I hinted at the idea of redefining instrumentation in a previous post, but I’ll try to go more in depth, or at least provide a more concrete expose now. The idea of sampling has come a long way since its inception. By a broad definition, sampling is extracting a piece of a audio signal for later replay – i.e. recording. Typically, though, sampling is considered to be using portions of audio that have been recorded for the sampling or otherwise recorded before hand in a purposeful, artistic way. Sampling has largely re-imagined what can be practically considered an instrument. We typically think of instruments as being producers of sound in a controllable way. By its very definition, sampling expands the set of sounds that can be controlled, such that the sampler is an instrument without bounds of sonic creation.
Synthetic instruments are often samplers at heart. While it is entirely possible to create an instrument using a solely algorithmic approach to the sound creation (i.e. creating a signal from scratch in response to user input), often samplers allow for more natural sounding results. Add this to on-the-fly modification and the resulting sampling instrument can be quite versatile in its ability to create new, inventive sounds that do not require individual programming.
Samplers also allow for more consistent and realistic accompaniment for acts that don’t have certain segments present – either by chance or by choice. Take for example the modern rapper. His beats are sampled from real drums, modified to be more fitting to the desired tone of the song, and looped over themselves. This is often sonically suiting to the genre – more so then traditional drums. Taking it to the extreme, artists such as iamerror are completely sample based, and have the freedom to create impossible segments, as far as traditional instruments are concerned.