7.10.12. STRUDEL-12 — Samples

Oscillators cover a lot of ground but most modern music is built from samples. daStrudel loads sample banks from disk and indexes them by name, so mini-notation patterns like s("bd sd hh cp") just work.

7.10.12.1. Part A: Sample Banks

A sample bank is a directory of folders — one folder per named sound. Every audio file inside the folder becomes an indexed variation:

drums/
  bd/            ← "bd" sound
    BT0A0A7.wav  ← variation 0 (bd:0)
    BT0AADA.wav  ← variation 1 (bd:1)
    ...
  sd/
  hh/
  cp/

strudel_load_sound loads one folder under a given name. Files are sorted alphabetically, so the :0, :1, … indices are stable:

require strudel/strudel_player

strudel_load_sound("{MEDIA}/drums/bd", "bd")
strudel_load_sound("{MEDIA}/drums/sd", "sd")
strudel_load_sound("{MEDIA}/drums/hh", "hh")
strudel_load_sound("{MEDIA}/drums/cp", "cp")

Once loaded, the sounds are visible to every pattern through the mini-notation:

let pat <- s("bd sd hh cp") |> gain(0.7)

7.10.12.2. Part B: Variations

Multiple files in one sound folder are addressable as name:index:

// Alternate two kick variations per cycle
let pat <- s("bd:0 bd:1 bd:0 bd:1") |> gain(0.7)

Omitting the index (just "bd") always picks :0. Use variations to keep machine-gun patterns from sounding mechanical — even a pair of slightly different kicks feels more human than a single repeated file.

7.10.12.3. Part C: Pitch Shifting with speed

speed(factor) scales the sample playback rate. speed(2.0) is one octave up and twice as fast; speed(0.5) is one octave down and half as fast. It is tape-style pitch shifting — shorter or longer than the original:

let normal <- s("bd bd bd bd") |> speed(1.0) |> gain(0.6)
let pitched <- s("bd bd bd bd") |> speed(2.0) |> gain(0.6)

speed only affects sample playback; oscillator voices (sine, sawtooth, …) ignore it and use note() for pitch instead.

7.10.12.4. Part D: Single-file Decoding with load_audio_file

Sometimes you just want to decode one audio file — to inspect it, to feed it into your own DSP chain, or to build a custom sample bank. load_audio_file decodes WAV / MP3 / FLAC / OGG into a Sample struct:

require strudel/strudel_samples

let sample <- load_audio_file("{MEDIA}/drums/bd/BT0A0A7.wav")
print("channels: {sample.channels}\n")
print("sample_rate: {sample.sample_rate} Hz\n")
print("frames: {length(sample.data) / sample.channels}\n")

An empty Sample (length(data) == 0) signals a decode failure — always check before using the data.

See also

Full source: tutorials/daStrudel/daStrudel_12_samples.das

Previous tutorial: STRUDEL-11 — Synthesis

Next tutorial: STRUDEL-13 — SF2 SoundFont Playback