Testing the effects of game music on cognition

For Jan Plass‘ class on research in games for learning, I’m working on an experiment testing the effects of game soundtracks on cognitive performance. The game in question is All You Can ET, developed by the NYU CREATE Lab.


Here’s the music:

You’re hearing four versions of the basic 32-bar loop: fast major, fast minor, slow major, and slow minor. We’ll be playtesting each of these versions to see how (or whether) they affect game performance.

This is my first game soundtrack. To get inspired, I searched for “best game soundtracks” on Google and listened through some lists. I quickly decided that the Hollywood film score feel of contemporary games was wrong for All You Can ET, so I focused my listening more on 1980s Nintendo and Sega soundtracks: Mario, Zelda, Metroid, the Mega Man series, Sonic. Game soundtracks of that era tend to use a rock beat rather than a “four on the floor” dance beat and almost never use swing, so I followed suit. I’m not an 8-bit purist, though, so I used the warmer and more analog timbre of Ableton’s Core 909 drum kit. The basic beat uses kick, snare, and clap, with hi-hat and rimshot parts that enter and exit to create textural variety.

For the melodic parts, I used the Helm synth. The bass is a modification of Helm’s Homebase preset, filtered to be less harsh-sounding. The lead synth is the Helm CM Pluck Time preset, which flawlessly evokes 1980s game sounds. I wanted a pad for background harmony, but was unable to find a satisfying one. Instead, I used an arpeggiated sound, a modification of Helm’s Chip Waltz preset, along with simple chords played using the Strings5 Quartet preset on Ableton’s Tension instrument. These synthetic timbres were chosen both to evoke retro games, and to provide a strong contrast in case we ever decide to test a variant using the acoustic/“real” instruments more typical of contemporary game soundtracks.

I composed the parts at a medium tempo of 110 beats per minute, and then set about trying to find the right tempo. For the “fast” tempo, I chose 140 beats per minute, typical of electronic dance music and energetic rock. Game soundtracks frequently use faster tempos than that, but it would be difficult to create music that would work both at such a frenetic pace and at a slower one. For the “slow” tempo, I chose 105 beats per minute, which is really a moderate tempo, but genuinely slow tempos make the music sound unnatural.

For a structure, I decided in advance to use an endlessly repeating 32 bar loop, which is 0:55 in length at the fast tempo, and 1:10 in length at the slow tempo. The loop is divided into four eight-bar sections, and each section is divided into front and back halves. All of the instruments repeat in identical four-bar loops, and I created musical interest by having the loops enter and exit and section boundaries, as is common practice in all forms of electronic pop and dance music.

I composed the melodic parts using the C Mixolydian mode. The C root was chosen to harmonize with pitched sound effects in All You Can ET. Mixolydian mode does not have the sense of tension and resolution evoked by the plain major scale, making it a better fit for the ambient feel of an endless loop. Mixolydian also shares the flat seventh scale degree with most minor modes. I drew all of the parts straight into the piano roll with the mouse pointer, wanting a rigidly quantized feel. To create the minor key version, I duplicated the entire loop and simply lowered each E to E-flat and each A to A-flat.

We’ll do an initial test on the fast and slow major versions on Monday, and possibly test more versions over the summer. I’ll keep you posted on the results.