Short Song – Class project

Posted by & filed under Academia.


For my MUS-K 361 course (Intro to MIDI / Electronic Music), my instructor gives us assignments for each unit where we have to create a short 30 sec – 2 min song and demo it in class.

reason rack

Cartouche (Precursor)

One thing I really like about these assignments, and this may seem counter-intuitive, is that we are restricted to only using the instrument covered in that unit. In this specific assignment, where we discussed synthesis (using the Subtractor in Reason), the assignment restricted us to using 3 – 6 Subtractors only — no effects processors, no drum machines, not even equalizers.
A game designer, Mark Rosewater,  that works for Wizards of the Coast once wrote an article about game development (specifically for Magic: the Gathering) where he debunked various myths about design. One point he made that has always stuck with me is “Restriction breeds creativity.”

Rosewater’s premise is that when we are completely unrestricted and have the full landscape of possibilities available to us, instead of exploring new paths we stick with what’s familiar. But when we are only able to use a pared-down set of resources, we are forced to step out of our comfort zone and explore unfamiliar / innovative solutions.

In this particular project, I used synthesized drum patches instead of drum machine samples; filter envelopes / low-frequency oscillators (LFOs) instead delay / reverb effects processors; noise synthesis instead of a distortion module. I was forced to be more creative with the main oscillators (Subtractor only uses 2 primary oscillators) rather than taking the easy way out and routing it through a series of effects.

Cartouche (Precursor) — the development

I wrote this song in about 3 or 4 hours this afternoon.  “Cartouche” is a word I picked up from the Stargate SG-1 series, but it’s in reference to Egyptian hieroglyphs — an ovular line enclosure around a series of glyphs, indicating that the glyphs are a royal name. The end result of the song reminded me of that musical mode that sounds middle-eastern, so it was the first thing that came to mind.

Since I had to use completely synthesized drums, I figured a trip-hop (Massive Attack, Tricky, Portishead) stylistic approach would be well-suited. I found a kick-bass patch and a snare patch. The arranged breakbeat was reminiscent of Phil Collin’s “In the Air Tonight” feedbacked drums.

Flipping through the bass synth samples, I eventually found a real dirty one called “Captain Black” – it was very dark and grungy. I didn’t have a melody or even an inkling of what it would sound like, so I just started tapping out some notes on my M-Audio until I found a bassline I liked. Record. Quantize. Fine-tune the arrangement. Same process with the Sonar sample.

Writing a melody is something that I still struggle with. I found a nice Polysynth sample (WaveTableClassic) that sounded like bells. I turned on Ring modulation and adjusted the filter a little bit to give it a slight delay / echo effect. Using the notes from the bassline, I tried to figure out a full 8-note scale to define the soundscape. This song was going to be short enough that no chord changes would really be feasible.

With the scale determined (I still have no idea what you would call it. It’s some minor or diminished scale, shifted to use C as the root note), I played around in that scale, with the drums and bass line looping in the background, until something fit. Record. Quantize. Fine-tune. It sounded a little busy, so I knocked out a few of the notes. It also didn’t sound like a real melody; it was playing near the front of the stage, but it didn’t have the spotlight on it. Need to add something else.

Flipped through the Polysynth patches again until I found one that fit, flavor wise, with the rest of the sounds. It was “Wheel” something or other. It was a nice sawtooth sample, somewhat smooth with a rough edge. I tried tapping out a risky melody on the keyboard. I say “Risky” because the notes were stepping into the spotlight. It blended nicely with the rest of the sounds. Nice. Record. Quantize. Fine-tune. Fine-tune some more. Fine-tune again. Fine-tune one last time. There we go.

I checked the project requirements — I was currently using 6 Subtractors and all were pre-fab patches; I was required to use one original patch. Whoops.

I added a new Subtractor, initialized it to clear it off, and started building a patch. I wanted something similar to the existing melody patch, but original. Oscillators 1 & 2 set to sawtooth for edginess. Detuned #2 down 2 octaves, and #1 up 3 semitones (minor third, I think?). Oscillator 2 set to multiply (“X”). Add some noise with a little color. Increase the Release on the Amp Envelope so it layers a little; increase the decay to make it little lazy sounding. Turn up Portamento to about 11 o’clock to make a bit lazier. Attach LFO #2 to the oscillator Phase — this will make it nice and gritty.  Slow the rate down, crank up the amount slightly to add a granular texture. For fun, route the LFO-1 control voltage to the Pan CV input on the main Mixer board, and turn up the rate and amount on LFO 1 with a triangle wave. Auto pan!

With a solid 8-bar loop, it was ready to be arranged properly. The target length is 1 to 2 minutes.

I arranged a minimal introduction, just drums and the sonar synth. Bring in the nasty bass line after 4 bars and let that hum for 4 more.

There are certain rules with conventional (teleological, as my Prof would say, as most mainstream electronica is predictably arranged) electronic music, the cardinal being the “rule of 4’s.” All main changes to the arrangement happen on bars that are some multiple of 4. Big changes are typically 16 bars apart, small changes are typically 4 or 8 bars apart. In the context of the Club / party setting, this predictable approach is absolutely critical so that the DJ can reliably introduce the song and have it mix well with the previous / next track.

At bar 9, bring in the bells, and let that simmer for 8 bars. At bar 13, bring in the main melody (8 bars long). For fun, play it twice. For bonus fun, automate the LFO-2 parameters on the second 8 bar segment (that control the phase) to really bring out the texturized grittiness of the patch.

The melody crosses the 1 minute mark — step it down after those 16 bars. Let the bass fall off after 4, and the drums after 3, letting the bells solo the last bar.

One minute, forty-five seconds.

I’ve been trying to learn to use the sound mastery tools in Reason. They’re pretty sophisticated; not as good as the uber-professional grade tools, I’ve read, but they do pretty well.

On each Subtractor, I added an M-Class Equalizer. For all but the kick-bass, I cut out the low-shelf; the kick has some sub-bass humming in it, so nothing needs to compete in that frequency range. The nasty bass line synth got a high-shelf cut, since it had no business being up in that neighborhood. For the Sonar and Bell samples, I used one of the Parametric EQ knobs, increased the Gain slightly, and then adjusted the frequency until I could hear some amplification in the playback. The “Q” knob adjusts the breadth of the frequency, to maximize the amount that is passed through the EQ. Passing the melody synthesizer through the EQ really helped shine the spotlight on it, and cleaned up the sample a lot; I guess it was getting muddied by some extraneous frequencies.

I added an M-Class Compressor as well, but I’m still learning how to use those appropriately, so it’s still a lot of guessing. 🙂

It’s not a complete song (hence “Precursor”), but it at least sounds resolved at this short length.

Let me know what you think (there’s a download link under the image up there — mp3 format).