Acoustic Startup Treble Secures $18 Million to Power Voice AI Systems
Technology company Treble from Iceland has managed to secure $18 million in a Series A+ round that was backed by Paladin Capital Group in order to develop Treble’s physics-based voice simulation platform. Alongside returning investors like KOMPAS VC and Frumtak Ventures, the investment brings Treble’s total investment to over $40 million. Investment is directed at development in North America, physical robots, smart wearables, and voice-controlled consumer devices.
Founded in 2020 by acoustics experts Finnur Pind and Jesper Pedersen, Treble is helping to solve severe training challenges in the speech recognition ecosystem. Standard voice models require lots of real-world audio samples and datasets obtained through web scraping but have difficulty working with reverberations or background noise. Treble gets around physical capturing of sound samples and instead uses digital twin technology to mimic the acoustic behavior of devices.
Leading manufacturers of hardware and software, such as Amazon and Logitech, use the cloud-based services provided by Treble to assess speech models and test micro-speakers before production. Through its collaboration with Hugging Face, Treble gives developers the tools to test speech software command processing capabilities in the most physically complicated environments.
This is an indication of a shift in the overall landscape from using static scraped Internet datasets for training to physics-based synthetic datasets. With the integration of autonomous agents, spatial computing headsets, and robotics technologies by tech firms, accurate acoustic simulation is essential.
Treble’s fundraising round sets the stage for physics-based acoustic simulation as a fundamental component of the coming generation of hardware and generative AI voice technology. In addition to this, using physics-based virtualization for gathering acoustic data, instead of collecting it from the real world, enables hardware manufacturers to reduce the time required for development from months to days.
