lachender Musikproduzent am Mischpult im Studio
© istock.com/gorodenkoff

A result of the long-term technology cooperation between Fraunhofer IDMT and Jamahook, the Swiss music technology company, is the Virtual Studio Technology (VST) plugin for digital audio workstations. Developed by Jamahook, the plugin allows music producers to find the most suitable audio loops for a production, based on examples extracted from an audio mix. The development is based on Fraunhofer IDMT’s SoundsLike AI-based music classification technologies.


Producing musical content is becoming increasingly easy with modern music production software. Music producers have access to huge quantities of audio samples and loops, but pinpointing the right material quickly is difficult, so the music must first be analyzed, organized and made searchable in a meaningful way.


AI-based music classification systems from Fraunhofer IDMT automatically annotate and categorize unlabeled music data in the Jamahook database and identify the audio samples and loops that fit a music producer's needs best, based on specific examples. The classification systems are based on Fraunhofer IDMT's SoundsLike technology which uses deep learning technologies to automatically extract musical properties and find similar songs.

The Jamahook VST plugin first analyzes a sample of the current music production in terms of instrument, tempo, harmony, mood, key or genre. The plugin then recommends a selection of loops automatically extracted from the Jamahook database that harmonizes well with the given music production. The user can select which characteristics the loops should match, resulting in a selection that can be immediately integrated into the current production at the appropriate tempo.

This might also be interesting for you

Research topic

Audio and Visual Content Analysis

Extracting meaningful data from audiovisual content


Research topic

Automatic Music Analysis



AI-based tagging and search for large music catalogs