Why Auditory Perception May Decide Which Robots Get Adopted

Humanoid robots are being developed with advanced locomotion and vision capabilities. They walk, navigate, and manipulate objects with increasing dexterity. Yet according to Gunnar Pétur Hauksson, co-founder and chief commercial officer at Treble Technologies, a critical dimension of machine perception remains underdeveloped: sound.

"For humanoid robots to operate in the real world, they'll need more than a sense of sight."

The robotics industry has prioritized vision and locomotion due to their tractability and clear progress metrics. Simulation platforms like NVIDIA Isaac Sim have enabled rapid iteration and large-scale training in robotics, reinforcing that focus.

"In my view, human-machine interaction will become the defining hurdle for widespread acceptance of robots by humans because it heavily impacts trust, safety, efficiency, and ease of use."

Real-world environments — factory floors, crowded public spaces, homes — are acoustically complex, and current robotics systems struggle with natural communication and auditory perception in exactly those conditions.

"Humans can communicate effectively in environments that are noisy, reverberant, and chaotic, extracting meaning from sound with a level of resilience that current systems still struggle to match."

Hauksson frames the issue biologically. Humans allocate significant energy to auditory processing, which is crucial for survival and communication — roughly in the range of 15% to 20% of the brain's sensory workload, depending on context. Vision dominates, but hearing's substantial share reflects how essential sound is for functioning in human environments.

"The reason this hasn't been solved isn't a lack of awareness, but a lack of infrastructure."

Audio systems in robotics are held to a higher standard due to the need for trust and natural interaction. A robot that moves imperfectly can still seem functional; one that mishears or responds out of sync erodes confidence quickly. The bottleneck, Hauksson contends, is structural:

High-quality audio data is difficult to obtain and scale for training AI systems. Unlike visual data, acoustic data depends heavily on spatial relationships, environmental acoustics, and device-specific characteristics.

"Treble is part of a growing shift toward physically accurate simulation of sound as a foundation for training audio systems."

"The difference between a robot that works and a robot that is accepted will not be subtle. It will largely come down to how it communicates."

Reykjavík, Iceland-based Treble Technologies is developing a platform for generating physically accurate acoustic data, with tools including the Treble SDK. The platform aims to reduce the sim-to-real gap in acoustic data generation. Hauksson's conclusion is pointed: as vision and locomotion capabilities converge across platforms, communication becomes the differentiator.

Get the long read at The Robot Report