Sampler Architecture

A sampler is a pool, a map, voices and envelopes working together, not a play button.

Related

SamplerSamplersDrum Kit

The problem

A sampler looks like it only plays recordings, so why does it also have filters, envelopes, modulation and routing?

How it works

A sampler is built from parts. The sample pool holds recordings and metadata; the map points key and velocity zones at samples; the voice engine handles transposition, looping and crossfades; envelopes and filters shape attack and brightness; the modulation matrix connects LFOs, velocity and controllers to parameters; effects and output handle pan, reverb and bus routing. Each keypress takes a voice, and the voice count sets how many notes sound at once. Disk streaming keeps large samples out of memory, at the cost of latency on the first trigger.

Practice and settings

Place samples in the map first, then shape envelopes, then add filtering and effects. Do not set an attack shorter than the sample's onset, or the transient is cut. Let filter cutoff follow velocity to restore brightness lost in transposition. Size the voice count for the densest passage and leave headroom. Streaming preload length affects first-note latency, so prioritise response in live use.

What to listen for

Hear a fast chord: with too few voices, the earliest notes cut off. Hear the low register, where poor streaming delays the first note by a beat. Hear the filter envelope: too low a cutoff dulls the attack, too high exposes transposition grit. Hear stacked layers and check whether envelopes start together.

Common misconceptions

That a sampler has one playback engine; that voice count only affects loudness; that the filter is decoration.

Sources

• Curtis Roads, The Computer Music Tutorial — 采样器声部与循环结构核对，未抄原文
• MIDI Association, General MIDI 1 Sound Set — 音色与鼓组编号核对，未抄原文

Used for fact-checking only; all prose is original.