Posted on Leave a comment

Do, Re, Mi, me, Mi

I decided I want to sing more.

One of the things I notice about singing in other countries / cultures is that people just do it. They don’t necessarily worry about the quality of their voices or pitch perfect tuning – the important thing is to sing, to express themselves. The current Western pop machine model we most commonly come into contact with is all about (post) production. Artificial processing techniques such as auto-tune, time-shifting, compression etc are available in any studio – the sonic equivalent of air-brushing. It’s not that the results aren’t pleasing, or shouldn’t be used, but it can become a problem when this approach is held up as a norm and anything outside these parameters is seen as sub-standard and should be disregarded. X-Factor and its ilk compound this problem by tearing people to pieces if they don’t sound identical to this narrow model. We have created, and for the most part accepted, an artificial ideal that requires you to remain silent and hang your head in shame if you don’t match up.

I decided to take a different approach.

To deliberately explore imperfection.

The idea behind the piece Meaning Without Talking was just to record myself singing and see what happened. I’m going to walk through how this came into existence and have included four versions during this article to illustrate parts of my process. Obviously the final version is the one that I’m happy with; the three before are to help illustrate the journey.

My process began with vocal improvisation – just for myself – as I woke up, got dressed and wandered around the house.

By the time the computer was switched on and Cubase loaded up I was ready to record. The melody wasn’t entirely fixed at this point but I knew more or less the shape of it, I had also decided that I wanted an interlude section but I didn’t know what that was going to be yet. I started recording but as I got to the end of the melody I wasn’t convinced that version was good enough, so I kept recording, left a gap and recorded it again. At the end of this I improvised an interlude line and then repeated the melody to finish. Listening back I thought maybe I’d keep all of it and perhaps add an instrumental following the first time through the melody. Although I hadn’t planned a structure for this piece, my own habits from years of listening and playing kicked in, and I ended up with an AABA structure.

My only initial rules were to come up with a melody quickly and be as natural as possible; I hadn’t decided on a harmonic structure but one emerged anyway. I know from past experience that if I start with chords my melodies tend to be a little aimless, I get distracted by jamming over the chords instead of focusing on the repetition necessary for a strong memorable line. Without planning, the melody contained variations and answering phrases based on the first phrase. There aren’t any amazing melodic surprises but I was quite pleased with it anyway. I liked the sound of it and melodically it describes a harmonic progression without needing chordal reinforcement for clarity.

Without giving myself rules to disrupt my habits, ones I didn’t realise I had asserted themselves anyway.

My intial solo vocal sounds like this :


[soundcloud url=”” params=”auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&visual=true” width=”100%” height=”100″ iframe=”true” /]

I now had one voice, not brilliantly sung, but I decided not to let that bother me. Having been doing some vocal experimentation for the Snow Giants album I’m working on (where I’ve recorded a four part choir, with three voices to each part for reinforcement), I decided to push that idea further and record twenty parts this time. I didn’t want to write down the melody but instead see if I could use the folk tradition of just singing along to pick it up. As I’d already heard this a few times I knew more or less how it went, but it was still not completely fixed in my head.

As I recorded the second and third voices these were a little more uncertain and I made noticeable mistakes as I strained to hear the first voice clearly and remember exactly what I’d sung the first time around. I hadn’t recorded with a click track or any guide note for pitching so that all needed to be judged by ear as I went along. This process is what would naturally happen with a large group learning something new, but because there’s only one of me I was artificially recreating problems that I didn’t really need to have. My thought behind doing it this way was that it should help create a more convincing impression of the ‘real thing’ when I was finished. I’d also made the decision that my interlude section would be a freshly improvised line every time I got to it.

The three voices sound like this :


[soundcloud url=”” params=”auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&visual=true” width=”100%” height=”100″ iframe=”true” /]

I didn’t want to ‘fix’ my mistakes but instead wait and see what happened if I just kept recording. A couple of coughs had crept in, but I wanted to leave them there for authenticity.

Subsequent voices gradually became more confident as a consensus formed on the phrasing, timing and pitch of the melody. Initial parts of the interlude were very rhythmically simple and harmonically unadventurous. With no chord sequence to work from I then began to try to establish some harmony.

Subsequent interlude parts became more elaborate – first harmonically, then rhythmically – with more moving parts and suspensions. Further into the process the rhythm started becoming less clear again so I returned to long notes. I spent some time taking more risks with register switching. I established two points of ‘harmonic rest’ where the chords simplify and hold. I doubled the bassline to try and add some harmonic strength at the bottom but all other parts were ‘fresh’ each time.

The twenty voices sound like this :


[soundcloud url=”” params=”auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&visual=true” width=”100%” height=”100″ iframe=”true” /]

So having recorded this twenty times, a consensus had gradually formed. A syllabic way of singing the initial scale movement in the first phrase developed over the performance, re-established into a habit after 20 recordings. I found myself singing the same way because ‘everybody else’ does, even though ‘everybody else’ is still me; I wanted to stick with the crowd and enjoyed doing so in this instance

Allowing the coughs to stay on the recordings made me want to cough too when I heard them – so several layers of coughing gradually built up as a response to previous layers. I also became aware at some point that I was watching the screen, where the wave forms of previous voices were providing visual clues to timing.

Voice Wavs
I became aware at some point that I was watching the screen for visual clues

I closed my eyes to record once I noticed what I was doing. I also noticed that one version of me was always rushing the phrases but that’s how it was so I left it in.

To increase the ‘live’ impression of the recording I then panned the 20 versions of ‘me’ 10 degrees at a time across the stereo field to try and exaggerate the effect of being in front of a choir. There was also only the reverb from my living room on the raw recording so I added an artificial one : ‘Church Hall Reverb’. This instantly created a different atmosphere and helped mask ‘early me’.

The coughing seemed pretty manic by this point so I initially tried reducing the volume of particularly loud coughs but it was still overwhelming. I didn’t want to lose it all as it maintains the illusion of being in a live space – people do cough – but to wipe it all seemed too sterile. I ended up taking most of the coughs out, leaving just one or two. I think it helps make sense of the space. We expect a certain level of random ambient noise in day to day life and it can seem strange without it, even if we aren’t consciously aware of its presence.

My final consideration was that the pulse during the improvised interlude was, unsurprisingly, quite loose. I decided to edit it slightly to conform more with my personal taste and removed some of the more uncertain notes. Although I had given myself a remit to be as natural as possible at the beginning of the exercise, this ‘rule’ was less important than ending up with a piece of music with which I was satisfied.

Here’s the final version :


[soundcloud url=”″ params=”auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&visual=true” width=”100%” height=”450″ iframe=”true” /]

First published on The Aware Musician Sept 25th 2014

Leave a Reply

Your email address will not be published. Required fields are marked *