The Sound of the Machines
PlayStation
One room. Twenty-four voices. An orchestra.
Sony SPU (CXD2922) 24 voices · Sony · 1994
The Chip
Imagine you are a composer in 1997, and you have just been handed the most powerful sound hardware of your career. Twenty-four voices — far more than the Super Famicom's eight. Each one can play a recorded instrument: a real violin, a real flute, a real timpani, sampled and stored in the machine.
But there is a catch, and it shapes everything. Those recordings are compressed — stored as 4-bit ADPCM, which rounds away the brightest frequencies to save space. And there is only one reverb unit in the entire machine. You cannot give the strings their own concert hall and the flute its own chapel. Everyone gets the same room.
A lesser composer hears two limitations there. A better one hears the secret of the orchestra. Because a real orchestra is also a group of players who have been compressed into one space — one stage, one hall, one shared acoustic. When you route all twenty-four of the PlayStation's voices through the single shared reverb, they stop being twenty-four separate sounds. They become an ensemble in a room.
That is the whole instrument list, and the whole trick. Compressed samples, seated in stereo, all breathing the same reverb. Everything the PlayStation is famous for — the lush, warm, cinematic RPG scores of the late nineties — comes out of those three facts.
Channel reference (technical)
Tomita's Track
Morning of Triumph
What to listen for
Listen for how the instruments share one space. You will not be able to point to where any single instrument "is" in a room of its own — they are all in the same hall. That shared reverb is the entire illusion. At the halfway mark a flute walks in; at the end the whole group arrives together. Nothing gets louder. More voices simply join the same room.
The PlayStation's SPU gives you 24 voices of 4-bit ADPCM and a single hardware reverb unit — one room for everyone. You cannot give each instrument its own acoustic space. At first that sounds like a limitation: no per-instrument reverb, no separate halls. But it is exactly why the result sounds like an orchestra. A real orchestra is also twenty-some players in one shared room. The constraint was the realism.
How this piece was made
When I tried to work out why Dragon Quest VII on the PlayStation sounds like a real orchestra, I expected the answer to be the samples — better recordings of better instruments. It was not. The samples on their own are small, compressed, 4-bit things. The orchestra is the reverb. The SPU has one reverb unit, and every one of its 24 voices passes through the same one. That single shared room is the trick: twenty-four separate sounds, all breathing the same air, stop sounding like twenty-four separate sounds and start sounding like an ensemble in a hall. So I wrote 'Morning of Triumph' to be heard that way — strings and a rolling harp first, a flute walking in at the halfway mark, and the whole group arriving together at the end. I did not make the sounds bigger to get there. I let the shared reverb make them one.
Techniques used in this track
- All 24 voices routed through one shared hall reverb (Schroeder type, wet 0.45) — the single global reverb is what fuses 24 separate samples into one orchestra in one room
- 4-bit ADPCM sample playback: the compression rounds the high frequencies, leaving the warmth that defines the PlayStation's orchestral sound
- Harp arpeggio in rolling eighth notes — the signature texture of late-90s Square orchestral scores, laid under the strings
- Per-voice stereo placement: strings spread left and right, flute centered, bass anchored — a real seating chart, not a mono stack
- L/R stereo reverb with the two channels delayed slightly against each other, widening the space without smearing the melody
- Section dynamics: A (strings + harp, bars 0–24) → B (flute enters, 24–48) → C (full ensemble, 48–72) — the arrival is built by addition, not volume
Three things about the Sony SPU
The PlayStation was not the most technically flexible audio hardware of its generation — the Sega Saturn could do more, and was harder to use. But the SPU's particular constraints, the compression and the single shared reverb, are exactly what gave the console its unmistakable warm, orchestral voice. Here is what composers were actually working with.
There is only one reverb on the whole machine — and that is the orchestra.
The PlayStation's SPU has a single hardware reverb unit. It cannot give each instrument its own acoustic space; every one of its 24 voices passes through the same reverb. That sounds like a limitation, and engineers at the time sometimes treated it as one. But it is the exact reason late-90s PlayStation scores sound like real orchestras. A real orchestra is also twenty-some players in one shared room, breathing the same air. When twenty-four separate samples all share one reverb, they stop sounding like twenty-four samples and start sounding like an ensemble. The machine had no choice but to put everyone in the same hall — and the same hall is what an orchestra is.
The warmth comes from throwing data away.
Every instrument on the PlayStation is stored as 4-bit ADPCM — a heavily compressed format that rounds off the highest, brightest frequencies to save space. On paper this is degradation: you are literally discarding part of the sound. But what survives the compression is the low and mid warmth, and what is lost is the brittle top end. The result is the soft, rounded, faintly nostalgic tone people associate with the console. The PlayStation does not sound warm despite the compression. It sounds warm because of it.
Each voice has its own seat.
Although they share one reverb, the SPU's 24 voices each have independent volume and stereo position. A composer can seat the strings to the left and right, place the flute in the center, anchor the bass — building a real seating chart rather than a flat mono stack. Combined with the single shared room, this is what produces the sense of a group of players arranged on a stage in front of you, rather than a pile of sounds stacked on top of one another.
What the shared room taught
There is a lesson hidden inside the PlayStation's biggest limitation. One reverb for everyone was a cost-saving decision — a single unit is cheaper than twenty-four. It was never meant as an artistic statement. It was an engineer trimming a budget.
But that one shared room is the reason a stack of compressed samples can sound like an orchestra playing in a hall. The limitation did not hold the music back. It supplied the one thing a pile of separate sounds is missing: a shared space to live in. An orchestra is not defined by having the best instruments. It is defined by everyone being in the same room at the same time.
One reverb was a way to save money. It became the sound of an orchestra. That is not despite the constraint. That is because of it.