RTP Packet Timestamp Difference for B-Frame Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RTP-based systems lack compatibility with bi-directional predicted frames (B-frames), limiting their support to only intra-coded frames (I-frames) and predicted frames (P-frames).
Innovation Solution
A method for visual data processing that involves converting a bitstream of a current frame into RTP packets, where each packet includes a first indication of the difference between the decoding timestamp and the presentation timestamp, enabling correct decoding of B-frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RTP-based system uses conventional conversion method, then I-frames and P-frames can be decoded correctly, but B-frames cannot be decoded correctly
Solution Approach 1:
The patent modifies the RTP packet structure by adding a new parameter (first indication field) that carries timestamp difference information. This parameter change enables the system to distinguish and correctly decode B-frames by providing the necessary timestamp adjustment data, thereby expanding frame type support while maintaining decoding correctness.
Solution Approach 2:
The patent introduces an intermediary conversion process between the visual data codec and RTP protocol. This intermediary layer handles the timestamp difference calculation and packet construction, acting as a mediator that translates B-frame decoding requirements into RTP-compatible packet structures with appropriate indication fields.
2Adaptability or versatility
If RTP-based system supports only I-frames and P-frames, then system complexity is low, but compatibility with state-of-the-art visual data coding is limited
Solution Approach 1:
The patent segments the timestamp information into separate components: the RTP timestamp and a separate first indication field containing the timestamp difference. This segmentation allows B-frame support to be added without fundamentally redesigning the entire RTP protocol, thereby improving compatibility while controlling complexity growth.
Solution Approach 2:
The patent applies partial action by adding only the necessary minimal indication field to support B-frames, rather than implementing a complete redesign of the RTP protocol. This partial enhancement provides sufficient functionality for B-frame decoding while avoiding excessive complexity changes to the overall system.
Data Source
AI summary
Embodiments of the present disclosure provide a solution for visual data processing. A method for visual data processing is proposed. The method comprises: performing a conversion between a bitstream of a current frame of visual data and at least one packet in accordance with a real-time transport protocol (RTP), wherein the at least one packet comprises a first indication indicating a difference between a decoding timestamp of the current frame and a presentation timestamp of the current frame.


