What a sound chip actually does
A sound chip takes instructions from a game's software — play this note at this pitch with this volume, start this effect, stop that channel — and translates them into electrical signals that speakers turn into sound waves. The translation requires generating waveforms, and the character of the sound depends almost entirely on what waveforms the chip can generate, how many it can generate simultaneously, and what processing it can apply to them.
The major approaches were synthesis — generating waveforms mathematically from oscillators and modifying them with filters and envelope generators — and sample playback, in which recorded audio was stored digitally and played back at variable speeds to produce different pitches. Synthesis was cheaper in terms of storage (no audio data to store) but produced characteristic electronic timbres. Sample playback could sound like any recorded instrument but required storage space that was expensive on early hardware. Most gaming platforms used synthesis exclusively, sample playback exclusively, or some combination, and each choice produced a distinctive sonic identity.
The chips that defined the sound of 1980s gaming were not generic electronic components. They were custom-designed integrated circuits, created specifically for their platforms by engineers who made specific decisions about what the chip needed to do. Those decisions — often driven by cost, storage constraints, and the technical capabilities of the era — produced the sonic landscapes that listeners now associate with entire childhoods.
The SID chip
The MOS Technology 6581 — universally known as the SID chip, for Sound Interface Device — was designed by Bob Yannes and included in the Commodore 64, released in 1982. Yannes was a musician as well as an engineer, and the SID reflected musical thinking: it was designed to produce sounds that could be useful for music, not merely sounds that could indicate game events.
The SID had three independent oscillators, each capable of generating four waveforms: triangle, sawtooth, pulse (with variable pulse width), and noise. Each oscillator had a dedicated amplitude envelope generator with separate attack, decay, sustain, and release parameters — the same ADSR structure used in professional synthesisers. The chip had a single multimode filter that could be applied to any combination of the oscillators, operating in low-pass, high-pass, or band-pass modes. One oscillator could ring-modulate another — a synthesis technique that produces metallic, bell-like timbres by multiplying two waveforms together.
The SID's musical range was exceptional for a chip of its era. Composers including Rob Hubbard, Martin Galway, and David Whittaker treated it as an instrument rather than as a sound effects generator, producing complex multi-voice arrangements that referenced jazz, classical, and pop music. Hubbard's compositions for Commando (1985), Monty on the Run (1985), and International Karate (1986) are still discussed as examples of what the SID could do in the hands of a composer who fully understood its architecture. The SID community — demoscene musicians, retro enthusiasts, chip music artists — remains active decades after the hardware it was written for became obsolete.
The NES APU and the Genesis YM2612
The NES's audio processing unit was integrated into the main 2A03 CPU rather than a separate chip. It provided five channels: two pulse wave oscillators with variable duty cycle (producing the characteristic NES "beep"), a triangle wave channel for bass lines and melodic lines with lower pitch, a noise channel for percussion and sound effects, and a DPCM (delta pulse-code modulation) channel for low-quality sample playback. The constraints were severe — the triangle wave had no volume control, the noise channel produced only white noise filtered through a shift register — but composers including Koji Kondo and Hirokazu Tanaka produced work that maximised the available channels through rhythmic complexity, melodic counterpoint, and creative use of the noise channel as a rhythmic element.
The Sega Genesis used two sound chips simultaneously. The principal chip was the Yamaha YM2612, an FM synthesis chip providing six channels of frequency modulation synthesis. FM synthesis — developed by John Chowning at Stanford and commercialised by Yamaha in the DX7 synthesiser (1983) — produces sound by using one waveform to modulate the frequency of another, creating complex harmonic spectra from simple sine waves. The results can approximate brass instruments, electric piano, bells, and aggressive electronic sounds with equal facility. The sixth channel of the YM2612 could operate in DAC mode — playing back 8-bit PCM samples rather than FM synthesis — which allowed limited sample playback for drums or voice snippets.
The YM2612's characteristic sound is closely associated with Yuzo Koshiro, whose arrangements for the Streets of Rage series and the Ys series explored the chip's range from electronic jazz to high-energy action scoring. The Mega Drive also included a Texas Instruments SN76489 chip — a simpler three-square-wave-plus-noise PSG — for backward compatibility with the Master System and as a supplementary sound source. Composers could layer both chips, though the results required careful balancing.
The SNES SPC700 and why it sounded different
The SNES SPC700, as discussed in the context of the console's hardware, used sample-based synthesis — ADPCM-compressed recordings of real instruments. The architectural difference from the other chips in this discussion was fundamental: where the SID, NES APU, and YM2612 were synthesisers that generated waveforms mathematically, the SPC700 was a sample player that reproduced recorded sounds. The sonic results were categorically different.
Sample-based synthesis could approximate the timbre of acoustic instruments more convincingly than any synthesis chip. A piano sample sounds like a piano in ways that a synthesised approximation doesn't. The limitation was fidelity: the ADPCM compression and the 32 kHz sample rate degraded the samples noticeably, particularly in high frequencies, giving SNES audio its characteristic warmth and slight muddiness. Composers who worked on SNES games were working with a sampler with imperfect samples and 64 kilobytes of sample RAM — tight constraints on how many instruments could be loaded simultaneously.
The compositional approach that worked best on the SPC700 treated the sample library as a collection of instrumental building blocks to be arranged, rather than a faithful reproduction system. Yasunori Mitsuda's score for Chrono Trigger (1995) used a carefully selected palette of samples — orchestral strings, synthetic textures, ethnic instruments — that sounded rich within the constraints of the chip while working around its limitations. The score is frequently cited as one of the finest in gaming, not because the hardware was unlimited but because the composer understood what the hardware did well and wrote for it specifically. The same is true of the SID compositions, the best NES music, and the best Genesis scores: each represents a composer internalising the character of a specific piece of hardware and producing work that could only have been made with that hardware.