FACTORSYNTH and FACTORmini – Max For Live Devices

Factorsynth FACTORmini Ableton Live Devices

Machine learning comes to Ableton Live with Factorsynth and its introductory version, Factormini. Two Max For Live devices that use a data analysis algorithm called matrix factorization to decompose any audio clip into a set of temporal and spectral elements.

By rearranging and modifying these elements it is possible to perform powerful transformations to any clip, such as removing notes or motifs, creating new ones, randomizing melodies or timbres, changing rhythmic patterns, remixing loops in real time, creating complex sound textures, etc.

Created by Paris-based independent researcher and developer J.J. Burred (https://www.jjburred.com), Factorsynth aims at bringing some of the ideas and methods of machine learning to a wider audience of composers, sound designers, and DJs.

Once an input clip is analyzed (“factorized”), a set of up to 30 temporal and spectral shapes are extracted from the clip. For instance, different pulsating temporal shapes can indicate the positions of the different instruments contained in a drum loop (kick, snare, hi-hat…). Spectral shapes can be understood as a “timbre palette” of the input sound.

Both temporal and spectral shapes can be modified and freely recombined, allowing not only to remove elements from the original sound, but also to create new ones by combining originally unrelated shapes (e.g. you can make the rhythm pattern of the snare drum play the timbre of the kick). This opens up many new ways to repurpose or generate random variations to the clips.

Factorsynth also allows an advanced form of cross-synthesis, where it is possible to combine temporal shapes extracted from a master sound with spectral shapes extracted from a second. Taking again a drum loop as example, the kick drum pattern would drive a certain note of a second sound, the hi-hat a different note, and so on.

Concerning computation time, Factorization usually takes a few seconds for short clips. Once the sounds are factorized, the modification and sound synthesis happen in real time.

Some demo videos:
– Presentation: https://www.youtube.com/watch?v=Vjbe_vJth38
– Demo: https://www.youtube.com/watch?v=3LodkAEPEY4
– Controller demo: https://www.youtube.com/watch?v=ly2HoxYUBVE
– Factormini demo: https://www.youtube.com/watch?v=PVRXblw8NTg

You can find more detailed information on the FAQ section:

or on the user manuals:

Available at https://www.jjburred.com/software/factorsynth/

Leave a Reply

Your email address will not be published. Required fields are marked *