This month's audio and sound category includes 3 filed patent applications from 2 companies: Sony with 2 filings and Meta with 1.
Meta's application describes a spatial audio system that adjusts sound placement based on head movements to maintain 3D positioning in virtual and augmented reality environments. Sony's filings cover a machine learning text-to-speech system that allows phoneme-level pitch adjustment for game character voices and an AI-powered system that dynamically balances in-game audio with external audio streams during gameplay.
Meta received 1 patent addressing a persistent challenge in VR and AR audio: keeping sound sources properly positioned as users move their heads. The system uses two complementary algorithms working together, one continuously updating audio placement based on head angle changes while the other applies adaptive recentering filters. This dual approach helps prevent the audio drift and disorientation that typically occur in head-tracked audio systems when users make natural head movements.
Sony's 2 patents both tackle audio challenges in gaming contexts. The first covers a text-to-speech system that gives developers granular control over synthesized character voices, allowing them to adjust pitch on individual phonemes rather than across entire utterances. The system generates initial pitch predictions using convolutional neural networks but lets developers override those predictions through interface controls, combining automated efficiency with manual creative direction for emotional dialogue delivery. The second patent describes an AI system that monitors gameplay to identify which audio elements matter most at any given moment, such as boss encounter outcomes or teammate communications, then automatically adjusts volume levels across competing audio sources to prevent important sounds from being obscured while keeping players immersed in the experience.