Klang.io launches Choral Mode for AI transcription of SATB recordings

Music software developer Klang.io has announced Choral Mode, an AI transcription feature designed to identify and notate soprano, alto, tenor and bass parts from a recording.

Announced on 28 October, the update is available in beta to existing users of the company’s Transcription Studio and Sing2Notes tools. Klang.io says the system can distinguish up to four simultaneous vocal parts and produce separate sheet music for each.

Users can record a performance directly or upload an audio file for analysis. The resulting notation can then be reviewed and edited, providing a starting point for preparing rehearsal materials, documenting vocal arrangements or studying individual lines within a performance.

The company describes Choral Mode as the first AI transcription tool capable of separating and notating four vocal parts in this way.

'Transcribing a single singing voice is extremely difficult for AI because of the infinite timbral variations available to the human voice,' said Sebastian Murgul, Klang.io’s founder and chief executive.

Murgul said the new mode is intended to help vocal groups, choirs and arrangers write, rehearse and analyse parts while reducing the need to transcribe each line manually.

The announcement comes as choirs prepare for the Christmas concert season. As a beta release, Choral Mode remains under development, with users invited to submit feedback and suggestions.

Based in Karlsruhe, Germany, Klang.io was founded in 2018 by Murgul and Alexander Lüngen. The company develops AI tools for transcribing live and recorded music into notation.