A new plugin from French developer Anemond utilizes machine learning to split sounds up into individual layers that can be remixed and randomized, generating thousands of rhythmic and melodic variations of any sample or loop.
Based on the same machine learning engine as Anemond's Factorsynth, a "fully-fledged sound design studio", Factoid 2 offers a simpler interface, reduced feature set and a lighter CPU load than its bigger and more expensive brother.
Described as a "loop revamper", Factoid's machine learning engine can extract between 2 and 8 layers from any sample. Anemond explains that Factoid isn't a stem separation tool, but instead a unique kind of processor that decomposes, or "factorizes" audio into a set of "temporal and spectral components".
This is "not intended to unmix instruments," Anemond says, "but to discover and extract unpredictable but interesting sound elements with a certain degree of structure", like notes, drum hits, and rhythmic or melodic motifs.
AI meets sampling: Coalescence uses machine learning to reimagine what a sampler plugin can do
Once these layers are extracted, Factoid can remix and manipulate them in a variety of ways to create new ideas. You're able to randomize sample slices on a quantized grid to create glitchy patterns, transform melodies into textures, and solo, mute and adjust the levels for each factorized layer. After manipulating your sample, you can drag and drop the results from Factoid into your DAW.
Factoid 2 is priced at €29 and is available as a standalone app or VST3/AU plugin for macOS and Windows.
Find out more on Anemond's website.
