VPS DPB and PTL Signaling for Multi-Layer Video Bitstreams

The patent segments the bitstream into multiple syntax structures (VPS, SPS, PPS, slice headers, etc.) with hierarchical organization. Each syntax structure contains specific parameter sets that can be independently coded and transmitted. This segmentation allows for selective transmission of essential information while omitting redundant data, thereby reducing the total amount of transmitted information while maintaining image quality.

The patent employs dynamic parameter adjustment through multiple syntax structures that allow different parameter sets to be applied to different regions or sequences of images. By changing parameters such as prediction modes, transform types, and quantization levels adaptively, the system achieves high image quality where needed while using more compressed representations where possible, thus reducing overall transmission requirements.

The patent creates a universal syntax structure framework where VPS contains multiple DPB and PTL syntax structures that can serve multiple output layer sets (OLS). Each syntax structure is designed to be multi-functional, supporting different layer configurations and prediction modes through a unified representation. This universality reduces the need for separate dedicated syntax structures for each OLS, thereby managing device complexity while maintaining encoding/decoding precision.

The patent performs preliminary signaling of DPB and PTL parameters in the VPS syntax structure before actual encoding/decoding operations. By pre-defining and signaling these parameter sets in advance, the system establishes a framework that guides subsequent processing without requiring complex real-time decisions. This preliminary action reduces runtime complexity while ensuring precise control over encoding/decoding operations.