Composing From a Model: Part 5

So, working from scaffolded composition activities for loops, riffs, and chord progressions, and with my additional work on vocal and lyric writing, I was now firmly enmeshed in studio time. Thankfully I had all the gear at my house, and so I was able to record over a few days.

Working in the Studio

My studio time with my band has led me to a few helpful points that I used to order my recording process.

The first one was about the order of recording. In my experience, it's best to record drums, then bass, then rhythm instruments, lead instruments, and vocals last. In a lot of popular music, drums and bass establish the rhythmic framework for everyone else to work in, which is why it's so important for drummers in particular to be able to play in time (though I would argue that it's almost as important for everyone else to be able to play in time).

For me, I decided to just use drum loops that were converted into WAV files, so drum recording was relatively painless, although I had to clip them slightly because they didn't quite match with the metronome. I actually didn't have bass for the first 2 minutes; that was an intentional decision based (no pun intended) on following Radiohead's model of macro-tension. I used a synth bass for the parts that were there, and worked off an improvisatory framework of following the chords but playing fairly melodically.

I recorded most of the other synth parts in one day, including the rhythm and lead lines. With each synthesiser, I took one of the preset sounds and then altered the filters, gain, overdrive, reverb, delay and EQ to create the sounds I was after. It was particularly fun turning the first sound of the song from a cheesy MIDI trumpet tone (like this) into something more workable, through messing with the envelope filter. It was also clear once I entered Logic that some of my harmonic ideas needed tweaking based on the sounds I'd chosen. It was nice to have this interplay between my ideas and technology, as both the ideas and the technology were able to be moulded to get a better musical outcome.

As you can see below, I also colour-coded the different sections of the piece so that I had a clearer idea where all the parts were.

Guitar Parts

Once all the synth parts were added, I recorded electric guitar using a single Rode NT1 in front of my amp. The rhythm part was fairly improvisatory, but was built on a few melodic ideas I had over the minor chord progression that I wrote earlier. The lead electric part was made by me trying to find the closest notes possible to move between each chord. It wasn't something I theoretically worked out; I used my ears after playing the first three notes, and then tried to find the right sounds for the other chords. I beefed up the overdrive for the lead part, both with my amp and in Logic.

Vocals

Once I reached the point of vocal recording, I had lyrics, melodies and harmonies decided for the verses and choruses, but nothing for the bridge. I tracked all the other vocals first (using the same NT1), using another helpful thing that I learned, both from the studio and also research.

I'm afraid I can't find the citation, but I read about a study that was done on a number of people who were regarded as "tone deaf" i.e. they could not sing the same pitch as what someone sung to them, or the pitch of an instrument. The study found that through a period of a number of weeks of intensive work with these tone deaf people, they were able to get empirically testable measurements of improvement. For one thing, it busts the myth of there being a certain percentage of the population who are just tone deaf for life with no hope of ever improving. But the interesting thing about the study from a music education perspective is that the method the scientists used most was recording people's voices and getting them to listen to it back.

Interestingly, when I recorded in a professional studio space with a well-known Sydney producer, he used the exact same technique with my band when we recorded vocals. Band members would sing, and then he would ask them to listen back, before making a comment and then getting us to re-track lines. I think it worked quite well for us.

So... when I recorded my own vocals, I made sure to listen to them back before re-tracking or going on. I think, again in part due to time constraints, these vocals were not my strongest, but by using the same technique as the producer I worked with, I think the results were improved.

While I recorded I also experimented with different vocal tones, and so this song featured vocals with more of a twang than how I had sung previously. Vocal tone is a whole other can of worms that I've been thinking about this year, but generally the aim was to attempt to communicate emotions that don't really fit into the bel canto style of singing that is popularised in opera and many choirs.

The Mystery Bridge Vocals

I arrived at the bridge vocals with an idea to sing long clear notes, and that was all. Given the parallel minor chords in the progression, I had to be a bit creative about finding a melody that would work. In the end I basically just sung over the top of my other recorded parts, and the words just came to me. Maybe you could say they were a part of the flavour of the song (dark and moody).

Once I had the main vocals in, I just experimented with singing lines over the top until I ran out of range. I ended up with three harmonies, but after recording them all in I decided the top two were a bit too full on for the first half of the bridge, which hasn't hit its full stride yet. So I cut them out for the first half.

After that was finished, it was just a matter of some more mixing, and then I was all done!

(You can listen to my finished track here)