Video Decoder Bit-Depth Reconstruction via Residual Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Decoders designed for lower bit-depth videos struggle to accurately reconstruct videos encoded at higher bit-depths, leading to reduced fidelity and visual quality issues when processing coefficients with higher precision.
Innovation Solution
A method is introduced that involves a residual process including inverse-quantization and inverse-transformation processes to compute (D+1)-bit residual samples, followed by a prediction process with rounding and/or truncation to produce d-bit reconstructed samples, allowing decoders to handle coefficients originally encoded at higher bit-depths by adapting bit-depth precision during the decoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a decoder designed for lower bit-depth (d-bit) processes coefficients encoded at higher bit-depth (D-bit), then device complexity and resource requirements are reduced, but manufacturing precision and visual quality deteriorate due to bit-depth mismatch
Solution Approach 1:
The patent changes the bit-depth parameter dynamically during the decoding process. The residual samples are computed at higher precision (D+1 bits) to maintain accuracy, then selectively truncated or rounded to the target bit-depth (d bits) for the final reconstructed samples. This parameter transformation allows the decoder to process high-precision coefficients while outputting low-precision results suitable for the target display device.
Solution Approach 2:
The decoding process is segmented into distinct stages with different precision requirements: (1) inverse quantization and inverse transformation stages that compute residual samples at high precision (D+1 bits), and (2) a final rounding/truncation stage that converts to target precision (d bits). This segmentation allows each stage to operate at optimal precision, maintaining quality where needed while reducing complexity at the output stage.
2Productivity
If higher bit-depth coefficients are processed with lower precision operations, then processing speed and energy consumption improve, but measurement precision and visual quality worsen
Solution Approach 1:
The patent applies different precision parameters to different stages of processing. Intermediate residual samples are maintained at high precision (D+1 bits) to preserve measurement accuracy, while final reconstructed samples are converted to lower precision (d bits) to improve processing speed and reduce energy consumption. This selective parameter application optimizes both speed and accuracy.
3Adaptability or versatility
If bit-depth is reduced from D-bit to d-bit during decoding, then adaptability to low-bit-depth devices improves, but loss of information increases due to precision reduction
Solution Approach 1:
The patent performs preliminary high-precision computation of residual samples before the final bit-depth conversion. By computing residuals at D+1 bits and only then truncating or rounding to d bits, the system preserves as much information as possible during intermediate calculations, minimizing information loss while achieving compatibility with low-bit-depth devices.
Solution Approach 2:
The high-precision residual samples act as an intermediary between the high-bit-depth encoded coefficients and the low-bit-depth reconstructed output. This intermediary representation maintains higher precision during processing, allowing accurate computation before the final conversion to the target bit-depth, thereby reducing information loss.
Data Source
AI summary
Methods are provided for reconstructing d-bit samples from coefficients, in an encoded video bit-stream, that had been encoded based on source samples having a bit-depth D, where D>d. A residual process, including inverse-quantization and inverse-transformation processes, is applied to the coefficients to compute (D+1)-bit residual samples. Then a prediction process, including (i) rounding and/or truncating intermediate samples and (ii) subsequently clipping intermediate samples, is applied to the residual samples to compute d-bit reconstructed samples.