Video Decoder Bit-Depth Reconstruction Rounding Error Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Decoders designed for lower bit-depth struggle to accurately reconstruct videos encoded at higher bit-depths due to accumulation of rounding errors and incompatibility issues, leading to reduced video fidelity.

Innovation Solution

A method that applies a residual process including inverse-quantization and inverse-transformation to compute (D + 1)-bit residual samples, followed by rounding and truncation to produce d-bit reconstructed samples, allowing decoders to handle coefficients encoded at higher bit-depths while maintaining acceptable video quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a decoder designed for lower bit-depth (d) is used to decode coefficients encoded at higher bit-depth (D), then device complexity and resource requirements are reduced, but video fidelity and reconstruction accuracy deteriorate due to accumulation of rounding errors

Engineering Contradiction:
Improvedecoder complexityVSAvoidvideo fidelity
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The decoding process is segmented into distinct stages: inverse quantization, inverse transformation, and rounding. By separating these operations, the patent applies different bit-depth handling strategies to each stage, allowing the residual process to maintain higher precision (D+1 bits) while the final output is rounded to match the decoder's native bit-depth (d bits), thus resolving the contradiction between device simplicity and reconstruction accuracy

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent dynamically adjusts bit-depth parameters throughout the decoding process. The residual process operates at (D+1) bits to preserve precision, intermediate values are rounded to (D+2) bits, and final output is truncated to d bits. This parameter transformation approach allows a simple d-bit decoder to accurately reconstruct D-bit encoded video by temporarily using higher precision in critical computation stages

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If rounding operations are applied during decoding to adapt to lower bit-depth, then compatibility with existing decoders is improved, but rounding errors accumulate and reduce video quality

Engineering Contradiction:
Improvedecoder compatibilityVSAvoidreconstruction accuracy
Core Design Contradiction:
Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent applies preliminary rounding only to specific intermediate values (the residual process output) rather than to all intermediate calculations. By rounding selectively at the optimal point in the decoding pipeline, the method maintains maximum precision during critical inverse transformation operations while still producing compatible d-bit output, thereby minimizing cumulative rounding errors

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The residual process acts as an intermediary between the high-precision encoded coefficients and the low-precision decoder. It computes residual samples at (D+1) bits, rounds them to (D+2) bits, and combines them with prediction samples to produce the final d-bit output. This intermediary approach allows precision to be maintained where needed while ensuring compatibility with simple decoders

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP2984835B1Methods for reconstructing an encoded video at a bit-depth lower than at which it was encoded
Publication Date: 2019.06.26 BLACKBERRY LTD
  • EP2984835B1 patent drawingFigure 1a
  • EP2984835B1 patent drawingFigure 1b
  • EP2984835B1 patent drawingFigure 1c

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.