Method, encoder and encoding method for effectively suppressing encoding ringing effect

Abstract

In a method of effectively suppressing a coding ringing effect provided by the invention, an actual quantized loss value is obtained by coding prediction, whether excessive quantized loss appears is judged to determine whether to carry out suppression of the coding ringing effect and a block effect, and extra resources are not occupied; filtering is carried out before coding reconstruction to obtain an optimal result of a current block y, the same is embedded in a block-based coding process of a coder, and the processing accuracy and real-time performance insufficiency caused by disengaging the coder is avoided; and an LMMSE criterion is used for filtering, and for low-bit-rate / very-low-bit-rate applications, the ringing effect and the block effect appearing near strong edges due to suppression on high-frequency signals and excessive quantized loss of high-frequency components can be effectively suppressed. The invention also provides the coder and a coding method, which carry out filtering on an optimal mode of a current block before coding reconstruction to obtain an optimal result, and have higher compression efficiency and better subjective quality performance.

Description

technical field [0001] The invention relates to the field of video encoding and processing, in particular to a method for effectively suppressing encoding ringing effects, an encoder and an encoding method. Background technique [0002] In commonly used video coding, in order to remove the correlation in image signals and reduce the dynamic range of image coding, transform coding and quantization techniques based on macroblocks are usually used. When the quantization parameter of encoding is large, due to the suppression of high-frequency signals and the excessive loss of high-frequency components, ringing effects may appear near strong edges, and serious smearing occurs in moving subjects in moving scenes. The artificial noise produced will reduce the subjective experience of the video. When encoding such video sequences with artificial noise, the encoding performance will also decrease. [0003] At present, the solutions to the ringing effect that may occur near strong e...

AI Technical Summary

Problems solved by technology

When the quantization parameter of encoding is large, due to the suppression of high-frequency signals and the excessive loss of high-frequency components, ringing effects may appear near strong edges, and serious smearing occurs in moving subjects in moving scenes. The artificial noise generated will reduce the subjective experience of the video

When encoding such video sequences with artificial noise, the encoding performance will also decrease

[0003] At present, the solutions to the ringing effect that may occur near strong edges mostly use global statistical information or time-domain correlation information, and use preprocessing methods. Since the entire process is separated from the encoder encoding process, the processing accuracy and real-time performance Insufficient, and additional frame-level preprocessing units need to be added

Images