Find the voice.
A karaoke experiment exploring voice separation from music.
ABOUT VOXFOCUS / FROM RESEARCH TO REAL-WORLD COMMUNICATION
From a karaoke experiment
to industrial AI.
Four years of research.
A voice worth hearing.
VoxFocus began with a question: could spectral analysis and machine learning separate a human voice from music? During the pandemic, a small interdisciplinary team explored that idea in a karaoke project.
What emerged was a broader opportunity: applying AI-based voice separation to communication in noisy environments. Our research draws on analysis methods inspired by astrophysics and brings together expertise in AI, physics and acoustics.
Over four years of research and development, this work has evolved towards a shared acoustic core for people and machines. Today, COMMAND and SENSE turn that foundation into focused industrial product paths.
Our mission is simple: make relevant speech usable where noise gets in the way of people working together and machines understanding them.
Build the AI.
Machine learning, spectral analysis and acoustic expertise brought together over four years.
Bring it to industry.
One core, developed into COMMAND for people and SENSE for machines.
AWARDS & INNOVATION PROGRAMMES
Research recognised.
Innovation supported.
Ideenwettbewerb Rheinland-Pfalz
First place in the Mainz region, idea maturity level 1.
View the awardGründungswettbewerb – Digitale Innovationen
Selected among the competition’s Top 50 teams.
View the selectionMISSION KI
Selected for a programme helping researchers turn AI research into deep-tech ventures.
Explore the fellowshipAI NATION
Supported through the AI NATION Grant programme.
Explore the programmeTHE FOUNDING TEAM
Acoustics meets industrial execution.
David Maksimovic
Founder & CEOPhysics, acoustic architecture and real-time processing.
Jens Schneider
Co-FounderCompany building, SAP product experience and B2B go-to-market.