Building the MMA Drone System
The engineering challenges of designing a drone system for a 7-note scale with 7 modes — how the SuperScales drone differs from the MetroDrone, and why mode-specific context changes the engineering requirements.
The MetroDrone, described in an earlier article, is a general-purpose tool: a sustained pitch at a specified frequency plus a metronome click. It works for any key center, any scale, any ear training context. The MMA drone system is a specialized tool: 864 carefully engineered audio files, each optimized for a specific mode of MMA in a specific key, with mode-specific context built into the drone’s timbre. The engineering requirements are substantially more demanding.
The Engineering Challenge — Seven Modes, One Key Center
The fundamental engineering challenge of the MMA drone system is this: the same key center (C) sounds different when you are practicing the MMA scale itself, the Lydian Dominant mode, and the Altered scale — even though all three share the same root pitch. The differences are not just theoretical. A student practicing the Altered scale against a neutral C drone hears the scale’s b9 and b13 as clashing against the clean root — because the neutral drone provides no context for why those altered tones are appropriate. A mode-specific drone that includes the b9 and b13 softly underneath the root provides that context immediately.
Building mode-specific drones required solving engineering problems that the simple MetroDrone does not encounter:
Mode-Specific Drone Specifications
| Mode | Root tone | Characteristic tones added | Engineering note |
|---|---|---|---|
| MMA (Mode 1) | C | Eb (♭3) at -12dB | The flatted 3rd is the defining difference from major — include it to immediately establish minor context |
| Dorian b2 | C | Db (♭2) at -12dB | The b2 is the most exotic note — hearing it immediately in the drone establishes the Phrygian-adjacent sound world |
| Lydian #5 | C | F# (#4) at -13dB, G# (#5) at -14dB | Both the #4 and #5 define this mode — include both but at slightly reduced levels to prevent whole-tone density in drone |
| Lydian Dominant | C | F# (#11) at -12dB, Bb (b7) at -13dB | The tritone (#11) plus the dominant 7th (Bb) immediately establishes the Lydian Dominant sound world |
| Mixolydian b13 | C | Ab (b13) at -12dB, Bb (b7) at -13dB | The b13 plus dominant 7th establishes the primary dominant context without the tritone tension of Lydian Dominant |
| Locrian nat2 | C | Gb (b5) at -12dB, D (nat2) at -14dB | The b5 establishes half-diminished context; the nat2 (vs. standard Locrian’s b2) is included softly to mark the distinction |
| Altered (Mode 7) | C | Db (b9) at -13dB, Ab (b13) at -14dB | Only two of the four alterations included — the b9 and b13 are the most harmonically informative without creating an overwhelming cluster |
The Loop Engineering Problem
Each of the 864 drone files is designed to be loopable — the end of the file connects seamlessly to the beginning so that a student can practice for longer than the file’s 4-minute duration without the drone cutting out. Achieving a seamless loop with sustained tones requires that the file begin and end at exactly the same phase in the drone’s sound wave — a zero-crossing loop point.
SuperCollider’s envelope system handles this: each drone sound begins with a short attack (0.01 seconds, inaudible) and ends with a short release (0.01 seconds, equally inaudible), creating a clean loop point without the click that would occur if the waveform was cut in the middle of a cycle. The metronome click at the loop point is synchronized to the beat grid so that the loop never creates a rhythmic hiccup.
This loop engineering detail is invisible to the student — they simply hear a continuous drone. But without it, every loop point would create a distraction that interrupts the practice session. Engineering that disappears into the background is the best engineering.
Audio engineering and pedagogical design — the same goals
Mode-specific context
Including mode characteristic tones in the drone is an engineering decision driven by pedagogical theory — students acquire mode sound worlds faster with context than without. The -12dB level is an engineering specification derived from listening tests, not from mathematical principles alone.
Invisible engineering
Good infrastructure engineering is invisible — seamless loops, consistent levels, reliable playback. When engineering works perfectly, the student is aware only of the practice material, not the infrastructure. Invisible engineering is the goal.
Constraint-driven design
The Altered mode’s characteristic tone selection (b9 and b13 only, not all four alterations) is a constraint-driven design decision — including all four created a drone that was acoustically counterproductive. Constraints often produce better design than unconstrained freedom.
Specification precision
864 files, each with a precise specification (-12dB characteristic tones, specific loop points, specific tempo variants). Engineering at this scale requires specifications precise enough to be automated — human memory cannot track 864 individual decisions reliably.
The final article in this series looks at the complete journey from discovery to publication — how the SuperScales book was assembled from verification, production, and design components, and what comes next in the SuperScales series.
Audio engineering for music education is discussed in Roads, The Computer Music Tutorial (1996). Loop point engineering: SuperCollider envelope documentation at supercollider.github.io. The SuperScales drone system and complete book are at muse-eek.com.