldn’t notice if you sang a song in a different key than the one they’re used to hearing it in. And I’m sure that’s probably true at a conscious level. But equally, I think they’ll have a stronger intuitive sense of recognition if they’re hearing it in the same key they’ve heard it in before. This was a major issue in my arrangements for Noteorious for the Radio 1 Masterpieces project – their versions were to be broadcast back-to-back with the originals, so I wanted to minimise the tonal jolt for the listeners going from the famous recordings to the a cappella version. But I think it exists in a dilute form even when there’s no opportunity for direct comparison.
And the reason I think this is because the original key also has a meaning for the singer. Anybody who chooses to sing a well-known popular song in an a cappella ensemble is likely also to have sung along with the song on the radio, whether absent-mindedly as they did something else or with great gusto as they revelled in the song. So, they will be used to the song fitting in that part of their voice, and to finding their own special meanings in the music as it sounds in that key. Arnie Cox has argued reasonably convincingly that a lot of the pleasures of musical listening come from covert mimesis – from mentally joining in with the performers – and this will be as true of audience members’ relationship with the a cappella version as of the singers’ relationship with the original.
So, if the ‘correct’ key in terms of the standard methods discussed at the start of the post is anywhere near the key of the original recording, these days I will do my damnedest to arrange it in that key. It may take some jiggery-pokery if that threatens to take the lead out of their best range, but actually you’ll often find that a lead who thinks that x note is too high for them will reach it quite happily in the context of a song they have sung along to. And anyway, jiggery-pokery is one of the jobs of the arranger – the other parts rarely seem to mind taking over bits of the tune if that becomes necessary. If the original key is going to be just impossible, of course, then you revert to the standard methods. Though even then, it’s good to keep the song in a related key – so it’s tonally close to the original, even if intervallically distant.
I’ve been expressing the connection that people have with the original key in quite idealistic terms: it speaks to them because they have both aural memory and motor habits of identifying with the song at that pitch. There is a practical benefit too, though: that habitual response is going to make it easier for singers to stay in tune in that key than in any other.