Dynamic Digital Signal Encoding for Stable Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current compression methods for high-definition digital television signals result in loss of high-frequency information, leading to image degradation, and existing encoding/decoding methods do not effectively stabilize compression rates without significant computational power.
Innovation Solution
A method and device for dynamically encoding digital multidimensional signals, such as image or audio signals, by performing localized encodings and on-the-fly analysis to adjust encoding parameters, allowing for stabilization of compression rates or noise levels at a target value, using techniques like blue noise masking and gradient-based adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional compression methods are applied to HDTV signals, then transmission efficiency is improved, but high-frequency information is lost causing image degradation
Solution Approach 1:
The patent transforms the image signal from spatial domain to frequency domain using 2D Discrete Cosine Transform (DCT), changing the representation parameters from pixel values to frequency coefficients. This allows selective compression of high-frequency components while preserving low-frequency information, achieving both compression and acceptable image quality
Solution Approach 2:
The patent extracts and separately processes different frequency components of the image signal. By identifying and isolating high-frequency components that contribute less to perceived image quality, the system can compress or discard these components more aggressively without significantly degrading the overall image quality
2Quantity of substance
If compression rate is increased to reduce memory capacity, then storage efficiency is improved, but image degradation increases
Solution Approach 1:
The patent applies quantization matrices with different scaling factors to compress the frequency domain representation. By adjusting the quantization step size, the system can control the compression ratio while managing the loss of high-frequency information, thus balancing memory usage and image quality
Solution Approach 2:
The patent processes different regions of the frequency spectrum with different compression aggressiveness. Low-frequency components that are critical for image quality are preserved with higher precision, while high-frequency components are compressed more heavily, achieving efficient memory utilization with acceptable overall quality
3Productivity
If localized encoding with sample-by-sample processing is applied, then compression rate is improved, but computational complexity increases
Solution Approach 1:
The patent divides the image signal into 8x8 pixel blocks and processes each block independently through the DCT transform. This segmentation allows parallel processing of multiple blocks and reduces the computational burden compared to processing the entire image at once, while still achieving effective compression
Solution Approach 2:
The patent applies zigzag scanning and run-length encoding to compress the DCT coefficients further. By exploiting the spatial correlation in the frequency domain and the fact that many high-frequency coefficients are zero or near-zero after quantization, the system achieves additional compression with minimal additional computational complexity
Data Source
AI summary
A method is provided for encoding a digital signal as an encoded signal. The method includes performing a plurality of localized encodings of a digital signal to generate a set of encoded local signals. Localized encodings are performed for a first sample of the digital signal. A plurality of physical quantities is assigned to the first sample. The set of encoded local signals includes an encoded local signal associated with each physical quantity of the plurality of physical quantities. The method further includes analyzing a characteristic associated with an encoded signal to determine a measured value of the characteristic. The encoded signal includes the set of encoded local signals. The method also includes adjusting a first encoding parameter associated with the plurality of localized encodings according to the measured value of the characteristic and a target value of the characteristic. The first encoding parameter is adjusted for a second sample of the digital signal. The second sample is processed after the first sample.