Recorded with the help of a software model programmed with 10 hours of improvised recordings using wood, metal and drum skins.

Emptyset, the experimental duo comprised of James Ginzburg and Paul Purgas, have developed a machine learning system for Blossoms, their second album for Thrill Jockey.

Programmed using a collection of existing material from the duo, as well as 10 hours of improvised recordings using wood, metal and drum skins, the system was designed to recognize coherent patterns within a large sonic data set.

“Blossoms is a work built on hybrids and mutations”, explains the label, “combining complexly synthesized audio with reverbs derived from impulses taken in architectural sites Emptyset have worked in previously.”

“The assembled compositions are emblematic of Emptyset’s dedication to forward-looking sound and examine patterns of emergence and augmentation, fragmentation and resilience, and the convolution of biotic and abiotic agency.”

Blossoms arrives on October 11 and is available to pre-order now. Check out the album artwork and tracklist below.

Tracklist: 

01. ‘Petal’
02. ‘Blossom’
03. ‘Bloom’
04. ‘Pollen’
05. ‘Blade’
06. ‘Axil’
07. ‘Filament’
08. ‘Bulb’
09. ‘Stem’
10. ‘Clone’

Read next: The 25 best albums of the last three months – April to June 2019