Video Codec Buffer Reduction via Mode-Specific Offsets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video encoding and decoding systems face challenges in reducing memory usage due to the large number of buffers required for various prediction modes, leading to increased memory consumption and costs.
Innovation Solution
Implementing a method that determines the prediction mode for frame data and selectively uses a mode agnostic buffer bank and mode specific buffers to retrieve and generate historical data, reducing the number of buffers needed by using fewer buffer banks than prediction modes, and generating historical data on the fly when necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a separate buffer is allocated for each prediction mode, then prediction accuracy is improved, but memory consumption increases
Solution Approach 1:
The patent implements a universal buffer structure where a single buffer can serve multiple prediction modes through mode-specific offset adjustments. The buffer is designed to accommodate different prediction modes (intra, inter, skip, etc.) by applying mode-dependent offsets to the same underlying storage structure, allowing one buffer to perform the functions previously requiring multiple separate buffers.
Solution Approach 2:
The patent changes the parameter organization within the buffer by introducing mode-specific offsets that modify how data is accessed and interpreted. Instead of allocating separate physical buffers for each mode, the system uses parameter adjustments (offsets) to differentiate mode-specific data within a unified buffer structure, thereby reducing memory usage while maintaining mode-specific prediction accuracy.
2Quantity of substance
If the number of buffers is reduced to lower memory usage, then memory efficiency is improved, but data retrieval complexity increases
Solution Approach 1:
The patent segments the unified buffer into mode-specific regions using offset calculations. Each prediction mode has an associated offset that segments the buffer space logically, allowing the system to retrieve data for specific modes without requiring physically separate buffers. This segmentation approach maintains organized data access while reducing overall buffer quantity.
Solution Approach 2:
The patent introduces mode-specific offsets as intermediary values that mediate between the unified buffer structure and mode-specific data retrieval needs. These offsets act as a translation layer that converts mode identification into specific buffer access positions, simplifying the retrieval process despite the unified buffer structure.
3Reliability
If mode-specific buffers are used for all prediction modes, then prediction quality is maintained, but device complexity increases
Solution Approach 1:
The patent merges multiple mode-specific buffer requirements into a single unified buffer structure. By combining what would traditionally be separate buffers for intra prediction, inter prediction, skip mode, and other prediction types into one buffer with mode-specific offset management, the system maintains prediction quality while reducing buffer management complexity.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reducing a quantity of buffers for a video codec. One of the methods includes determining, from a plurality of prediction modes, a prediction mode for data that represents frame data in a frame; in response to determining the prediction mode, selecting, using the prediction mode, one or more buffers from a plurality of buffers, each buffer of which is for a prediction mode from the plurality of prediction modes, a first quantity of buffers in the plurality of buffers being less than a second quantity of prediction modes in the plurality of prediction modes; retrieving, from each of the one or more buffers, historical data for the frame data; and in response to retrieving the historical data, generating, using the historical data, updated data for the frame data in the frame of the video sequence.