Lavender Earl Gray (Alexander Dove, 2017)

Lavender Earl Gray (Alexander Dove, 2017)


This song is the B-side to Lonesome George, a single I put out at the end of July. I think arguably the single itself has some production / mixing mistakes that I hope to improve on in the eventual album version – it’s all a learning process, after all. This B-side, however, was a lot of fun and involved a lot of embracing happy accidents.

First of all, the distorted, almost 8-bit sounding bits at the beginning and end are from my first attempt at recording the guitar, when my Focusrite Scarlett 2i4 interface was having an error that it would often have, where audio recorded through it was crushed in this way until I unplugged and reconnected it. This seems to be something that happens when these interfaces are used just for input and something else (in that case, my laptop’s built-in soundcard) is used for output. Ever since I’ve had my speakers connected to the Focusrite, there’s been no problem. In this particular instance, I was tickled by the distorted audio and felt the need to preserve some of it  for the intro and outro.

I ended up not recording the song until much later, and recorded guitar and vocals together with a somewhat silly three-mic setup, with two ribbon mics (Cascade Fatheads) in stereo around the guitar, and my Rode NTK in front my face for vocals. Ribbon mics always record in a figure-8 pattern, capturing sound in two directions and rejecting sound from the sides – I pointed the sides of the Fatheads towards my face to try and keep the main vocal sound out of the guitar, while they would inevitably catch some of the echoes from the room.

All three mics were going into my old Mackie mixer, with the ribbons hard left and right and the vocal mic panned center. The stereo outputs went into the Focusrite to record a stereo track into Audacity. The idea of recording this way occurred to me partly because I’d recently discovered the excellent Variety of Sound plugins such as Density. These plugins have the capability for “Mid/Side” processing, aka distinguishing between the information that overlaps between the stereo channels (the mid) and the information that differs between them (the sides) in order to process them separately. This means I could rebalance between the middle mic (mostly vocals) and the sides (guitar and room sound) and treat them differently with EQ, compression, and saturation.

Density by Variety of Sound

The other happy accident came when two of my two takes magically turned out incredibly close together in timing, so much so that I was able to use the slightly weaker take as a kind of jerry-rigged slap delay (the slight echo effect you can hear on the lead vocals and particularly the more percussive guitar strums). I took the alternate take, compressed it pretty hard, rolled off the high and low frequencies, and placed it slightly after the main take with an amount of delay that I found to have a fun rhythmic energy to it. Because the “delay” is actually another organic performance, there is a slight ebb and flow in the amount of delay that feels (to my ear) less robotic while still not sounding like anything found in nature.

What more needs to be said? It’s a silly little song, and I wanted to keep the feeling of a rough bedroom recording. A couple happy accidents and some playing around made the rest of the track. Oh, and thanks to my friend SJ for sharing the tea, and Jeremiah for the mug, that made the song exist in the first place.