Inverse transforming method and device for transform scanning table in video decoding

Abstract

The present invention relates to the field of multimedia video technology, and discloses an AVS inter-frame forecasting pixel generating device, including a main state controller, a dual-port FIFO, a data cache register pile, a brightness forecasting pixel generating unit, chrominance forecasting pixel generating unit and a weighting forecasting unit. The device integrates a dual-port memory, can simultaneously read data, store the input predicted data and calculate 1 / 4 or 1 / 8 accuracy forecasting pixel; 1 / 4 pixel accuracy forecast adopts a three-stage flowing water structure, 1 / 8 pixel accuracy forecast adopts a two-stage flowing water structure; each clock period produces a forecast pixel generating result; the adoption of the three-stage flowing water pixel generating structure improves the performance frequency; and a three-stage multiplexed filtering unit is adopted to calculate 1 / 4 pixel forecast. The inter-frame forecast pixel generating structure and method, which has high-performance performance and smaller area, is applied to 1 / 4 accuracy and 1 / 8 accuracy inter-frame forecast pixel generating process in other video standards.

Description

technical field [0001] The invention relates to the technical field of multimedia video, in particular to an inverse transformation method and device for transforming a scanning table in video decoding. Background technique [0002] Two-dimensional transformation is usually used in image and video transformation coding blocks, such as digital audio and video codec standards JPEG, MPEG, H.264, AVS, etc. domain, reducing the temporal and spatial correlation between adjacent pixels. [0003] Two-dimensional transformation is further divided into discrete cosine (DCT) transformation and integer transformation. The DCT transform is usually used in image coding standards such as JPEG and MPEG, and it is closest to the statistically optimal Karhunen-Loeve transform in most signal classifications. DCT maps an N-dimensional vector x to a new vector X, and the transformation parameters are determined by linear transformation X=Hx. However, one disadvantage of DCT transformation is t...

AI Technical Summary

Problems solved by technology

[0005] On the decoder side, the inverse transformation includes multiple multiply-accumulate operations, which has a large amount of calculation, especially for high-definition video decoding. In order not to increase the burden on the processor, it is often implemented by very large-scale integrated circuit (VLSI) hardware. The hardware implementation is fast and efficient. However, a large number of multipliers are required for parallel multiplication and accumulation operations and a large amount of memory is used to store the intermediate results of the operation.

In the general inverse transformation module, it is necessary to perform two operations of saving the result and converting the intermediate result, which greatly increases the storage unit required for the inverse transformation, and the two operations of the transformation also greatly increase the delay of the inverse transformation operation.

Images