I wrote a few weeks ago about creating a version of Sandy Denny's song, "Nothing more". I am working, albeit slowly, to a full cd of SD covers, and the pleasure that I received from creating NM (and the positive responses) encouraged me to find other songs to work on.
It's not very easy trying to create personalised versions of Sandy Denny songs. Frequently they are in strophic form - one of my favourites, "Full moon" has five verses, each consisting of 24 bars, and most of the material on her "Old fashioned waltz" is similar in structure - which makes it hard on the arranger to find an interesting way of performing the material. Some of the earlier songs, though, are in different forms which makes them attractive. One song which I started looking at, "The sea", is strophic, but the micro-structure within each verse is far from straight-forward. At the moment, I have put this song to one side but I may well return to it.
I decided to try and sequence a version of the title track of Sandy's first solo album. I've probably mentioned this several times beforehand, but it's worth repeating: I bought this record on its release at the beginning of September 1971 and it had a very strong influence on how I viewed harmony. The title track has long been a favourite of mine but I've never really felt the urge before now to create my own version.
In terms of macro structure, there are three repeats of a structure composed of two verses and a chorus. The verse is fairly straight-forward: although it starts on a minor chord, it swiftly moves to the relative major: (in E minor) Em | G | D | C | C | G | D | C | Em | Em; all of these chords are in the key of G. The chorus, though, is something different; it starts out by modulating to the major subdominant of Em (or the supertonic of G) : A | G | C | A | A | G C | A | D | Em | Em. But the strength of the chorus is not only in its harmony but also in the bar lengths. One bar has two chords in it, but more importantly, both the third and the seventh would appear to have a beat dropped from them (in other words, the time signature changes from 4/4 to 3/4 for those two bars). I had always been aware of a strong accent on the fourth beat of the fourth bar; it was only when I was creating a chord map that I realised that this would better be represented by having a shortened bar. The second shortened bar arose when I tried singing the song; again, the accent was falling in 'the wrong place'.
My MIDI sequencer doesn't have a problem with multiple time signatures (although it is slightly difficult to copy such passages) but Reason can't handle them (or rather, ignores them). This is a problem when using drum loops, but this song required no drumming.
I worked out an interesting beginning (Em7|Em6|Em7|Em6) which made pitching the first note problematic but this could be fixed. Over this intro, I had a french horn playing a lick; a different lick was used for all the transitions between sections (two bars of Em) with the lick varying on each invocation. The french horn also played a solo, but after hearing this many times, I got bored and decided to transfer the solo to a different instrument. After 'auditioning' several instruments, I decided on a concertina, which certainly adds a different sound.
Singing the song went fairly well, although as usual it took several attempts to find a good vocal sound. I ended up using a fairly strange - for me - equalisation setting, with a boost at 200 Hz as well as the more conventional 2 KHz and 5 KHz. At the moment, there's quite a deep reverb which gets hidden by the music (that's why I allowed myself a deeper reverb than usual).
After listening for about a week, I hit on two improvements. In the link between the final verse and chorus, I used a triplet chord stream (ie Em F#m G) to lead into the A chord: this makes quite a difference. I also noticed that my vocal on the final chorus was slightly out of time; instead of rerecording it, I was able to lift the second chorus and paste this in at the end. To my pleasure, the edit worked the first time and I was spared the agony of inserting (or deleting) milliseconds of silence.
Having written all of this, I am struck by how much easier a musician has things in the 2000s: instruments can be inserted and/or exchanged willy nilly, bad notes can be fixed, solos can be changed, vocals can be moved in time and more importantly, put into tune. I am awed and humbled by how much harder it was in the 60s and 70s to create music: for that, the musicians had to be good and get things right almost first time. There was little room for second thoughts, which meant a great deal of pre-production. Nowadays, there's little pre-production but plenty of post-production.