Following a conversation can be difficult when background noise overwhelms the speaker’s voice. Existing wearable devices, including hearing aids and smart glasses with multi-microphone arrays, use techniques such as beamforming and denoising to improve speech comprehension, but these systems can struggle in environments exceeding 75 dB or when the signal-to-noise ratio is particularly poor.
Meta’s new patent application describes a smart-glasses system that takes a different approach. The glasses capture and clean the speaker’s voice before sending the enhanced audio directly to another person’s earbuds, hearing aids or other hearables. The system could make conversations easier to follow in noisy environments while reducing the need to rely on excessive amplification, which the application identifies as a potential risk to hearing health.
The system combines smart glasses, microphones, audio processing and a wireless connection with a wearable audio device. Rather than relying primarily on the listener’s device to pick up a speaker’s voice from a noisy environment, the glasses worn by the speaker capture and process the voice before transmitting it to the listener.
The smart glasses can establish a direct wireless connection with another user’s wearable audio device, including through a low-latency Bluetooth link. The patent also describes compatibility across different generations and manufacturers through negotiation of supported communication protocols and audio codecs. The glasses can use a mouth-directed beamformer to emphasize sound around the wearer’s mouth while reducing sounds from other directions. They can then apply denoising techniques, including deep neural network-based noise reduction and spatial noise reduction, to further improve speech clarity.
The processed voice can be transmitted with control-channel metadata containing information such as ambient noise levels, SNR estimates or recommended mixing parameters. The receiving device can use this information to adjust active noise cancellation, transparency and the balance between the transmitted voice and locally captured ambient sound. In some configurations, the system can support multiple speakers and listeners. Enhanced voice audio can be transmitted to several wearable devices through a group communication session, while an AI assistant can help with device pairing, playback modes and group conversations.
Why it matters
The application points to a broader role for Meta’s smart glasses as an interface between people and other wearable devices.
Meta’s AI glasses lineup includes the Ray-Ban Meta and Oakley Meta HSTN. The Ray-Ban Meta combines cameras, microphones, speakers, Bluetooth connectivity and Meta AI for functions such as capturing photos and videos, making calls and interacting with the assistant. The Oakley Meta HSTN builds on this platform with a sport-oriented design, longer battery life, higher-resolution cameras and Meta AI features tailored to athletic use.
The proposed system uses these capabilities in a different way. Instead of serving primarily as a device for the wearer, the glasses can capture and process one person’s voice and deliver enhanced audio to another person’s wearable device. This positions the glasses as a communication bridge between people and their connected wearables, adding another potential use for Meta’s wearable platform.
U.S. Pat. App. Pub. No. 2026/027630, titled “Smart Glasses Systems for Hearable Conversations,” was filed on March 2, 2026 and published on September 10, 2026. The listed investors are Sam Alex Padinjaremannil, Thomas Ivan Harvey, Kanji Mavji Kerai, Christopher Corson, Guoqing Li, Nils Thomas Fritiof Lunner, Morteza Khaleghi-Meybodi, Poornima Shankar, Bhramara Tirupati, Vladimir Tourbabin, and Kevin Winter.