Family 8 — Audio · rung 8.9

The S-DSP has a hardware echo (reverb) unit. This example isolates it — one short "pop" on a timer, turned into a decaying tail by the echo — with START to toggle it off and on, because you only hear reverb next to the dry sound.

(No screenshot — the lesson is audible. Run it with sound and listen.)

What you'll learn

audioSetEcho(delay, feedback, volL, volR) sizes the echo ring and sets how long the tail rings (feedback) and how loud the wet signal is. Call it before enabling.
audioSetEchoFilter(fir[8]) — the 8-tap FIR on the echo path; tap 0 = 127 passes the echo straight through.
audioEnableEcho(voiceMask) routes voices into the echo; audioDisableEcho() turns it off.
Echo costs ARAM — the ring buffer is real memory, sized by the delay, so a bigger hall leaves less room for samples.

SNES concepts

Echo is a DSP feature, not a CPU one: the S-DSP keeps a delay ring in APU RAM, feeds a fraction of the output back into it (the feedback), and mixes the delayed signal (the wet volume) back over the dry. audioSetEcho configures all of that in one call; audioEnableEcho chooses which of the 8 voices are sent in.

How to build

make -C examples/audio/echo

Run echo.sfc in luna with audio and press START to hear dry vs wet.

Modules used

console, audio, input

Ladder

The focused counterpart to soundboard (which uses echo as one of many features). It builds on the audio v2 engine — see soundboard first for the sample-loading basics.