It should be noted that I have tried my best to make this blog post accessible to the general reader. I have tried to avoid technical music theory jargon and flowery, arty-farty language as much as possible. This is the story of the composition of this piece, revealing my processes and thoughts at each stage, put as straightforwardly as I can. I hope that you find it interesting...
There is a link to the web page containing the recording of the composition at the end of the blog, in case you want to skip ahead and listen to the piece first. Or just listen to the piece and avoid all the detailed description. Up to you...
This piece started after a chance conversation with Simon Taylor - a friend and also my bandmate in Dr Zebo’s Wheezy Club. Simon is also very involved in nature conservation, particularly in his home town of Warminster. He is a member of the ecological group Sustainable Warminster and was organising a series of new compositions created around recordings of bats made in and around Warminster to publicise their presence in the area. These recordings had been slowed down to bring the pitch of the bat calls down into the range of human hearing. Slowing them down also allows the detail of the calls to be appreciated. I volunteered to write a piece as part of this project because I liked the compositional challenge this would involve and also supported the cause.
My first thoughts were to create a kind of ‘soundscape’ featuring the bat calls accompanied by strings. As this piece was written in lockdown here was no opportunity to use real string players, only synthesised string samples on a computer. I have become interested in writing for strings for two reasons; as they are the backbone of an orchestra and therefore it is a necessary skill for a composer, and also, following eight years of listening to and writing big band music I need a change! The composing was done on a mac laptop running Logic software and also a very basic midi sound source - an ancient Roland Sound Canvas.
So according to the original concept, the piece should have consisted of just two distinct layers, the bat calls and the strings. (NB: This actually became three layers later on. Read on!) Having listened to many of the bat calls I chose a recording of a pipistrelle bat and then cleaned up the sound by filtering out (in Logic) a lot of the extraneous noise and hiss in the original field recording. I copied this track and pitch-shifted the copies of the recording, eventually ending up with three layers of similar but slightly different-sounding bat calls which entered one after another to create a kind of musical canon. The sound of one bat building up to three bats and then fading back to just one bat gave a pleasing sparse-busy-sparse texture and overall structure to the piece. At this point in the composition process, the title The Last Pipistrelle also presented itself, as the final single bat disappeared into the distance at the end...
To this soundscape of bat calls, I wanted to add a layer of strings but not a conventional string arrangement. The strings should float along without any sense of pulse but still be unpredictable enough to maintain the listener’s interest throughout. To achieve this I used a couple of techniques that I had come across as part of my MA studies in classical composition.
Firstly, I composed five ‘chords’ that I liked. Let’s call these 1, 2, 3, 4, and 5. The numbers also correspond to the numbers of notes in each ‘chord’ (Yes, I know that 1 and 2 notes don’t make a chord, but please bear with me) and finally to the number of beats that each chord lasts. I also added a ‘chord’ 6 which was 6 beats of silence(?!). The only ‘rule’ in choosing the notes in the chords was that the top note of each chord rose by a semitone - chord 1 = G, chord 2 = Ab, etc. The rest of the notes were chosen freely and follow no plan or conventional chord progression. I just trusted my ears! I had created a row of 5 chords which gradually got longer, thicker and seemed to rise up (at least the top note rises). This was followed by chord 6, six beats of silence. Let’s call this: 1, 2, 3, 4, 5, and 6.
The late British composer Sir Peter Maxwell Davies had first used a technique in the 1970s which derived from the permutations used in the British tradition of change-ringing church bells. Here the bells start playing in one set musical pattern and then the bell-ringers gradually change the pattern over time following a set of rules for developing the original pattern. These rules vary depending on the particular ‘piece’ they are playing. If I apply this technique to my set of chords, eventually the order of chords reverses (and if continued further, they come back around again to the original starting order). So, if I repeat the six chords but move on to the next permutation each time that gives me the following pattern of chords. (I’ll leave you to work out exactly how the rows permutate!):
1 2 3 4 5 6
2 1 4 3 6 5
2 4 1 6 3 5
4 2 6 1 5 3
4 6 2 5 1 3
6 4 5 2 3 1
6 5 4 3 2 1
Eventually, I end up with six beats of silence followed by 5 chords which gradually get shorter, thinner and seem to descend (at least the top note falls). This is the mirror image of the opening set of chords. Again, another pleasing compositional shape.
The final piece of compositional ‘fun and games’ is a technique that I learned from Philip Cashian - my composition teacher for my MA - that involves creating a palate of similar chords from one initial chord. This is done by keeping the same number of notes but swapping the positions of the ‘vertical’ gaps (or intervals) between the notes in the chord.
Here is a simple example using a chord of C. This chord has three notes: C, E and G. The gap (or interval) between C and E is 4 semitones, and the gap between E and G is 3 semitones. So the original chord of C, in terms of intervals, is 4,3. If I keep the same intervals but swap them over (3,4), I get C, Eb, and G. (C to Eb = 3 semitones and Eb to G = 4 semitones). So I now have two chords with the same ‘intervallic content’ that sound similar but slightly different. The more notes in a chord, the more possible permutations, and the greater the number of similar-sounding chords generated.
Cashian uses this technique as part of his pre-compositional process (but with much less conventional ‘chords’), to generate possible musical material to use in each new piece. He then treats this material merely as a starting point and is very free in the way that he plays around with the resulting groups of notes in the final composition.
In my composition, the result of using this technique is that every time chords 3, 4 or 5 appear they might be in their original form or in one of the permutations sharing the same intervallic content. Again, I just trusted my ear… As a result, each set of six chords sounds similar but slightly different and appears in an unpredictable order with the ‘silent chord’ appearing at any time, apparently at random. All of this keeps the listener guessing…
With all this raw material and a basic plan, I composed the string orchestra parts using Sibelius notation software. I decided that the dynamics of the string chords should align with the sparse-busy-sparse outline of the bat recordings by starting quietly building up and then dying away. This arc is the basic form of the final piece.
When I listened back to the first version of the strings layer alone I spotted two problems. Firstly, I decided that the flow of the chords was still too predictable, so I broke up some of the long note textures with plucked (pizzicato) notes. Secondly, the slow six beats of silence (chord 6) created too long a gap and this caused the piece to lose momentum entirely. I filled the gap with a single high note that changes in length throughout the course of the piece. This single held note also worked well as an introduction to the string chords, and as an ending.
Up until this point, the two layers (bat calls and strings) had been entirely separate, so I imported all the files into Logic, and shuffled them about against each other until I was happy with the way the two textures aligned. As neither layer has a strong pulse, this was purely by ear, gut feeling, and trial and error. The raw sound of the bat recordings and the Sibelius string samples were improved by adding reverb, filtering, changing the envelopes and eq-ing, and in the case of the strings, by adding copies of each part played on the much better string samples available in Logic. The individual parts were panned left and right to provide a wider, bigger soundscape and the complete track was finally mixed to make everything balance.
I think of this process as painting with sound, and that includes all the false starts, changes of mind, failed experiments and alternative versions we are familiar with from the visual arts world. The final piece evolves out of the initial sketches, and I start the process without knowing exactly what will emerge. I have an inkling of what I’m after, and hopefully, something close to that ‘sound-picture’ in my head gradually comes into focus over time. In this case, I wanted the bat calls to sit against the evolving string backing but neither layer should dominate the music. The strings definitely mustn’t distract the listener from the stars of the piece - the pipistrelle bats.
At this stage, I sat back, closed my eyes and just listened to the piece from start to finish, several times, to see if I thought it ‘worked’ and if I was happy with the outcome. I wasn’t… Something was missing. It needed another layer of something. Some other sound in the mix to make it ‘come alive and sparkle’. But what, exactly…?
This led to three very frustrating weeks of trial and error. I tried a big cello tune. No! A layer of random percussion sounds. No! Some piano flourishes - getting closer… Eventually, I discovered the slightly distorted piano sound in the banks of keyboard samples in Logic. I just improvised a layer of piano ‘stuff’ against the backdrop of the bats and strings, just having fun, with no critical filter on the quality of the playing and the success of the ideas. The good thing about digital recording is that, unlike when you improvise jazz, you can easily go back and edit improvisations and fix mistakes and bad decisions. By deleting notes, copying and pasting phrases and moving sounds around I eventually came up with the keyboard part on the finished version. Phew!
I listened to it all playing back together and found the missing ‘sparkle’ had appeared. The piece was finally finished.