This month's Audio & Sound category includes 3 granted patents from Activision Blizzard (1), NetEase (1), and Voyetra Turtle Beach (1).
Activision Blizzard's patent details a neural network trained on recorded player voice reactions that generates contextual audio for non-player characters during similar gameplay events. Voyetra Turtle Beach's filing covers a headset system that analyzes game audio streams, identifies specific sound patterns, and delivers real-time voice command alerts to players. NetEase's patent describes spatial audio processing for projectiles that adjusts sound properties based on object orientation as items move in and out of a player's detection range.
Activision Blizzard received 1 patent for a system that records authentic player voice reactions during gameplay, then uses those recordings to train a neural network that generates realistic NPC audio. The system separates the voice capture process from the game instance itself, correlating player audio with specific in-game events over time. Instead of relying on scripted dialogue, the trained network can produce event-triggered NPC sounds based on how real humans actually reacted in those same situations.
Turtle Beach's patent covers a headset that performs its own audio analysis and delivers voice coaching to players without requiring any integration with the game software. The device listens to the game's audio output, detects specific sound characteristics like directionality, intensity, and frequency, and responds by speaking pre-mapped voice commands or alerts. This headset-native intelligence loop operates independently of the game, allowing the hardware to coach players in real time based purely on what it hears.
NetEase received 1 patent addressing spatial audio for fast-moving projectiles and items in games. The system calculates how sound should change as an object travels through a player's detection zone, encoding the item's orientation and movement path into the audio signal rather than playing a static sound at a single fixed point. This approach solves the problem of representing high-speed virtual objects that move too quickly for traditional position-based audio playback to accurately capture.
All data sourced from USPTO patent filings. Google Patents may take several weeks to index recent publications. If a link is unavailable, search for the patent number at USPTO Patent Public Search.