CN103491373B - A kind of level Four flowing water filtering method of deblocking filter suitable for HEVC standard - Google Patents

Abstract

The invention belongs to the level Four flowing water filtering method of high-definition digital video compression coding and decoding technical field, specially a kind of deblocking filter suitable for HEVC standard.If in original video stream, each image includes a luminance component Y, each luminance component corresponds to two chromatic components Cb, Cr；In deblocking filter module, to all right filtering process of Y, Cb and Cr, minimum basic unit per treatment is the side of a 4x4 pixel block；The present invention is filtered processing by level Four pipeline mode for luminance component Y, and two chromatic components Cb, Cr to the side of 4x4 pixel blocks.The present invention can effectively reduce processing time, the maximum operating frequency and chip processing capabilities of raising property chip, so as to efficiently realize the real-time coding of HD video.

Description

A four-stage pipeline filtering method suitable for HEVC standard deblocking filter

technical field

The invention belongs to the technical field of high-definition digital video compression encoding and decoding, and is aimed at the HEVC video encoding and decoding standard, and in particular relates to a four-stage pipeline filtering method applicable to a deblocking filter of the HEVC standard.

Background technique

HEVC (High Efficiency Video Coding) is a next-generation video codec standard proposed by JCTVC, an organization jointly established by the International Telecommunications Organization (ITU) and the Motion Picture Experts Group (MPEG). The goal is to double the compression rate compared to the previous generation standard H.264/AVC under the premise of the same visual effect.

The HEVC-based video encoder, as shown in Figure 1, is mainly composed of the following modules: intra prediction, inter prediction, transformation, quantization, inverse quantization, inverse transformation, reconstruction, deblocking filter, auto Compatible with sample point compensation and other modules. The basic process of video compression coding can be summarized as follows: 1. Use intra prediction or inter prediction to predict the pixels of the current original video stream; 2. Subtract the original pixel value from the predicted pixel value to obtain the residual value; 3. Transform and quantize the residual to obtain the output residual coefficient and then enter CABAC (Context-based Adaptive Binary Arithmetic Coding) entropy coding to form the final compressed output code stream; 4. The residual coefficient is dequantized and inversely transformed Processing, and then added to the previously obtained predicted pixels to obtain reconstructed pixels, which are stored as predicted reference frame pixels.

The processing unit block in HEVC introduces a quadtree structure. The maximum size of an image processing block is 64×64. Blocks are combined and processed separately. The encoding end needs to traverse all division situations of the block once to determine which division situation is best handled.

The deblocking filter module is shown in Figure 1. This module mainly judges the characteristics of the pixel values on both sides of the boundary between the PU (prediction unit) and the TU (transformation unit), and filters the pixels on both sides of the boundary to eliminate the distortion caused by the block coding method. Some border effects are caused, so that the reconstructed image has better quality and PSNR value (Peak Signal-to-Noise Ratio).

In the standard reference software HM9.0, this module first processes the vertical edge of the entire image falling on the 8x8 boundary, and then processes the horizontal edge of the entire image falling on the 8x8 boundary. However, this kind of filtering process is very complicated, which will cause the maximum operating frequency of the chip to be very low, and it is impossible to efficiently encode the video stream in real time.

Contents of the invention

The purpose of the present invention is to propose a four-stage pipeline filtering method that can overcome the deficiencies of the prior art and can be effectively applied to the deblocking filter of the HEVC standard.

In the usual original video stream, each image includes a luminance component, denoted as Y, and each luminance component corresponds to two chrominance components, denoted as Cb and Cr respectively. In the deblocking filter module, Y, Cb and Cr are all filtered. The minimum basic unit of each processing is a 4x4 block side, as shown in Figure 2, where the thick black line is the 4x4 block side that needs to be filtered, and the P block and Q block are adjacent 4x4 pixel point blocks on both sides of the side. If the current edge is a vertical edge, the P block and the Q block are the adjacent 4x4 pixel blocks on the left and right sides of the edge respectively; if the current edge is a horizontal edge, the P block and the Q block are the adjacent 4x4 pixel blocks above and below the edge respectively Blocks of pixels.

As shown in Figure 2, set:

p3,0, p2,0, p1,0, p0,0, q0,0, q1,0, q2,0, q3,0,

p3,1, p2,1, p1,1, p0,1, q0,1, q1,1, q2,1, q3,1,

p3,2, p2,2, p1,2, p0,2, q0,2, q1,2, q2,2, q3,2,

p3,3, p2,3, p1,3, p0,3, q0,3, q1,3, q2,3, q3,3,

The square matrix on the left is the value of 16 corresponding pixel points in the 4x4 P block; the square matrix on the right is the value of the 16 corresponding pixel points in the 4x4 Q block; among the two square matrices, the 8 in the first row The pixels are the first group, the 8 pixels in the second row are the second group, the 8 pixels in the third row are the third group, and the 8 pixels in the fourth row are the fourth group;

The four-stage flowing water filtering method proposed by the present invention can not only filter the Y component, but also filter the Cb and Cr components. The specific steps of the four-stage flowing water filtering method of the present invention are as follows:

(1) Level 1

Send a signal to the memory, read the pixel point data of the P block and Q block of the corresponding 4x4 block side, and calculate the filter strength BS value of the current 4x4 block side, BS can be judged from Table 1; calculate the threshold value of the currently processed 4x4 block side β and threshold tc values.

Table 1 Method of taking BS value BS value condition 2 At least one of P block and Q block is intra prediction 1 At least one of the P block and the Q block has a non-zero residual transform coefficient, and the edge is the edge of the transform unit 1 The absolute value of the motion vector difference between the P block and the Q block is greater than or equal to an integer pixel 1 The inter-frame prediction of P block and Q block refers to different reference frames, or the number of motion vectors between them is different 0 other situations

The values of β and tc can be obtained by equations (1) and (2), where Bitdepth refers to the number of bits per pixel of the image, and it is a constant for an image. β', tc' can be obtained by looking up Table 2 with Q as an input, and Q is an input quantity.

(1)

(2)

Table 2 Lookup table of β′ and tc′ Q 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 β' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 7 8 tc' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Q 19 20 twenty one twenty two twenty three twenty four 25 26 27 28 29 30 31 32 33 34 35 36 37 β' 9 10 11 12 13 14 15 16 17 18 20 twenty two twenty four 26 28 30 32 34 36 tc' 1 1 1 1 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 Q 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 β' 38 40 42 44 46 48 50 52 54 56 58 60 62 64 - - tc' 5 5 6 6 7 8 9 10 11 13 14 16 18 20 twenty two twenty four

(2) Second level

Calculate the filter decision value dE of the currently processed 4x4 block edge, p-point decision value dEp, q-point decision value dEq, filter decision condition value CON, weak filter decision condition value CON_W, brightness threshold , the chromaticity threshold value of C.

Calculate the dE value, if formula (3) is established, then dE is 1, otherwise it is 0, where dp0, dp3, dq0, dq3 can be obtained from formulas (4), (5), (6), (7), where p2, 0, p1,0, p0,0, p2,3, p1,3, p0,3, q2,0, q1,0, q0,0, q2,3, q1,3, q0,3 are the corresponding in Figure 2 The value of the location pixel.

(3)

(4)

(5)

(6)

(7)

Calculate the values of dEp and dEq, if formula (8) is established, then dEp is 1, otherwise it is 0. If formula (9) is established, then dEq is 1, otherwise it is 0.

(8)

(9)

Calculate the value of CON, if formulas (10), (11), and (12) are established at the same time, then the value of CON is 1, otherwise it is 0.

(10)

(11)

(12)

Calculate CON_W and If the value of (13) is established, then the value of CON_W is 1, otherwise it is 0. in The expression of is shown in formula (14).

(13)

(14)

calculate The value of _C can be obtained by formula (15).

(15).

(3) The third level

For the data corresponding to the P block and Q block, filter processing is performed according to the data obtained in the previous stage; the details are as follows:

If the current processing is the 4x4 block edge of the Y component, proceed as follows:

. If the BS value of the current 4x4 block side is 0, or dE is equal to 0, then the pixel values of the P block and Q block will not be processed, and the original value will be kept. Otherwise proceed to steps II and III.

. If the CON value of the current 4x4 block edge is 1, strong filtering is performed on the P block and Q block. We take the first group of 8 pixels shown in Figure 2 as an example for processing. Replace the original value of p0,0,q0,0,p1,0,q1,0,p2,0,q2,0 with the value of p0,q0,p1,q1,p2,q2,p0,q0,p1,q1, The values of p2 and q2 can be obtained from formulas (16), (17), (18), (19), (20), and (21).

(16)

(17)

(18)

(19)

(20)

(twenty one)

The strong filtering method of the 2nd, 3rd, and 4th groups of 8 pixels of P block and Q block is similar to this.

. If the CON value of the current 4x4 block side is 0, and the value of CON_W is 1, weakly filter the P block and Q block, otherwise the pixel values of the P block and Q block will not be processed, and the original value will be kept. We take the first group of 8 pixels shown in Figure 2 as an example for processing. Replace the original values of p0,0, q0,0 with the values of p0, q0, the values of p0, q0 can be obtained by formula (22), (23), the Clip1 operation means that if the value in the expression is less than 0, then the value is equal to 0, if it is greater than 255, it is equal to 255, otherwise it is equal to the original value. ₁ can be obtained from formula (19). The third value in the brackets of the Clip3 operator is less than the first value, it is equal to the first value, if it is greater than the second value, it is equal to the second value, otherwise it is equal to the third value value.

(twenty two)

(twenty three)

(twenty four)

If dEp is 1, replace the original values of p1 and 0 with the value of p1, and the value of p1 can be obtained by formula (26), where the intermediate amount p can be obtained by formula (27).

(26)

(27)

If dEq is 1, replace the original value of q1,0 with the value of q1, and the value of q1 can be obtained by formula (28), where the intermediate quantity q can be obtained by formula (29).

(28)

(29)

The weak filtering method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P and Q blocks is similar to this.

If the current processing is a 4x4 block edge of the Cb or Cr component, proceed as follows.

If the BS value of the current 4x4 block side is equal to 2, perform chroma filtering on the P block and the Q block, otherwise, the pixel values of the P block and the Q block are not processed, and the original values are kept. Here, the processing method of the first group of 8 pixels shown in FIG. 2 is taken as an example for description. Replace the original p0,0, q0,0 values with the values of p0, q0. The values of p0 and q0 can be obtained by equations (30) and (31).

(30)

(31)

The processing method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P block and the Q block is similar to this.

(4) Fourth level

According to the input requirements, the P block and Q block that have been processed by filtering are written back to the corresponding memory.

The present invention adopts an effective four-stage pipeline filtering method applicable to the deblocking filter of the HEVC standard, effectively reduces the processing time, and improves the maximum operating frequency and processing capacity of the chip.

Description of drawings

Figure 1: Diagram of HEVC video encoder structure.

Figure 2: Illustration of a 4x4 pixel block edge that needs to be filtered.

Detailed ways

The method of the present invention will be further specifically described below by way of examples and in conjunction with the accompanying drawings.

The four-stage pipeline structure proposed by the present invention can not only filter the Y component, but also filter the Cb and Cr components. The specific process of each stage in the four-stage pipeline structure is as follows.

(1) Level 1

Send a signal to the memory, read the pixel point data of the P and Q blocks corresponding to the 4x4 block side, and calculate the BS value of the current 4x4 block side at the same time, and the BS value can be judged from Table 1. Calculate the β and tc values of the currently processed 4x4 block side. The values of β and tc can be obtained from formulas (1) and (2), where Bitdepth refers to the number of bits per pixel of the image. For an image, it is a constant. β′, tc′ can be obtained by looking up Table 2 from Q as the input, and Q is the only input quantity.

(1)

(2)

Table 2 Lookup table of β′ and tc′ Q 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 β' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 7 8 tc' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Q 19 20 twenty one twenty two twenty three twenty four 25 26 27 28 29 30 31 32 33 34 35 36 37 β' 9 10 11 12 13 14 15 16 17 18 20 twenty two twenty four 26 28 30 32 34 36 tc' 1 1 1 1 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 Q 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 β' 38 40 42 44 46 48 50 52 54 56 58 60 62 64 - - tc' 5 5 6 6 7 8 9 10 11 13 14 16 18 20 twenty two twenty four

(2) Second level

Calculate the dE value, if formula (3) is established, then dE is 1, otherwise it is 0, where dp0, dp3, dq0, dq3 can be obtained from formulas (4), (5), (6), (7), where p2, 0, p1,0, p0,0, p2,3, p1,3, p0,3, q2,0, q1,0, q0,0, q2,3, q1,3, q0,3 are the corresponding in Figure 2 The value of the location pixel.

(3)

(4)

(5)

(6)

(7)

Calculate the values of dEp and dEq, if formula (8) is established, then dEp is 1, otherwise it is 0. If formula (9) is established, then dEq is 1, otherwise it is 0.

(8)

(9)

Calculate the value of CON, if formulas (10), (11), and (12) are established at the same time, then the value of CON is 1, otherwise it is 0.

(10)

(11)

(12)

Calculate CON_W and If the value of (13) is established, then the value of CON_W is 1, otherwise it is 0. in Can be obtained by formula (14).

(13)

(14)

calculate The value of _C can be obtained by formula (15).

(15).

(3) The third level

If the current processing is the 4x4 block edge of the Y component, proceed as follows.

. If the BS value of the current 4x4 block side is 0, or dE is equal to 0, then the pixel values of the P and Q blocks will not be processed, and the original values will be kept. Otherwise proceed to steps II and III.

. If the CON value of the current 4x4 block edge is 1, strong filtering is performed on P and Q blocks. We take the first group of 8 pixels shown in Figure 2 as an example for processing. Replace the original value of p0,0,q0,0,p1,0,q1,0,p2,0,q2,0 with the value of p0,q0,p1,q1,p2,q2,p0,q0,p1,q1, The values of p2 and q2 can be obtained from formulas (16), (17), (18), (19), (20), and (21).

(16)

(17)

(18)

(19)

(20)

(twenty one)

The strong filtering method of the 2nd, 3rd, and 4th groups of 8 pixels of the P and Q blocks is similar to this.

. If the CON value of the current 4x4 block side is 0, and the value of CON_W is 1, perform weak filtering on the P and Q blocks, otherwise the pixel values of the P and Q blocks will not be processed, and the original values will be kept. We take the first group of 8 pixels shown in Figure 2 as an example for processing. Replace the original values of p0,0, q0,0 with the values of p0, q0, the values of p0, q0 can be obtained by formula (22), (23), the Clip1 operation means that if the value in the expression is less than 0, then the value is equal to 0, if it is greater than 255, it is equal to 255, otherwise it is equal to the original value. ₁ can be obtained from formula (19). The third value in the brackets of the Clip3 operator is less than the first value, it is equal to the first value, if it is greater than the second value, it is equal to the second value, otherwise it is equal to the third value value.

(twenty two)

(twenty three)

(twenty four)

If dEp is 1, replace the original value of p1,0 with the value of p1, the value of p1 can be obtained by formula (26), where p can be obtained by formula (27).

(26)

(27)

If dEq is 1, replace the original value of q1,0 with the value of q1, and the value of q1 can be obtained by formula (28), where q can be obtained by formula (29).

(28)

(29)

The weak filtering method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P and Q blocks is similar to this.

If the current processing is a 4x4 block edge of the Cb or Cr component, proceed as follows.

If the BS value of the current 4x4 block side is equal to 2, perform chroma filtering on the P block and the Q block, otherwise, the pixel values of the P and Q blocks are not processed, and the original values are kept. Here, the processing method of the first group of 8 pixels shown in FIG. 2 is taken as an example for description. Replace the original p0,0, q0,0 values with the values of p0, q0. The values of p0 and q0 can be obtained by equations (30) and (31).

(30)

(31)

The processing method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P and Q blocks is similar to this.

(4) Fourth level

Write back the P and Q blocks that have been filtered and processed to the corresponding memory according to the input requirements.

Claims (1)

1. A four-stage pipeline filtering method applicable to the deblocking filter of the HEVC standard. In the original video stream, each image includes a luminance component, denoted as Y, and each luminance component corresponds to two chrominance components, respectively Denoted as Cb and Cr; in the deblocking filter module, Y, Cb and Cr are all filtered, and the smallest basic unit of each processing is a side of a 4x4 pixel block, and the P block and Q block are the sides For adjacent 4x4 pixel blocks on both sides, if the current edge is a vertical edge, the P block and Q block are respectively the 4x4 pixel block adjacent to the left and right of the edge; if the current edge is a horizontal edge, the P block and Q block are respectively is the adjacent 4x4 pixel block above and below the side; set: The square matrix on the left is the value of 16 corresponding pixel points in the 4x4 P block; the square matrix on the right is the value of the 16 corresponding pixel points in the 4x4 Q block; among the two square matrices, the 8 in the first row The pixels are the first group, the 8 pixels in the second row are the second group, the 8 pixels in the third row are the third group, and the 8 pixels in the fourth row are the fourth group; It is characterized in that the specific steps are as follows: (1) Level 1 Send a signal to the memory, read the pixel point data of the P block and Q block corresponding to the 4x4 block side, and calculate the filter strength BS value of the current 4x4 pixel point block side, BS is obtained by judging from Table 1; Table 1 Method of taking BS value By formula (1), (2) calculates the threshold value β and the threshold value tc value of the 4x4 pixel block edge of current processing; Wherein, Bitdepth refers to the number of bits of each pixel of the image, which is a constant for an image; β', tc' is obtained by Qp as an input look-up table 2, and Qp is an input quantity; β=β'×(1<<(Bitdepth-8)) (1) t _c ＝t _c '×(1<<(Bitdepth-8)) (2) Table 2 Lookup table of β′ and tc′ Qp 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 β' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 7 8 tc' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Qp 19 20 twenty one twenty two twenty three twenty four 25 26 27 28 29 30 31 32 33 34 35 36 37 β' 9 10 11 12 13 14 15 16 17 18 20 twenty two twenty four 26 28 30 32 34 36 tc' 1 1 1 1 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 Qp 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 β' 38 40 42 44 46 48 50 52 54 56 58 60 62 64 - - tc' 5 5 6 6 7 8 9 10 11 13 14 16 18 20 twenty two twenty four (2) Second level Calculate the filter decision value dE, p-point decision value dEp, q-point decision value dEq, filter decision condition value CON, weak filter decision condition value CON_W, brightness threshold Δ, and chroma threshold ΔC of the currently processed 4x4 pixel block edge : ①Calculate the dE value, if formula (3) is established, then dE is 1, otherwise it is 0, where dp0, dp3, dq0, dq3 are obtained from formulas (4), (5), (6), (7): dp0+dp3+dq0+dq3<β (3) dp0＝|p _2,0 -2p _1,0 +p _0,0 | (4) dp3＝|p _2,3 -2p _1,3 +p _0,3 | (5) dq0＝|q _2,0 -2q _1,0 +q _0,0 | (6) dq3＝|q _2,3 -2q _1,3 +q _0,3 | (7) ②Calculate the values of dEp and dEq, if formula (8) is true, then dEp is 1, otherwise it is 0; if formula (9) is true, then dEq is 1, otherwise it is 0; dp0+dp3<(β+(β>>1))>>3 (8) dq0+dq3<(β+(β>>1))>>3 (9) ③Calculate the value of CON, if formulas (10), (11), and (12) are established at the same time, then the value of CON is 1, otherwise it is 0; dp0+dq0<β/8, dp3+dq3<β/8 (10) |p _3,0 -p _0,0 |+|q _0,0 -q _3,0 |<β/8,|p _3,3 -p _0,3 |+|q _0,3 -q _3,3 |＜β/8 (11) |p _0,0 -q _0,0 |＜2.5t _c ,|p _0,3 -q _0,3 |＜2.5t _c (12) ④ Calculate the values of CON_W and Δ, if formula (13) is established, then the value of CON_W is 1, otherwise it is 0; where, Δ is shown in formula (14); |Δ|＜10×t _c (13) Δ=(9×(q _0,0 -p _0,0 )-3×(q _1,0 -p _1,0 )+4)>>4 (14) ⑤ Calculate the value of ΔC, which can be obtained by formula (15); among them, the _Clip3 operation means that if the third value in the brackets is less than the first value, then the value of the expression is equal to the first value, if the third value in the brackets is greater than the second value, the value of the expression is equal to the second value, otherwise the value of the expression is equal to the third value; Δ _C ＝Clip3(-t _c ,t _c ,((((q _0,0 -p _0,0 +1)<<2+p _1,0 +q _1,0 +4)>>3))) (15) (3) The third level For the data corresponding to the P block and Q block, filter processing is performed according to the data obtained in the previous stage; the details are as follows: ①If the current processing is the 4x4 pixel block edge of the Y component, proceed as follows: I. If the BS value of the current 4x4 pixel block edge is 0, or dE is equal to 0, then the pixel point values of the P block and the Q block are not processed, and the original values are kept, otherwise, proceed to steps II and III; II. If the CON value of the current 4x4 pixel block side is 1, strong filtering is performed on the P and Q blocks, and the processing method for the first group of 8 pixels is: use p0, q0, p1, q1, p2, q2 replace the original values of p0,0, q0,0, p1,0, q1,0, p2,0, q2,0, the values of p0, q0, p1, q1, p2, q2 are given by formula (16), ( 17), (18), (19), (20), (21) get: p ₀ =(p _2,0 +2p _1,0 +2p _0,0 +2q _0,0 +q _1,0 +4)>>3 (16) q ₀ =(p _1,0 +2p _0,0 +2q _0,0 +2q _1,0 +q _2,0 +4)>>3 (17) p ₁ =(p _2,0 +p _1,0 +p _0,0 +q _0,0 +2)>>2 (18) q ₁ =(p _0,0 +q _0,0 +q _1,0 +q _2,0 +2)>>2 (19) p ₂ =(2p _3,0 +3p _2,0 +p _1,0 +p _0,0 +q _0,0 +4)>>3 (20) q ₂ =(p _0,0 +q _0,0 +q _1,0 +3q _2,0 +2q _3,0 +4)>>3 (21) The strong filtering method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P block and the Q block is similar to this; III. If the CON value of the edge of the current 4x4 pixel block is 0, and the value of CON_W is 1, perform weak filtering on the P block and the Q block, otherwise the pixel values of the P block and the Q block are not processed; for the first The processing method for a group of 8 pixels is: replace the original values of p0,0, q0,0 with the values of p0, q0. The value in is less than 0, then this value is equal to 0, if it is greater than 255, then it is equal to 255, otherwise it is equal to the original value; the intermediate quantity Δ ₁ is obtained by formula (24): p ₀ =Clip1(p _0,0 +Δ ₁ ) (22) q ₀ ＝Clip1(q _0,0 +Δ ₁ ) (23) Δ ₁ =Clip3(-t _c ,t _c ,Δ) (24) If dEp is 1, replace the original value of p1,0 with the value of p1, the value of p1 is obtained by formula (26), where the intermediate value Δp is obtained by formula (27): p ₁ =Clip1(p _1,0 +Δp) (26) Δ _p ＝Clip3(-(t _c ＞＞1), t _c ＞＞1, (((p _2,0 +p _0,0 +1)＞＞1-p _1,0 +Δ)＞＞1) ) (27) If dEq is 1, replace the original value of q1,0 with the value of q1, and the value of q1 is obtained by formula (28), where the intermediate quantity Δq is obtained by formula (29): q ₁ ＝Clip1(q _1,0 +Δq) (28) Δ _q ＝Clip3(-(t _c ＞＞1), t _c ＞＞1, (((q _2,0 +q _0,0 +1)＞＞1-q _1,0 +Δ)＞＞1) ) (29) The weak filtering method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P block and the Q block is similar to this; ②If the current processing is a 4x4 pixel block edge of Cb or Cr component, proceed as follows: If the BS value of the current 4x4 pixel block edge is equal to 2, perform chroma filtering on the P block and the Q block, otherwise the pixel values of the P block and the Q block will not be processed, and the original value will be kept; for the first group of 8 pixels The processing method of the point is: replace the original value of p0,0, q0,0 with the value of p0, q0; the value of p0, q0 is obtained by formula (30), (31): p ₀ =Clip1(p _0,0 +Δ _C ) (30) q ₀ =Clip1(q _0,0 +Δ _C ) (31) The processing method of the 8 pixels of the 2nd, 3rd, and 4th groups of the P block and the Q block is similar to this; (4) The fourth level According to the input requirements, the P block and Q block that have been processed by filtering are written back to the corresponding memory.