tags: music sampling electroacoustic composition synthesis

Sampling (composition)

History

Broadly, there are two categories of [electroacoustic] [composition]:

  1. [synthesis] (generated audio)
  2. sampling (recorded audio)

Composed manipulations of recorded sound via variable-speed phonographs date back to 1920s (D. Milhaud, E. Toch, P. Hindemith).

Magnetic tape (1930s) enabled more fluid manipulation of recorded audio

Musique concrète was an early form of sampling, primarily concerned with tape splicing, manipulation, and looping.

Early Sampling Instruments

Operating principles include:

In general, all these instruments work to achieve the same thing; creative expression through manipulation of recorded audio.

Digital Sampling Instruments

Sampling techniques were greatly enhanced by [midi] (1983). Numerous other sampling instruments were introduced by Akai, Roland, Kurzweil, Ensoniq.

Techniques and Terminology

keymapping: assigning samples to [midi] note numbers

root key: the [midi] note number that plays back a sample at its original speed/pitch

key zone/key range: a contiguous region of [midi] note numbers that plays back a single sample, usually transposing pitch based on note number

multisampling: recording multiple pitches across an instrument's range and keymappping to smaller key zones

chromatic multisampling: creating a separate audio sample for each chromatic pitch on an instrument, and using 1:1 keymapping

velocity switching: using [midi] velocity values to select from multiple samples assigned to the same [midi] note number

keyswitching: using a [midi] note number outside an instrument's typical range to switch between banks of samples

looping: see: [looping]

sample patch: a pre-configured sampler instrument which has undergone many of the previously mentioned techniques (multisampling, keymapping, velocity switching, keyswitching, predetermined loop points)

sample library: a software package containing several pre-configured sample patches

downsampling: when an audio sample is transposed (pitch-shifted) upwards, some sampler [algorithm]s will skip a number of sample values in the audio file (see: [sampling in signal processing]), in proportion to the amount of pitch-shift specified. Also called decimation.