Temporal Similarity Quantization for Scalable Signal Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal processing systems face inefficiencies in bitrate allocation and quality enhancement for compressed signals, particularly in scalable encoding techniques, leading to inefficient use of bandwidth and potential losses in visual quality due to inadequate quantization strategies.
Innovation Solution
The system determines quantization parameters based on the similarity between data elements across different time samples of a signal, allowing for dynamic bitrate allocation and improved quality reconstruction by generating output data using these parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If quantisation is used to reduce data amount, then data compression is improved, but signal quality deteriorates
Solution Approach 1:
The patent applies dynamic quantisation parameter adjustment based on temporal similarity analysis. The system continuously adapts quantisation parameters for different time samples based on their similarity to reference samples, making the quantisation process dynamic rather than static. This allows the system to use coarser quantisation for similar samples (reducing data amount) while maintaining finer quantisation for dissimilar samples (preserving signal quality).
Solution Approach 2:
The patent changes the quantisation parameter based on the derived measure of temporal similarity. When temporal similarity is high, the system applies stronger quantisation (larger parameter values) to reduce data amount. When temporal similarity is low, the system applies weaker quantisation (smaller parameter values) to preserve signal quality. This parameter adaptation resolves the contradiction between compression efficiency and quality preservation.
2Adaptability or versatility
If scalable encoding techniques are used to encode at multiple quality levels, then quality flexibility is improved, but data amount increases
Solution Approach 1:
The patent segments the encoding process into base layer and enhancement layer components. The base layer contains essential signal information encoded with coarser quantisation, while the enhancement layer contains additional detail information encoded with finer quantisation. This segmentation allows the system to provide multiple quality levels without transmitting all data at maximum quality, thus reducing overall data amount while maintaining quality flexibility.
Solution Approach 2:
The patent applies different quantisation qualities to different temporal segments based on their similarity characteristics. Time samples with high temporal similarity are encoded with coarser quantisation, while those with low similarity are encoded with finer quantisation. This local quality adaptation provides quality flexibility where needed while reducing data amount where possible.
3Manufacturing precision
If higher quality video encoding is used, then visual quality is improved, but bandwidth usage increases
Solution Approach 1:
The patent dynamically adjusts the encoding quality for different time samples based on temporal similarity analysis. For time samples that are similar to reference samples, the system uses coarser quantisation parameters, reducing bandwidth usage. For time samples that are dissimilar and require higher visual quality, the system uses finer quantisation parameters. This dynamic adaptation maintains visual quality where necessary while reducing overall bandwidth consumption.
Solution Approach 2:
The patent changes quantisation parameters based on the derived measure of temporal similarity between current and reference time samples. When similarity is high, larger quantisation step sizes are applied, reducing bandwidth usage. When similarity is low, smaller quantisation step sizes are applied, preserving visual quality. This parameter adaptation resolves the contradiction between visual quality and bandwidth usage.
Data Source
AI summary
A first value of a first data element in a first set of data elements is obtained, the first set of data elements being based on a first time sample of a signal. A second value of a second data element in a second set of data elements is obtained, the second set of data elements being based on a second, later time sample of the signal. A measure of similarity is derived between the first value and the second value. Based on the derived measure, a quantisation parameter useable in performing quantisation on data based on the first time sample of the signal is determined. Output data is generated using the quantisation parameter.