Treble has raised an $18 million extension to its Series A for software that simulates how speech, noise and audio devices behave in realistic spaces. The Icelandic company wants voice-model and hardware developers to test difficult listening conditions without recording every environment in the physical world.
Paladin Capital Group led the round, with KOMPAS VC, Frumtak Ventures, EIC and Omega also participating. Treble previously raised $12 million in 2024 and has now collected more than $40 million in total. It counts Amazon and Logitech among its customers.
The platform can generate synthetic audio for speech enhancement, noise suppression and model training, then evaluate voice systems across different acoustic conditions. Treble also provides virtual prototypes for headphones and speakers, including tests of how microphone placement affects a smart speaker’s ability to hear commands. A partnership with Hugging Face produced a benchmark for speech recognition under realistic conditions.
The company is expanding toward smart glasses, vehicles, drones and robots that depend on sound. Its approach is based on physics simulation, but simulated data still needs validation against real microphones, rooms and users. The funding will help Treble broaden those tests while avoiding the cost and limited coverage of recordings gathered solely from the internet or staged environments.