Encoder CTU Quantization Parameter Control for Buffer Overflow Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In low-latency applications, existing bit rate control methods for video encoders, such as CBR and VBR, face challenges in managing the varying bit counts of CTUs with complex or simple textures, leading to potential buffer overflow due to limited buffer capacity.
Innovation Solution
An encoder system that calculates energy parameters for each CTU and determines suitable quantization parameters to control the bit rate by adjusting quantization operations across CTU sets and rows, using methods like row-level, CTU-set-level, and CTU-level quantization parameter determination to ensure the bit rate remains within buffer limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CBR or VBR control method is used, then bit rate control is achieved, but buffer overflow occurs due to limited buffer capacity in low-latency applications
Solution Approach 1:
The patent divides the frame into multiple rows, and each row into multiple CTUs, enabling hierarchical bit rate control. By controlling bit rates at the row level and CTU level separately, the system can manage buffer usage more granularly, preventing buffer overflow while maintaining overall bit rate control in low-latency scenarios.
Solution Approach 2:
The patent calculates energy parameters for each CTU before encoding and determines quantization parameters in advance based on these energy parameters and row-level bit rate constraints. This preliminary determination of quantization parameters ensures that the bit rate for each CTU is controlled before encoding occurs, preventing buffer overflow.
2Reliability
If quantization parameter is not properly set for CTU with complex textures, then encoding quality is maintained, but bit count exceeds buffer capacity upper limit
Solution Approach 1:
The patent determines quantization parameters independently for each CTU based on its local energy parameters and texture complexity. CTUs with complex textures receive appropriately higher quantization parameters to control their bit counts, while simple CTUs use lower quantization parameters to maintain quality, achieving localized bit rate control that prevents buffer overflow.
Solution Approach 2:
The patent dynamically adjusts quantization parameters based on energy parameters calculated from the image data. By changing the quantization parameter according to the local content characteristics (energy parameters), the system controls the bit count of each CTU to stay within buffer capacity limits while maintaining encoding quality.
3Measurement precision
If row-level and CTU-set-level quantization parameter determination is used, then bit rate control precision is improved, but device complexity increases
Solution Approach 1:
The patent implements a segmented control structure with multiple levels: frame-level, row-level, and CTU-level quantization parameter determination. Each level operates independently with its own control logic, allowing precise bit rate control at each granularity level while keeping the overall system manageable through modular organization.
Solution Approach 2:
The patent uses a unified energy parameter calculation approach that works across all CTUs regardless of their position or complexity. The same basic methodology (calculating energy parameters and determining quantization parameters based on these energies and bit rate constraints) is applied universally at row-level and CTU-set-level, simplifying the control mechanism while achieving precise bit rate control.
Data Source
AI summary
The present invention provides an encoder including a quantization circuit, an encoding circuit, an energy parameter calculation circuit and a quantization parameter determination circuit. The quantization circuit is configured to perform quantization operations on a plurality of CTUs in image data in sequence to generate quantized data respectively corresponding to the plurality of CTUs. The encoding circuit is configured to perform encoding operations on the quantized data of the plurality of CTUs in sequence to generate encoded data. The energy parameter calculation circuit is configured to receive the image data, and calculate a plurality of energy parameters respectively corresponding to the plurality of CTUs in the image data. The quantization parameter determination circuit is configured to determine a plurality of quantization parameters of the plurality of CTUs according to at least a portion of the plurality of energy parameters, for the quantization circuit to perform the quantization operations.