This session introduces the development of a new Supplemental Enhancement Information (SEI) message by the Joint Video Experts Team (JVET), set to be included in the upcoming version of the Versatile Supplemental Enhancement Information (VSEI) standard, expected in 2025. The Constituent Rectangles SEI message offers a standardized approach to describing video content composed of multiple frame-packed rectangles, which could otherwise have been encoded separately in a multi-layer bitstream. This SEI message supports various use cases, such as VR/AR, 3D, free viewpoint video, and cloud gaming, by enabling efficient signaling of parameters like size, position, and type for each constituent rectangle within a picture. It eliminates the need for multi-layer codec profiles, providing a flexible, general-purpose method for handling frame-packed content in a single-layer video bitstream. The session will cover the SEI message’s design, its software implementation, and example use cases, showcasing its potential to enhance video coding workflows across a wide range of applications.
JVET is developing a Constituent Rectangles SEI message for inclusion in a future version of VSEI, the companion standard to VVC The Constituent Rectangles SEI message provides a standardized way to describe a video containing pictures formed by frame packing of multiple constituent rectangles of a variety of types Example usages include multi-view, video + alpha, video + depth The SEI message enables use of a single layer video codec/profile, without requiring use of a multi-layer video codec/profile