Factoid 2: Redefining Sound Manipulation with Machine Learning

  • Factoid 2, developed by French innovator Anemond, employs machine learning to deconstruct sounds into individual layers for remixing and randomization.
  • It shares its machine learning foundation with Factorsynth but offers a simpler interface, reduced features, and lower CPU load.
  • Described as a “loop revamper,” Factoid 2 can extract 2 to 8 layers from any audio sample, focusing on temporal and spectral components rather than instrument separation.
  • Users can remix and manipulate layers to create new sonic ideas, including glitchy patterns, textured melodies, and adjustable volume levels.
  • Priced at €29, Factoid 2 is available as a standalone app or VST3/AU plugin for macOS and Windows.

Main AI News:

A recent innovation by French developer Anemond leverages machine learning to dissect sounds into distinct layers, allowing for remixing and randomization, thereby generating myriad rhythmic and melodic variations from any sample or loop. Drawing from the same machine learning prowess as Anemond’s Factorsynth, a comprehensive sound design studio, Factoid 2 presents a streamlined interface, pared-down feature set, and lighter CPU load compared to its more extensive counterpart.

Termed a “loop revamper,” Factoid’s machine learning engine can discern between 2 and 8 layers within a sample. Anemond clarifies that Factoid isn’t a stem separation tool but rather a unique processor that dissects, or “factorizes,” audio into a collection of “temporal and spectral components.”

This isn’t about separating instruments,” Anemond explains, “but rather uncovering and extracting unforeseeable yet intriguing sonic elements with a certain level of organization,” such as notes, drum hits, and rhythmic or melodic motifs.

Once these layers are isolated, Factoid can remix and manipulate them in diverse ways to spark new concepts. Users can randomize sample slices on a quantized grid to produce glitchy patterns, convert melodies into textures, and control individual layers’ volume, muting, and solo functions. Following sample manipulation, results from Factoid can seamlessly integrate into your DAW via drag-and-drop functionality.

Factoid 2, priced at €29, is accessible as both a standalone application and a VST3/AU plugin compatible with macOS and Windows.

Conclusion:

Factoid 2, with its innovative approach to sound manipulation through machine learning, represents a significant advancement in the audio production market. Its ability to dissect and reassemble audio samples opens up new creative possibilities for producers and musicians alike. With its accessible pricing and compatibility across platforms, Factoid 2 is poised to become a staple tool in the arsenal of digital audio workstations, catering to a wide range of users seeking innovative ways to shape and mold sound.

Source