Adeia

Two Streaming patents tackle different aspects of the viewer experience during gameplay. One application describes a system that monitors a player's real-time performance metrics and automatically surfaces relevant tutorial videos from a dynamically updated database, removing the need for players to pause and search for help. The system stitches together video segments across different sub-levels and continuously prunes its library by deleting tutorials when player performances exceed stored benchmark metrics. The other patent enables viewers watching a livestream to scan the screen and immediately launch into the same game session at the identical state being shown, without requiring account logins or Platforms authentication. This approach embeds game-state metadata directly into video frames, allowing stateless session reconstruction that sidesteps traditional barriers like QR codes or pre-registration.

Cloud gaming receives 2 patent applications focused on reducing latency through predictive encoding architectures. Both describe a system that runs multiple encoder instances in parallel, each configured with different quantization parameters, while a variational autoencoder analyzes incoming frames to predict scene complexity and determine the optimal number of encoders to activate. Instead of using traditional trial-and-error rate control methods, the deep learning model forecasts encoding requirements before frames arrive, allowing the system to select the best-performing output without re-encoding. This approach targets real-time video delivery scenarios where maintaining consistent quality within strict bitrate caps is essential, particularly in environments where even small delays degrade the user experience.