Video Coding DPB Parameter Inference for Multi-Layer Bitstream Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in efficiently signaling decoded picture buffer information, leading to suboptimal video decoding and display processes, particularly in multi-layer video coding scenarios where redundant signaling of DPB parameters occurs.
Innovation Solution
The proposed solution involves inferring the index of the dpb_parameters syntax structure for output layer sets when the number of output layer sets with more than one layer equals the number of dpb_parameters syntax structures in the VPS, eliminating the need for explicit signaling and reducing redundant information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If decoded picture buffer parameters are explicitly signaled for each output layer set in multi-layer video coding, then the decoding accuracy is improved, but the bitstream redundancy increases and decoding efficiency deteriorates
Solution Approach 1:
The patent merges the DPB parameter signaling for multiple output layer sets into a single shared syntax structure. When multiple output layer sets reference the same dpb_parameters syntax structure, they share the same parameter values (max_dec_pic_buffering_minus1, max_num_reorder_pics, max_latency_increase_plus1), eliminating redundant signaling while maintaining accurate DPB management for each layer set.
Solution Approach 2:
The dpb_parameters syntax structure is designed to serve multiple output layer sets simultaneously. A single dpb_parameters structure can be referenced by multiple ols_dpb_params_idx syntax elements across different output layer sets, making the parameter signaling universal and applicable to various layer configurations without requiring separate signaling for each layer set.
2Productivity
If the number of dpb_parameters syntax structures equals the number of output layer sets with more than one layer, then the signaling efficiency is improved, but the device complexity increases
Solution Approach 1:
The patent introduces dynamic adaptation in the syntax structure configuration. The number of dpb_parameters syntax structures is not fixed but adapts to the actual number of output layer sets with more than one layer. When this number changes, the system dynamically adjusts the syntax structure count and corresponding indexing, allowing flexible configuration while maintaining manageable complexity through standardized adaptation rules.
3Loss of substance
If redundant DPB parameter signaling is eliminated through inference, then the bitstream size is reduced, but the decoding process complexity increases
Solution Approach 1:
The patent implements preliminary action by pre-defining the relationship between output layer sets and dpb_parameters syntax structures. The ols_dpb_params_idx syntax element is pre-configured to indicate which dpb_parameters structure each output layer set should reference. This preliminary configuration allows the decoder to infer parameters without redundant signaling while following straightforward lookup procedures that minimize decoding complexity.
Data Source
AI summary
This disclosure relates to video coding and more particular to techniques for signaling decoded picture buffer information for coded video data. The method of decoding video data comprises, determining a number of output layer sets with more than one layer; determining a number of decoded picture buffer parameters syntax structures in a video parameter set; in a case that the number of the decoded picture buffer parameters syntax structures is not equal to the number of the output layer sets with more than one layer, decoding a syntax element specifying an index of the decoded picture buffer parameters syntax structures; and in a case that the syntax element is not present, inferring a value of the syntax element.