When a single company collects several patents in the same issue week, the cluster can say more than any one document. On July 7, 2026, the US Patent and Trademark Office issued a set of grants assigned to Meta Platforms Technologies, LLC — the Reality Labs entity — and the group reads like a parts list for a head-mounted display: the ears, the eyes, and the hands.

At the center is US12677106B1, titled "Statistical provisioning of perceptual audio cues to enhance speech." The granted patent is directed to a computer-implemented method for generating spatial audio from a plurality of sources, calculating a pitch similarity and an interaural level difference for each source, and re-rendering the mix so that a target talker is easier to follow. The disclosure describes capturing audio through the microphones of a user's head-mounted display and applying perceptual cues — whispered backgrounds, roughly ten-percent time-dilated vowels, and steepened sound onsets — to separate a voice from competing sound in an artificial-reality environment.

The core of the claimed method is stated plainly in the record:

A computer-implemented method for generating spatial audio, comprising: receiving audio data from a plurality of sources; for each source of the plurality of sources: calculating a pitch similarity for the audio data, the pitch similarity determined in real time to dynamically adjust perceptual audio cues, wherein the perceptual audio cues comprise one or more of whispered backgrounds, time-dilated vowels, and enhanced sound onsets; and calculating an interaural level difference based at least in part on at least one of an attenuation level or a dynamic range of interaural time differences of the audio data; generating spatial audio based at least in part on at least one of the pitch similarity of the audio data or the interaural level difference of the audio data; and causing output of audio, through one or more speakers, based at least in part on the generated spatial audio.— Statistical provisioning of perceptual audio cues to enhance speech, US12677106B1

The named inventors are Antje Ihlefeld and William Owen Brimijoin, II, and the grant is classified under CPC codes including H04S 7/303 and G10L 21/0232 — spatial-audio processing and speech-signal enhancement. In practical terms, the patent describes a computational approach to the cocktail-party problem, using interaural time and level differences as the spatial cues between a listener's left and right ears. It signals that audio intelligibility, not just immersive sound, is a defined workstream inside the Reality Labs audio group.

The cohort maps to a whole device

The audio thread continues in the same week. US12676139B1, issued July 7, is directed to mixed-reality text narration with dynamic text detection — a headset reading on-screen text aloud and re-syncing when that text changes. A week earlier, US12669974B1 (June 30) covered creating custom audio mixes for artificial-reality environments. Placed beside the hero grant, the three describe a stack that captures sound, narrates content, and composes a personalized mix — audio treated as an end-to-end subsystem rather than a single feature.

The optics and hardware side of the cohort fills in the rest of the device. US12674991B2 (July 7) is directed to augmented-reality-glasses temple-arm components, describing a temple housing in which a speaker sits between front and rear battery cells — a packaging detail that ties the audio work back to physical eyewear. US12663619B2 (June 23) claims a wide-field-of-view optical lens assembly with a piezo-actuated tunable lens exceeding 100 degrees of field of view, addressing the display optics a headset presents to the eyes.

Input rounds out the picture. US12663864B2 (June 23) is directed to EMG-based control for interacting with vehicles, describing a wrist-wearable that reads electromyography signals for in-air and surface gestures. Together, the temple hardware, the tunable lens, and the EMG wristband correspond to the ears, eyes, and hands — the three interfaces an AR/VR product has to solve simultaneously.

None of this fixes a delivery date or a product name, and a patent grant records what a company has protected, not what it will ship. But the shape of a same-week cohort is itself information. This one concentrates on the sensor-and-output layer of head-mounted hardware, with a specific emphasis on making speech legible in noisy, multi-source scenes. For a portfolio watcher, the grants suggest that Reality Labs continues to file broadly across the physical device while sharpening a distinct claim to psychoacoustic audio processing — the part of the experience that decides whether a wearer can actually understand what they hear.