AV1 Encoder Prediction Circuit Intra-Frame Optimization

Solution Overview

Problem

In the Alliance for Open Media (AOMedia) Video 1 (AV1) encoding format, the need to repeatedly partition super blocks into multiple smaller blocks for prediction, transform, and reconstruction leads to long encoding times and low efficiency due to the requirement for multiple iterations of pixel value reconstruction.

Innovation Solution

An encoder with a prediction circuit that includes a specific partition mode processing circuit and a best partition mode determining circuit, which partitions a super block into specific blocks for intra-frame prediction, generating initial reconstructed pixel values, and then uses these values for subsequent blocks, allowing for reduced waiting time and fewer iterations of transform, quantization, and reconstruction by determining the best partition mode among different partition modes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If the encoder performs pixel value reconstruction multiple times for each partition manner to determine the best coding scheme, then the encoding accuracy is improved, but the encoding time increases and encoding efficiency decreases

Engineering Contradiction:

Improveencoding accuracyVSAvoidencoding time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs pixel value reconstruction in advance for a specific partition mode before determining the best partition mode. By pre-computing the reconstruction data, the encoder avoids repeating this time-consuming operation for each partition manner, thus reducing encoding time while maintaining the ability to achieve accurate reconstruction when the best mode is selected.

Inventive Principle:

Principle #10Preliminary action

2Manufacturing precision

If the encoder performs pixel value reconstruction for each partition manner to ensure accurate reference pixels, then the prediction accuracy is improved, but the device complexity and processing overhead increase

Engineering Contradiction:

Improveprediction accuracyVSAvoidprocessing overhead

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent performs pixel value reconstruction in advance for a specific partition mode before determining the best partition mode. By pre-computing the reconstruction data, the encoder avoids repeating this time-consuming operation for each partition manner, thus reducing encoding time while maintaining the ability to achieve accurate reconstruction when the best mode is selected.

Inventive Principle:

Principle #10Preliminary action

3Manufacturing precision

If the encoder waits for best partition mode determination before performing intra-frame prediction, then the reference pixel accuracy is improved, but the hardware waiting time increases

Engineering Contradiction:

Improvereference pixel accuracyVSAvoidhardware waiting time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs pixel value reconstruction in advance for a specific partition mode before determining the best partition mode. By pre-computing the reconstruction data, the encoder avoids repeating this time-consuming operation for each partition manner, thus reducing encoding time while maintaining the ability to achieve accurate reconstruction when the best mode is selected.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11750800B1Encoder and associated signal processing method

Publication Date: 2023.09.05 REALTEK SEMICON CORP

AI summary

A prediction circuit in an encoder utilizes a specific partition mode to process a super block for generating a plurality of reconstructed pixel values for each block in the super block, and the reconstructed pixel values of each block are directly utilized as reference pixels for other blocks to perform intra-frame prediction, so as to improve the efficiency of the encoder.