Augment has three main components. First is a routing framework that works behind the scenes and allows a performer to add chains of sources and effects without having to manage low level bus routing between the sounds. Secondly, it generalizes different types of sound sources to a single interface. SuperCollider recognizes synths and patterns that create synths. Patterns are meant to provide a high-level take on synths, running the algorithms that play and organize them. However, I found that the ‘instruments’ I wanted to play were sometimes best expressed in low level synths, and sometimes in high level patterns. The SynthInstrument and PatternInstrument classes in Augment generalize this to a single interface and make it easy to interchange these types and bind their controls to physical controllers. The last component is the collection of basic synth graphs. The synth graphs are mostly the same as base SuperCollider. They describe the computation that will be executed to create sounds. There are some argument keywords that are introduced to synths by Augment to help with generalizing bus routing and buffer allocation.

Req is a single track that displays many of my recent uses of Augment. It has a lot of intensive sample processing, manipulation of live synths and voices, and completely invented sounds from unitGens within the language. Most of the sampling source material is taken from recordings of Maurice Durufle’s Requiem. The sections below provide more specifics and details about different uses of the system.

Modular SYnthesis

Augment can emulate classic modular synth setups. SuperCollider is generally good at emulating the separate units of modular synthesis, but has an awkward routing system. Augment helps simplify the routing problem so you can focus on the sound sources.

Experimental Sampling

Sampling is a strength of the system. It can accomplish most glitch effects (stuttering, undersampling, abusive modulation), and can also deploy these samples in multi-layered textures. Broken Drums slows down and slices a live drum track, and MultiSampling slices pieces from a track and algorithmically builds a new song from them.

Algorithmic sample manipulation

Algorithmic sampling is especially helpful for transforming material. The tracks in this section were all drawn from a new age spoken track taken from YouTube, but were transformed in unexpected ways through slicing and resampling the source material.

source transformation

Both of these pieces reference the same source material, but have completely different outcomes. The original is a simple vocal track. Quick Scrub manipulates the pointer that reads the audio data, resulting in a randomly fluctuating read speed that completely changes the tone of the material. Voice Cloud uses a granular sampling technique where pieces are ripped from the original and overlaid at different speeds, turning a single voice into a choir.