Treble raises $18m Series A-2 led by Paladin Capital for acoustic simulation

What Treble does

Treble sells cloud software that models how sound travels through a space. It combines acoustic simulation, digital twins of rooms and devices, and generation of synthetic audio data that can be used to train and test AI models.

The platform combines two kinds of simulation, according to audioXpress. A wave-based method models effects such as diffraction and interference, while a geometrical method handles higher-frequency sound in large or complex spaces. A Python software development kit lets engineers model microphone arrays, directional sources and device shapes, and build datasets that can be filtered by distance, reverberation and other properties.

Companies use it to prototype audio hardware virtually and to test how products such as speakers, headsets and voice assistants perform in different rooms and noise conditions. Treble says this can cut prototype testing and audio data collection from months to days.

The company names Amazon and Logitech among the firms it works with. Wontak Kim, a senior audio research manager at Amazon, said Treble provides an accurate and scalable way to model real-world acoustic environments. Andy Harper, head of audio AI and signal processing at Logitech, said the fidelity of the underlying physics gave his team confidence in the simulations.

What the money is for

Treble said it would use the funds to grow in the US and to meet demand from consumer technology companies, enterprise and industrial customers, and developers building physical AI and world models.

audioXpress reported that the company is starting a push on the US West Coast under Vineet Ganju, a former Synaptics executive who joined as US general manager. Target uses in physical AI include robots, wearables, autonomous vehicles and smart-home devices.

Physical AI refers to AI systems that act in the real world, such as robots. Treble wants its simulations to be used to train and test these systems across many acoustic conditions.

"Audio and voice interaction is becoming a fundamental interface for the next generation of intelligent products," said Finnur Pind, Treble's co-founder and chief executive. He said devices and AI models that work well in a lab can struggle with reverberation, background noise and unfamiliar settings.

Who invested

Paladin Capital Group, the lead investor, made the case that robots today rely mostly on sight. François Ruether, a vice president at Paladin, said robots are largely limited to visual perception and cannot make sense of sounds such as someone falling in another room or a crash around a corner.

KOMPAS VC, Frumtak Ventures and the EIC Fund were already shareholders. Treble did not give the dates or sizes of its earlier rounds. The Next Web reported that Pind co-founded the company with Jesper Pedersen, and that both are acoustic engineers.