The NES sound chip - the Ricoh 2A03 APU, integrated into the 6502-based CPU - was the
universal platform for game music across the NES’s commercial lifespan (1983–1995
in Japan; 1985–1995 in North America). Every NES composer worked within the same
constraints.
The two pulse wave channels - Square 1 and Square 2 - were the primary
melodic voices. They supported programmable duty cycles (12.5%, 25%, 50%, 75%) giving
timbre variation, and hardware-assisted pitch sweeps enabling vibrato effects. Melody
and countermelody were typically split between these two channels.
The triangle wave channel operated at fixed amplitude with no volume
envelope - making it ideal for bass lines. Its stepped waveform produced an
aliasing artefact at lower pitches that gave NES bass a characteristic gritty quality.
The noise channel used a linear feedback shift register to generate
pseudo-random noise used for percussion and sound effects.
The DPCM channel - Delta Pulse Code Modulation - could play back
1-bit delta-encoded PCM samples at selectable rates from approximately 4.2 kHz to
33.5 kHz. Memory and fidelity constraints limited its use in music, but it was
employed for drum hits and short instrument patches.
In this context, Kageyama’s use of borrowed chords in the Ghosts ‘n Goblins
Stage 1 theme was notable. Most NES composers of 1985–1986 used straightforward
diatonic harmony in major or relative minor. Modal mixture - borrowing chords from a
parallel mode - was a more harmonically complex technique that gave the GnG NES
soundtrack a sense of doom distinct from contemporaneous platformer soundtracks.