CN113891089B - Method, device and equipment for constructing motion information candidate list - Google Patents

Abstract

The present application provides a method, device and equipment for constructing a motion information candidate list, including: if the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame ; Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion of the time domain adjacent blocks information is not available, and the backward motion information of the neighboring blocks in the time domain is available, then determine the backward motion information in the time domain motion information according to the backward motion information of the neighboring blocks in the time domain; based on the current Time-domain motion information of the block, and constructing a motion information candidate list of the current block. Improve prediction accuracy with this application.

Description

Method, device and equipment for constructing motion information candidate list

technical field

The present application relates to the field of encoding and decoding, and in particular relates to a method, device and equipment for constructing a motion information candidate list.

Background technique

In order to achieve the purpose of saving space, video images are transmitted after being coded, and complete video coding may include processes such as prediction, transformation, quantization, entropy coding, and filtering. For the prediction process, the prediction process can include intra-frame prediction and inter-frame prediction. Inter-frame prediction refers to using the correlation of the video time domain to predict the current pixel using the pixels of the adjacent encoded image, so as to effectively remove video time domain redundancy. Purpose. Intra-frame prediction refers to using the correlation of the video spatial domain to predict the current pixel using the pixels of the encoded block of the current frame image, so as to achieve the purpose of removing video spatial redundancy.

In the inter-frame prediction process, it is usually necessary to construct a motion information candidate list for the current block. The motion information candidate list may include multiple candidate motion information. The candidate motion information in the motion information candidate list may be spatial motion information or Time domain motion information. Obviously, in order to construct the motion information candidate list, it is necessary to obtain the time-domain motion information of the current block. The time-domain motion information obtained in the related art has problems such as poor prediction effect and poor coding performance.

Contents of the invention

The present application provides a method, device and equipment for constructing a motion information candidate list, so as to improve prediction accuracy.

The present application provides a method for constructing a motion information candidate list, the method comprising:

If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame;

Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks If it is not available, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain; based on the current block The time-domain motion information of the current block is constructed to construct a motion information candidate list of the current block.

The present application provides a device for constructing a motion information candidate list, the device comprising:

A determining module, configured to determine the temporal adjacent blocks of the current block from the co-located frame of the current frame if the current frame where the current block is located is a B frame; and determine the current block based on the motion information of the temporal adjacent blocks time domain motion information, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent block is not available, the time domain adjacent block after If the backward motion information is available, the determination module determines the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain; the construction module is used to determine the backward motion information based on the time domain of the current block Motion information, constructing a motion information candidate list of the current block.

The present application provides a decoding device, including: a processor and a machine-readable storage medium, where the machine-readable storage medium stores machine-executable instructions that can be executed by the processor;

The processor is used to execute machine-executable instructions to implement the following steps:

If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame;

Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks If it is not available, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain; based on the current block The time-domain motion information of the current block is constructed to construct a motion information candidate list of the current block.

The present application provides an encoding device, including: a processor and a machine-readable storage medium, where the machine-readable storage medium stores machine-executable instructions that can be executed by the processor;

The processor is used to execute machine-executable instructions to implement the following steps:

If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame;

Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks If it is not available, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain; based on the current block The time-domain motion information of the current block is constructed to construct a motion information candidate list of the current block.

It can be seen from the above technical solutions that in the embodiment of the present application, when the forward motion information of the adjacent blocks in the time domain is not available and the backward motion information of the adjacent blocks in the time domain is available, the backward motion information of the adjacent blocks in the time domain can be used. The motion information determines the backward motion information in the time-domain motion information of the current block, thereby obtaining the time-domain motion information of the current block, improving the accuracy of prediction, improving the prediction performance, improving the coding performance, and making the prediction value of the current block closer to Raw pixels and brings improved encoding performance.

Description of drawings

FIG. 1 is a schematic diagram of a video coding framework;

FIG. 2 is a flowchart of a method for constructing a motion information candidate list in an embodiment of the present application;

3A-3C are schematic diagrams of deriving time-domain motion information in an embodiment of the present application;

4A-4C are schematic diagrams of weight prediction angles in an embodiment of the present application;

FIG. 4D is a schematic diagram of weight prediction positions in an embodiment of the present application;

FIG. 4E is a schematic diagram of reference weight values in an embodiment of the present application;

FIG. 4F is a schematic diagram of adjacent blocks of the current block in an embodiment of the present application;

FIG. 4G is a schematic diagram of an angle partition in an embodiment of the present application;

FIG. 5A is a schematic diagram of a UMVE model prediction process in an embodiment of the present application;

FIG. 5B is a schematic diagram of the AFFINE model prediction process in an embodiment of the present application;

FIG. 6A is a schematic structural diagram of an apparatus for constructing a motion information candidate list in an embodiment of the present application;

FIG. 6B is a hardware structural diagram of a decoder device in an embodiment of the present application;

FIG. 6C is a hardware structural diagram of an encoding end device in an implementation manner of the present application.

detailed description

The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, rather than limiting the present application. The singular forms of "a", "said" and "the" used in the embodiments and claims of this application are also intended to include plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein is meant to include any and all possible combinations of one or more of the associated listed items. It should be understood that although terms such as first, second, and third may be used in the embodiment of the present application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of the embodiments of the present application, first information may also be called second information, and similarly, second information may also be called first information, depending on the context. Additionally, the use of the word "if" may be construed as "at", or "when", or "in response to determination".

The embodiment of this application proposes a method, device and equipment for constructing a motion information candidate list, involving the following concepts:

Video coding framework: see Figure 1, the video coding framework can be used to implement the encoding end processing flow of the embodiment of the application, the schematic diagram of the video decoding framework is similar to Figure 1, and will not be repeated here, and the video decoding framework can be used to implement the application The processing flow of the decoding end of the embodiment. Exemplarily, in the video encoding framework and video decoding framework, it may include but not limited to: intra prediction/inter prediction, motion estimation/motion compensation, reference image buffer, in-loop filtering, reconstruction, transformation, quantization, inversion Transform, inverse quantization, entropy encoder and other modules. At the encoding end, through the cooperation between these modules, the processing flow at the encoding end can be realized, and at the decoding end, through the cooperation between these modules, the processing flow at the decoding end can be realized.

Intra prediction: Considering that there is a strong spatial correlation between adjacent blocks in the image, the surrounding reconstructed pixels can be used as reference pixels to predict the current unencoded block. Therefore, only It is necessary to perform subsequent encoding processing on the residual signal (original signal-predicted signal), instead of encoding the original signal, which effectively removes the redundancy in the spatial domain, achieves the purpose of removing video spatial redundancy, and greatly improves the compression of the video signal efficiency.

Inter prediction (inter prediction): Based on the correlation of the video time domain, since the video sequence contains strong temporal correlation, the pixels of the current image are predicted by using the pixels of the adjacent encoded image, so as to effectively remove the redundancy of the video time domain .

Transform: Convert an image described in the form of pixels in the spatial domain to the transform domain, expressed in terms of transform coefficients. Since most images contain more flat areas and slowly changing areas, transformation can convert the scattered distribution of image energy in the spatial domain into a relatively concentrated distribution in the transformed domain, remove the frequency domain correlation between signals, and cooperate with the quantization process to effectively Compressed stream.

Entropy coding: A lossless coding method based on the principle of information entropy, which converts a series of element symbols used to represent video sequences into binary code streams for transmission or storage. The input symbols may include quantized transformation coefficients and motion vectors Information, prediction mode information, transform and quantization related syntax, etc. The output data is the final code stream after the original video compression. Entropy coding can effectively remove the statistical redundancy of these video element symbols, and is one of the important tools to ensure the compression efficiency of video coding.

Motion Vector (Motion Vector, MV): In inter-frame prediction, the motion vector can be used to represent the relative displacement between the current block of the current frame and the reference block of the reference frame. Each divided block has a corresponding motion vector transmitted to the decoder. If the motion vector of each block is encoded and transmitted independently, especially for a large number of small-sized blocks, a lot of bits will be consumed. In order to reduce the number of bits used to encode the motion vector, the spatial correlation between adjacent blocks can be used to predict the motion vector of the current block according to the motion vector of the adjacent encoded block, and then encode the prediction difference, which can Effectively reduces the number of bits representing motion vectors. When encoding the motion vector of the current block, the motion vector of the adjacent coded block can be used to predict the motion vector of the current block, and then the predicted value of the motion vector (MVP, Motion VectorPrediction) and the real estimate of the motion vector The difference (MVD, Motion Vector Difference) between them is encoded.

Motion Information: Since the motion vector represents the position offset between the current block and a reference block, in order to accurately obtain the information pointing to the block, in addition to the motion vector, the index information of the reference frame image is also required to represent the current block Which reference frame image to use. In video coding technology, for the current frame, a reference frame image list can usually be established, and the reference frame image index information indicates which reference frame image in the reference frame image list is used for the current block.

Exemplarily, many encoding technologies also support multiple reference picture lists, therefore, an index value may also be used to indicate which reference picture list is used, and this index value may be called a reference direction. To sum up, in the video coding technology, information related to motion, such as a motion vector, a reference frame index, and a reference direction, may be collectively referred to as motion information.

Skip mode (skip mode) and Direct mode (direct mode): In the inter-frame prediction process, due to the strong temporal correlation of video, that is, there are many similar blocks in two adjacent frames of images, the current block of the current frame is often Motion search is performed in adjacent reference images, and the block that best matches the current block is found as a reference block. Due to the high similarity between the reference block and the current block, the difference between the two is very small. Therefore, the code rate overhead of encoding the difference value is usually much smaller than the code rate overhead of encoding the pixel value of the current block.

In order to indicate the position of the block that best matches the current block, a lot of motion information needs to be encoded and sent to the decoder, so that the decoder knows the position of the best match block. The motion information, especially the motion vector, needs to consume a lot of code rate for transmission. In order to save this bit rate overhead, special modes that save motion information are designed: Skip mode and Direct mode.

In Skip mode and Direct mode, the motion information of the current block fully reuses the motion information of a neighboring block in the time domain or space domain, that is, selects a motion information from the motion information sets of multiple surrounding blocks as the current block. Sports information.

Therefore, this mode only needs to encode an index value to indicate which motion information in the motion information set the current block uses, and the difference between Skip mode and Direct mode is: Skip mode does not need to encode residuals, Direct mode requires encoding residuals, Direct mode is a direct mode in inter-frame prediction. It needs to transmit residual information. The decoder can derive the motion information of the current block by analyzing the index. After obtaining the motion information, add the predicted value to the residual value to obtain the reconstruction value.

Rate-Distortion Optimized (Rate-Distortion Optimized): There are two indicators for evaluating coding efficiency: code rate and PSNR (Peak Signal to Noise Ratio, peak signal-to-noise ratio). The smaller the bit stream, the greater the compression rate and the greater the PSNR , the better the quality of the reconstructed image is, the discriminant formula is essentially a comprehensive evaluation of the two when selecting the mode. For example, the cost corresponding to the mode: J(mode)=D+λ*R, where D represents Distortion (distortion), which can usually be measured using the SSE index, and SSE refers to the average difference between the reconstructed image block and the source image Square sum; λ is the Lagrangian multiplier, and R is the actual number of bits required for image block coding in this mode, including the sum of bits required for coding mode information, motion information, residuals, etc. In mode selection, if the RDO principle is used to make comparison decisions on encoding modes, the best encoding performance can usually be guaranteed.

In the inter-frame prediction process, it is usually necessary to construct a motion information candidate list for the current block. The motion information candidate list may include multiple candidate motion information. The candidate motion information in the motion information candidate list may be spatial motion information or Time domain motion information. Obviously, in order to construct the motion information candidate list, it is necessary to obtain the time-domain motion information of the current block. The time-domain motion information obtained in the related art has problems such as poor prediction effect and poor coding performance.

In view of the above findings, in this embodiment, when the forward motion information of adjacent blocks in time domain is not available and the backward motion information of adjacent blocks in time domain is available, it can be determined according to the backward motion information of adjacent blocks in time domain Forward motion information in the time domain motion information of the current block, determine the backward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain, so as to obtain the time domain motion information of the current block, Improve prediction accuracy, improve prediction performance, and improve encoding performance.

The method for constructing the motion information candidate list will be described in detail below in conjunction with several specific embodiments.

Embodiment 1: Referring to FIG. 2 , it is a schematic flowchart of a method for constructing a motion information candidate list. This method can be applied to a decoding end (also called a video decoder) or an encoding end (also called a video encoder). Methods can include:

Step 201, if the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frames of the current frame.

Exemplarily, if the current frame where the current block is located is a B frame, based on the preset position of the current block, the time domain corresponding to the preset position can be determined from the co-located frame (also called collocated frame) of the current frame Neighboring blocks (may also be called collocated locations).

Exemplarily, the preset position of the current block can be configured according to experience, and there is no restriction on the preset position of the current block. For example, the preset position of the current block can be: the pixel position of the upper left corner of the current block, or the current The upper right pixel position of the block, or the lower left pixel position of the current block, or the lower right pixel position of the current block, or the center pixel position of the current block, etc.

Exemplarily, the collocated frame (collocated frame) may be the corresponding frame of the reference frame index in the backward reference frame list of the current frame, such as the corresponding frame whose reference frame index is 0 in the backward reference frame list, or, the backward reference frame The corresponding frame whose reference frame index is 1 in the frame list, or the corresponding frame whose reference frame index is 2 in the backward reference frame list, is not limited. The co-located frame can also be the corresponding frame of the reference frame index in the forward reference frame list of the current frame, such as the corresponding frame whose reference frame index is 0 in the forward reference frame list, or the reference frame index in the forward reference frame list The corresponding frame is 1, or the corresponding frame whose reference frame index in the forward reference frame list is 2 is not limited. Of course, the above are just a few examples, which are not limited.

In a possible implementation, the way to determine the co-located frame is that if the current frame is a P frame, then the co-located frame of the current frame is a forward reference frame list (that is, the reference frame list List0, and the forward reference frame list can also be called It is the frame whose reference frame index is 0 in the reference image queue (0). If the current frame is a B frame, the co-located frame of the current frame is a backward reference frame list (i.e. reference frame list List1, and the backward reference frame list can also be referred to as The frame whose reference frame index is 0 in the reference image queue 1).

Exemplarily, regarding the acquisition method of the collocated frame, it may be a preset collocated frame, such as using the first reference frame in the forward reference frame list of the current frame as the collocated frame, or using the The first reference frame in the forward reference frame list is used as the co-located frame; or, it can also be an derived co-located frame, such as taking the reference frame closest to the current frame in the forward reference frame list of the current frame as the co-located frame, or using the current The reference frame closest to the current frame in the frame's backward reference frame list is used as the same frame; or, it can also be a reference frame parsed from the coded bit stream. For example, for the decoder, the decoder can parse it from the code stream output the indication information of the co-located frame, and determine the co-located frame according to the indicated information of the co-located frame.

Exemplarily, the collocated position is a neighboring block in time domain corresponding to the preset position of the current block in the colocated frame.

Step 202, determine time-domain motion information of the current block based on the motion information of temporal-domain adjacent blocks, and the time-domain motion information may include forward motion information and backward motion information. Exemplarily, if the forward motion information of the adjacent blocks in the time domain is not available, but the backward motion information of the adjacent blocks in the time domain is available, then the backward motion information.

In another possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is available but the backward motion information of the neighboring blocks in the time domain is not available, the forward motion information in time domain motion information, and determine backward motion information in the time domain motion information according to forward motion information of adjacent blocks in time domain.

In another possible implementation manner, if the forward motion information of the adjacent blocks in the time domain is available and the backward motion information of the adjacent blocks in the time domain is available, then the time domain is determined according to the forward motion information of the adjacent blocks in the time domain. forward motion information in the time domain motion information, and determine the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain.

In another possible implementation manner, if the forward motion information of the adjacent blocks in the time domain is available and the backward motion information of the adjacent blocks in the time domain is available, then the time domain is determined according to the forward motion information of the adjacent blocks in the time domain. forward motion information in the time domain motion information, and determine backward motion information in the time domain motion information according to the forward motion information of adjacent blocks in the time domain.

In another possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is available but the backward motion information of the neighboring blocks in the time domain is not available, the Forward motion information in temporal motion information.

In another possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is not available, but the backward motion information of the neighboring blocks in the time domain is available, the The forward motion information in the time-domain motion information, and determine the backward motion information in the time-domain motion information according to the backward motion information of the adjacent blocks in the time domain.

In the above embodiments, the motion information of the adjacent blocks in the time domain includes the forward motion information of the adjacent blocks in the time domain and the backward motion information of the adjacent blocks in the time domain. The forward motion information of the neighboring blocks in the temporal domain includes the forward motion vector of the neighboring blocks in the temporal domain and the forward reference frame index of the neighboring blocks in the temporal domain, and the forward reference frame index corresponds to the forward reference frame index of the neighboring blocks in the temporal domain frame. The backward motion information of the temporal adjacent block includes the backward motion vector of the temporal adjacent block and the backward reference frame index of the temporal adjacent block, and the backward reference frame index corresponds to the backward reference of the temporal adjacent block frame. The time-domain motion information of the current block includes forward motion information and backward motion information. The forward motion information in the time-domain motion information includes the forward motion vector of the current block and the forward reference frame index of the current block. The forward reference The frame index corresponds to the forward reference frame of the current block. The forward reference frame of the current block is the reference frame in the forward reference frame list of the current block, such as the reference frame whose reference frame index is 0 in the forward reference frame list of the current block, or the forward reference frame list of the current block The reference frame index in is the reference frame of 1, which is not limited. The backward motion information in the temporal motion information includes the backward motion vector of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the backward reference frame of the current block. The backward reference frame of the current block is the reference frame in the backward reference frame list of the current block, such as the reference frame whose reference frame index is 0 in the backward reference frame list of the current block, or the backward reference frame list of the current block The reference frame index in is the reference frame of 1, which is not limited.

Exemplarily, the forward motion information in the time domain motion information of the current block is determined according to the forward motion information of the time domain adjacent blocks, which may include but not limited to: based on the co-located frame and the forward reference frame of the time domain adjacent blocks and the distance between the current frame and the forward reference frame of the current block, stretch the forward motion vector of the adjacent block in the time domain to obtain the stretched motion vector, and determine the stretched motion vector as A forward motion vector of the current block; determining the forward motion information in the temporal motion information of the current block based on the forward motion vector of the current block and the forward reference frame index of the current block. For example, the forward motion information in the temporal motion information of the current block includes the forward motion vector of the current block and the forward reference frame index of the current block.

Exemplarily, the backward motion information in the time domain motion information of the current block is determined according to the backward motion information of the time domain adjacent blocks, which may include but not limited to: based on the co-located frame and the backward reference frame of the time domain adjacent blocks and the distance between the current frame and the backward reference frame of the current block, stretch the backward motion vector of the adjacent block in the time domain to obtain the stretched motion vector, and determine the stretched motion vector as the backward motion vector of the current block; determining the backward motion information in the temporal motion information of the current block based on the backward motion vector of the current block and the backward reference frame index of the current block. For example, the backward motion information in the temporal motion information of the current block includes the backward motion vector of the current block and the backward reference frame index of the current block.

Exemplarily, determining the forward motion information in the time domain motion information of the current block according to the backward motion information of the time domain adjacent blocks may include but not limited to: based on the co-located frame and the backward reference frame of the time domain adjacent blocks and the distance between the current frame and the forward reference frame of the current block, stretch the backward motion vector of the adjacent block in the time domain to obtain the stretched motion vector, and determine the stretched motion vector as A forward motion vector of the current block; determining the forward motion information in the temporal motion information of the current block based on the forward motion vector of the current block and the forward reference frame index of the current block. For example, the forward motion information in the temporal motion information of the current block includes the forward motion vector of the current block and the forward reference frame index of the current block.

Exemplarily, the backward motion information in the time domain motion information of the current block is determined according to the forward motion information of the time domain adjacent blocks, which may include but not limited to: based on the co-located frame and the forward reference frame of the time domain adjacent blocks and the distance between the current frame and the backward reference frame of the current block, stretch the forward motion vector of the adjacent block in the time domain to obtain the stretched motion vector, and determine the stretched motion vector as the backward motion vector of the current block; determining the backward motion information in the temporal motion information of the current block based on the backward motion vector of the current block and the backward reference frame index of the current block. For example, the backward motion information in the temporal motion information of the current block includes the backward motion vector of the current block and the backward reference frame index of the current block.

Step 203, based on the temporal motion information of the current block, construct a motion information candidate list of the current block.

It can be seen from the above technical solutions that in the embodiment of the present application, when the forward motion information of the adjacent blocks in the time domain is not available and the backward motion information of the adjacent blocks in the time domain is available, the backward motion information of the adjacent blocks in the time domain can be used. The motion information determines the forward motion information in the time domain motion information of the current block, and determines the backward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain, thereby obtaining the time domain of the current block Motion information improves prediction accuracy, prediction performance, and coding performance, which can make the predicted value of the current block closer to the original pixel and bring about an improvement in coding performance.

Embodiment 2: In order to obtain the time-domain motion information of the current block, it can be implemented in the following manner:

If the current frame where the current block is located is a P frame, that is, the current image is a P image, the following steps are used to determine the temporal motion information:

Step a1. Based on the preset position of the current block, determine the temporal adjacent block corresponding to the preset position from the co-located frame of the current frame. If the motion information of the temporal adjacent block is not available, the current block's Time domain motion information is zero motion information.

Exemplarily, the co-located frame may be an image whose reference index is 0 in the forward reference frame list of the current frame, and the forward reference frame list may also be referred to as reference image queue 0, that is, List0. The preset position of the current block may be the upper left luminance sample position. If the reference frame index of the motion information of the adjacent blocks in the time domain is -1, it means that the motion information of the adjacent blocks in the time domain is not available.

To sum up, step a1 can also be expressed as: if the luminance sample corresponding to the luminance sample position in the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 0 is located in the temporal domain The reference frame index stored in the information storage unit is -1 (indicating that the motion information is unavailable), then the L0 motion vector MvE0 of the current block is a zero vector, the L0 reference index value RefIdxL0 of the current block is 0, and the time domain motion information derivation process ends.

Step a2, if the motion information of the neighboring blocks in time domain is available, the L0 reference index RefIdxL0 of the current block is 0, the distance index of the image corresponding to the L0 reference index is recorded as DistanceIndexL0, and the block distance of the image corresponding to the L0 reference index is recorded as BlockDistanceL0 . The motion vector of the adjacent block in the time domain is recorded as mvRef (mvRef_x, mvRef_y), the distance index of the image where the adjacent block in the time domain is located is recorded as DistanceIndexCol, and the distance index of the motion vector pointing to the image is recorded as DistanceIndexRef.

Exemplarily, if the reference frame index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block (i.e. adjacent block in the instant domain) is located in the image whose reference index is 0 in the reference image queue 0 If it is not -1, if the stored reference frame index is greater than or equal to 0, it means that the motion information of the neighboring blocks in the time domain is available.

Exemplarily, the L0 reference index RefIdxL0 of the current block is 0, indicating the first image in the reference image queue 0, that is, the L0 reference index corresponds to the first image in the reference image queue 0. The distance index DistanceIndexL0 of the picture corresponding to the L0 reference index indicates the distance index of the first picture in the reference picture queue 0, such as POC (display order). The block distance BlockDistanceL0 of the image corresponding to the L0 reference index indicates the distance between the image corresponding to the L0 reference index and the current frame, such as the difference between the distance index DistanceIndexL0 of the first image in the reference image queue 0 and the distance index of the current frame.

Exemplarily, the L0 motion vector of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image with the reference index 0 in the reference image queue 0 is recorded as mvRef(mvRef_x, mvRef_y) , the motion vector of the adjacent block in the instant domain. The distance index of the image where the motion information storage unit is located is marked as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the motion vector is marked as DistanceIndexRef.

Step a3, based on the DistanceIndexCol and DistanceIndexRef, the distance BlockDistanceRef between the image where the temporal adjacent block is located and the image pointed to by the motion vector of the temporal adjacent block can be determined in the following manner.

BlockDistanceRef = DistanceIndexCol – DistanceIndexRef

Step a4: Set the L0 reference index RefIdxL0 of the current block to 0, and determine the L0 motion vector mvE0 (mvE0_x, mvE0_y) of the current block using the following formula, and the L0 reference index and the L0 motion vector constitute the time domain motion information of the current block.

mvE0_x＝Clip3(-32768,32767,Sign(mvRef_x*BlockDistanceL0*BlockDistanceRef)*(((Abs(mvRef_x*BlockDistanceL0*(16384/BlockDistanceRef)))+8192)>>14));

mvE0_y=Clip3(-32768,32767,Sign(mvRef_y*BlockDistanceL0*BlockDistanceRef)*(((Abs(mvRef_y*BlockDistanceL0*(16384/BlockDistanceRef)))+8192)>>14)).

In the above formula, Clip3(-32768,32767) is used to limit the value range of mvE0_x and mvE0_y between -32768 and 32767, that is, when the value is less than -32768, let the value of mvE0_x and mvE0_y be -32768 , when the value is greater than 32767, set the values of mvE0_x and mvE0_y to 32767. Sign represents a sign function, Abs represents an absolute value, and >> represents a right shift. mvRef_x represents the horizontal motion vector of the adjacent blocks in the time domain, and mvRef_y represents the vertical motion vector of the adjacent blocks in the time domain. mvE0_x represents the horizontal motion vector of the temporal motion information of the current block, and mvE0_y represents the vertical motion vector of the temporal motion information of the current block. For the definitions of BlockDistanceL0 and BlockDistanceRef, refer to the above steps, and details are not repeated here. Of course, the numerical values in the above formulas can also be adjusted according to actual needs, and there is no limitation on this.

Step a5, set the value of interPredRefMode equal to 'PRED_List0'.

Based on the above steps, if the current frame is a P frame, the temporal motion information of the current block can be obtained.

If the current frame where the current block is located is a B frame, that is, the current image is a B image, the following steps are used to determine the temporal motion information:

Step b1, based on the preset position of the current block, determine the temporal adjacent block corresponding to the preset position from the co-located frame of the current frame, if the forward motion information of the temporal adjacent block is not available, according to the MVP method Get the temporal motion information of the current block.

Exemplarily, the co-located frame may be an image whose reference index is 0 in the backward reference frame list of the current frame, and the backward reference frame list may also be referred to as reference image queue 1, that is, List1. The preset position of the current block may be the upper left luminance sample position. If the reference frame index of the motion information of the adjacent blocks in the time domain is -1, it means that the forward motion information of the adjacent blocks in the time domain is not available.

Exemplarily, obtaining the temporal motion information of the current block according to the MVP method may include but not limited to: taking the size and position of the coding unit where the current block is located as the size and position of the current block, and using the L0 motion vector predictor obtained according to the MVP and L1 motion vector predictors as the L0 motion vector MvE0 and L1 motion vector MvE1 of the current block respectively, and set the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block to be equal to 0, and end the motion information derivation process.

To sum up, step b1 can also be expressed as: if the reference image stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block in the image whose reference index value is 0 in the reference image queue 1 is The frame index is -1 (the storage unit stores the reference frame index of the forward motion information, and the reference frame index is -1 indicating that the forward motion information is not available), then the L0 reference index and L1 reference index of the current block are both equal to 0. The size and position of the coding unit where the current block is located are used as the size and position of the current block, and then the L0 motion vector predictor and the L1 motion vector predictor obtained according to the MVP are respectively used as the L0 motion vector MvE0 and the L1 motion vector MvE1 of the current block, And make the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block both equal to 0, and end the time-domain motion information derivation process of the current block.

Step b2, if the forward motion information of the neighboring block in the time domain is available, both the L0 reference index and the L1 reference index of the current block may be equal to 0. The distance index of the image corresponding to the L0 reference index is recorded as DistanceIndexL0, the distance index of the image corresponding to the L1 reference index is recorded as DistanceIndexL1; the block distance of the image corresponding to the L0 reference index is recorded as BlockDistanceL0, and the block distance of the image corresponding to the L1 reference index is recorded as BlockDistanceL1.

Exemplarily, if the reference frame index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image with a reference index value of 0 in the reference image queue 1 is not -1, such as If the reference frame index is greater than or equal to 0, then the forward motion information of the temporal adjacent block is available.

Exemplarily, the L0 reference index of the current block is 0, indicating the first reference image in the reference image queue 0 of the current block, that is, the L0 reference index corresponds to the first image in the reference image queue 0. The distance index DistanceIndexL0 of the picture corresponding to the L0 reference index indicates the distance index of the first picture in the reference picture queue 0, such as the POC of the first picture. The block distance BlockDistanceL0 of the image corresponding to the L0 reference index indicates the distance between the image corresponding to the L0 reference index and the current frame, such as the difference between the distance index of the first image in the reference image queue 0 and the distance index of the current frame.

Exemplarily, the L1 reference index of the current block is 0, indicating the first image in the reference image queue 1 of the current block, that is, the L1 reference index corresponds to the first image in the reference image queue 1. The distance index DistanceIndexL1 of the image corresponding to the L1 reference index indicates the distance index of the first image in the reference image queue 1, such as the POC of the first reference image. The block distance BlockDistanceL1 of the image corresponding to the L1 reference index indicates the distance between the image corresponding to the L1 reference index and the current frame, such as the difference between the distance index of the first image in the reference image queue 1 and the distance index of the current frame.

Step b3, record the motion vector of the adjacent block in the time domain as mvRef (mvRef_x, mvRef_y), the distance index of the image where the adjacent block in the time domain is located is recorded as DistanceIndexCol, and the distance of the reference image pointed to by the motion vector of the adjacent block in the time domain The index is denoted as DistanceIndexRef. For example, the L0 motion vector of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef(mvRef_x, mvRef_y), Motion vectors of neighboring blocks in the instant domain. The distance index of the image where the motion information storage unit is located is marked as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the motion vector is marked as DistanceIndexRef. Based on the DistanceIndexCol and DistanceIndexRef, the distance BlockDistanceRef between the image where the temporal adjacent block is located and the reference image pointed to by the motion vector of the temporal adjacent block can be determined in the following manner.

BlockDistanceRef = DistanceIndexCol – DistanceIndexRef

Step b4, set the L0 reference index RefIdxL0 of the current block to be 0, use the following formula to determine the L0 motion vector mvE0 (mvE0_x, mvE0_y) of the current block, and the L0 reference index and the L0 motion vector constitute the time domain motion information of the current block Forward motion information. Let the L1 reference index RefIdxL1 of the current block be 0, use the following formula to determine the L1 motion vector mvE1(mvE1_x, mvE1_y) of the current block, and the L1 reference index and L1 motion vector constitute the backward motion in the time domain motion information of the current block information, and the forward motion information and the backward motion information constitute the time-domain motion information of the current block.

mvE0_x＝Clip3(-32768,32767,Sign(mvRef_x*BlockDistanceL0*BlockDistanceRef)*(((Abs(mvRef_x*BlockDistanceL0*(16384/BlockDistanceRef)))+8192)>>14))

mvE0_y＝Clip3(-32768,32767,Sign(mvRef_y*BlockDistanceL0*BlockDistanceRef)*(((Abs(mvRef_y*BlockDistanceL0*(16384/BlockDistanceRef)))+8192)>>14))

mvE1_x＝Clip3(-32768,32767,Sign(mvRef_x*BlockDistanceL1*BlockDistanceRef)*(((Abs(mvRef_x*BlockDistanceL1*(16384/BlockDistanceRef)))+8192)>>14))

mvE1_y＝Clip3(-32768,32767,Sign(mvRef_y*BlockDistanceL1*BlockDistanceRef)*(((Abs(mvRef_y*BlockDistanceL1*(16384/BlockDistanceRef)))+8192)>>14))

In the above formula, the meaning of each symbol can be referred to step a4, and will not be repeated here.

Based on the above steps, if the current frame is a B frame, the temporal motion information of the current block can be obtained.

In Embodiment 2, if the current frame is a B frame, it is L0 of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index is 0 in the reference image queue 1 The motion information is used as the reference motion information, without considering the L1 motion information of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is located as the reference motion information.

Embodiment 3: In order to obtain the temporal motion information of the current block, if the current frame where the current block is located is a P frame, that is, the current image is a P image, then the determination method of the temporal motion information refers to Embodiment 2, which will not be repeated here. If the current frame where the current block is located is a B frame, that is, the current image is a B image, the following steps are used to determine the temporal motion information of the current block. Exemplarily, a P frame is a picture that can be decoded using a knowledge picture and past pictures in display order as reference pictures in inter-frame prediction. A B frame is a knowledge image that can be used in inter prediction, and displays sequentially past and future images as reference images to be decoded.

Step c1. Based on the preset position of the current block, determine the temporal adjacent block corresponding to the preset position from the co-located frame of the current frame. The motion information of the temporal adjacent block includes forward motion information and backward motion information. Sports information. If the forward motion information of the neighboring blocks in the time domain is not available, and the backward motion information of the neighboring blocks in the time domain is not available, the time domain motion information of the current block is obtained according to the MVP method.

Exemplarily, the co-located frame (collocated frame) of the current frame may be the corresponding image (image may also be called a frame) of the reference frame index in the backward reference frame list of the current frame, such as the reference frame index in the backward reference frame list The corresponding image of 0 may also be the corresponding image of the reference frame index in the forward reference frame list of the current frame, such as the corresponding image of the reference frame index of 0 in the forward reference frame list, and there is no restriction on this co-located frame. For convenience of description, take the image whose reference index is 0 in the backward reference frame list of the current frame as an example. The backward reference frame list may also be called reference image queue 1, namely List1.

Exemplarily, the preset position of the current block can be configured based on experience, for example, the upper corner luminance sample position, the upper right luminance sample position, the lower left luminance sample position, the lower right luminance sample position, the center luminance sample position, etc. The default position is not limited. For the convenience of description, take the upper-left luminance sample position of the current block as an example.

Exemplarily, if the reference frame index of the forward motion information of the adjacent blocks in the time domain is -1, it means that the forward motion information of the adjacent blocks in the time domain is not available. If the reference frame index of the forward motion information of the neighboring blocks in the time domain is not -1, if the reference frame index is greater than or equal to 0, it means that the forward motion information of the neighboring blocks in the time domain is available. If the reference frame index of the backward motion information of the adjacent blocks in the time domain is -1, it means that the backward motion information of the adjacent blocks in the time domain is not available. If the reference frame index of the backward motion information of the adjacent blocks in the time domain is not -1, if the reference frame index is greater than or equal to 0, it means that the backward motion information of the adjacent blocks in the time domain is available.

Exemplarily, obtaining the time-domain motion information of the current block according to the MVP method may include but not limited to the following methods: using the size and position of the coding unit where the current block is located as the size and position of the current block, the L0 motion vector obtained according to the MVP The predicted value and the L1 motion vector predictor are respectively used as the L0 motion vector MvE0 and the L1 motion vector MvE1 of the current block, and the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block are both equal to 0.

The L0 motion vector MvE0 and the L0 reference index RefIdxL0 form the forward motion information in the time domain motion information of the current block, the L1 motion vector MvE1 and the L1 reference index RefIdxL1 form the backward motion information in the time domain motion information of the current block, The forward motion information and the backward motion information constitute time-domain motion information of the current block.

The L0 motion vector MvE0 indicates a forward motion vector, the L0 reference index RefIdxL0 indicates a forward reference index, the L1 motion vector MvE1 indicates a backward motion vector, and the L1 reference index RefIdxL1 indicates a backward reference index.

Step c2, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion information of the current block can be determined according to the forward motion information of the neighboring block in the time domain and determine the backward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain.

Exemplarily, if the reference frame index of the forward motion information of the adjacent blocks in the time domain is not -1, if the reference frame index is greater than or equal to 0, then the forward motion information of the adjacent blocks in the time domain is available. If the reference frame index of the backward motion information of the adjacent blocks in the time domain is not -1, if the reference frame index is greater than or equal to 0, it means that the backward motion information of the adjacent blocks in the time domain is available.

The forward motion information of the neighboring block in the time domain may include the forward motion vector mvRef0(mvRef0_x, mvRef0_y) of the neighboring block in the time domain and the forward reference frame index of the neighboring block in the time domain, and the forward reference frame index corresponds to the time domain The forward reference frame of the adjacent block, the backward motion information of the adjacent block in the time domain includes the backward motion vector mvRef1(mvRef1_x, mvRef1_y) of the adjacent block in the time domain and the backward reference frame index of the adjacent block in the time domain, the The backward reference frame index corresponds to the backward reference frame of the neighboring block in time domain.

The forward motion information in the temporal motion information of the current block includes the forward motion vector mvE0(mvE0_x, mvE0_y) of the current block and the forward reference frame index of the current block, which corresponds to the forward reference frame index of the current block frame, the forward reference frame of the current block is the reference frame in the forward reference frame list of the current block, taking the reference frame whose reference frame index is 0 in the forward reference frame list of the current block as an example, that is, the previous frame of the current block The reference frame index of the forward reference frame (also referred to as the L0 reference index) is 0, indicating the first picture in the forward reference frame list (ie reference picture queue 0) of the current block. The backward motion information in the temporal motion information of the current block includes the backward motion vector mvE1 (mvE1_x, mvE1_y) of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the backward reference of the current block frame, the backward reference frame of the current block is the reference frame in the backward reference frame list of the current block, taking the reference frame whose reference frame index is 0 in the backward reference frame list of the current block as an example, that is, the backward reference frame of the current block The reference frame index (also referred to as L1 reference index) of the forward reference frame is 0, indicating the first picture in the backward reference frame list (ie reference picture queue 1) of the current block.

Exemplarily, the distance index (such as POC) of the forward reference frame of the current block can be recorded as DistanceIndexL0, and the distance index of the backward reference frame of the current block can be recorded as DistanceIndexL1, and the forward reference frame of the current frame and the current block can be recorded as DistanceIndexL1. The distance between frames is recorded as BlockDistanceL0, BlockDistanceL0 represents the difference between the distance index of the current frame and the distance index of the forward reference frame of the current block, and the distance between the current frame and the backward reference frame of the current block is recorded as BlockDistanceL1, BlockDistanceL1 represents the difference between the distance index of the current frame and the distance index of the backward reference frame of the current block.

Exemplarily, the distance index of the image where the time-domain adjacent block is located (ie, the co-located frame of the current frame) can be recorded as DistanceIndexCol, and the image pointed to by the forward motion vector mvRef0 of the time-domain adjacent block (ie, the time-domain adjacent block The distance index of the forward reference frame) is recorded as DistanceIndexRef0, and the distance index of the image pointed to by the backward motion vector mvRef1 of the adjacent block in the temporal domain (the backward reference frame of the adjacent block in the temporal domain) is recorded as DistanceIndexRef1. Based on this, the distance between the co-located frame and the forward reference frame of the time-domain adjacent block can be recorded as BlockDistanceRef0, and the distance between the co-located frame and the backward reference frame of the time-domain adjacent block can be recorded as for BlockDistanceRef1.

BlockDistanceRef0 represents the difference between the distance index of the co-located frame and the distance index of the forward reference frame of the temporal adjacent block, for example, it can be expressed by the following formula: BlockDistanceRef0=DistanceIndexCol−DistanceIndexRef0.

BlockDistanceRef1 represents the difference between the distance index of the co-located frame and the distance index of the backward reference frame of the temporal adjacent block, for example, it can be expressed by the following formula: BlockDistanceRef1=DistanceIndexCol−DistanceIndexRef1.

Based on the above definition, the forward motion information in the time domain motion information of the current block is determined according to the forward motion information of the time domain adjacent blocks, which may include but not limited to: based on the co-located frame and the forward reference frame of the time domain adjacent block The distance BlockDistanceRef0 between, and the distance BlockDistanceL0 between the current frame and the forward reference frame of the current block, stretch the forward motion vector mvRef0 (mvRef0_x, mvRef0_y) of the adjacent block in the time domain to obtain the stretched motion vector, Use the stretched motion vector as the forward motion vector mvE0(mvE0_x, mvE0_y) of the current block; determine the forward motion vector in the temporal motion information of the current block based on the forward motion vector of the current block and the forward reference frame index of the current block The motion information, for example, the forward motion information in the temporal motion information includes mvE0 (mvE0_x, mvE0_y) and the forward reference frame index of the current block.

Exemplarily, the following formula (1) and formula (2) can be used to stretch the forward motion vector mvRef0 (mvRef0_x, mvRef0_y) of the adjacent block in the time domain to obtain the stretched motion vector mvE0 (mvE0_x, mvE0_y).

mvE0_x＝Clip3(-32768,32767,Sign(mvRef0_x*BlockDistanceL0*BlockDistanceRef0)*(((Abs(mvRef0_x*BlockDistanceL0*(16384/BlockDistanceRef0)))+8192)>>14))(1)

mvE0_y＝Clip3(-32768,32767,Sign(mvRef0_y*BlockDistanceL0*BlockDistanceRef0)*(((Abs(mvRef0_y*BlockDistanceL0*(16384/BlockDistanceRef0)))+8192)>>14))(2)

In formula (1) and formula (2), Clip3(-32768,32767) is used to limit the value range of mvE0_x and mvE0_y between -32768 and 32767, that is, when the value is less than -32768, let mvE0_x and The value of mvE0_y is -32768, and when the value is greater than 32767, set the values of mvE0_x and mvE0_y to 32767. Sign represents a sign function, Abs represents an absolute value, and >> represents a right shift. mvRef0_x represents the horizontal motion vector of the forward motion vector of the neighboring block in the temporal domain, and mvRef0_y represents the vertical motion vector of the forward motion vector of the neighboring block in the temporal domain. mvE0_x indicates the horizontal motion vector of the forward motion vector of the current block, and mvE0_y indicates the vertical motion vector of the forward motion vector of the current block. BlockDistanceL0 represents the distance between the current frame and the forward reference frame of the current block, and BlockDistanceRef0 represents the distance between the co-located frame and the forward reference frame of the temporal adjacent block. The numerical values in formula (1) and formula (2) can also be adjusted according to actual needs, and there is no limit to this. Of course, formula (1) and formula (2) are just examples, as long as it can be based on the distance between the co-located frame and the forward reference frame of the temporal adjacent block, and the distance between the current frame and the forward reference frame of the current block , stretching the forward motion vector mvRef0 of the adjacent block in the time domain to obtain the stretched motion vector mvE0. For example, based on the distance between the co-located frame and the forward reference frame of the temporal adjacent block, and the distance between the current frame and the forward reference frame of the current block, the forward motion vector of the temporal adjacent block The horizontal motion vector of the current block is stretched to obtain the horizontal motion vector in the forward motion vector of the current block; and, based on the distance between the co-located frame and the forward reference frame of the temporal adjacent block, and the previous To the distance between the reference frames, the vertical motion vector in the forward motion vector of the adjacent block in the time domain is stretched to obtain the vertical motion vector in the forward motion vector of the current block.

Based on the above definition, the backward motion information in the time domain motion information of the current block is determined according to the backward motion information of the temporal adjacent blocks, including but not limited to: The distance BlockDistanceRef1 between, and the distance BlockDistanceL1 between the current frame and the backward reference frame of the current block, stretch the backward motion vector mvRef1 (mvRef1_x, mvRef1_y) of the adjacent block in the time domain to obtain the stretched motion vector, and Use the stretched motion vector as the backward motion vector mvE1(mvE1_x, mvE1_y) of the current block; determine the backward direction in the temporal motion information of the current block based on the backward motion vector of the current block and the backward reference frame index of the current block The motion information, for example, the backward motion information in the temporal motion information of the current block includes mvE1 (mvE1_x, mvE1_y) and the backward reference frame index of the current block.

Exemplarily, the following formula (3) and formula (4) can be used to stretch the backward motion vector mvRef1(mvRef1_x, mvRef1_y) of the adjacent block in the time domain to obtain the stretched motion vector mvE1(mvE1_x, mvE1_y).

mvE1_x＝Clip3(-32768,32767,Sign(mvRef1_x*BlockDistanceL1*BlockDistanceRef1)*(((Abs(mvRef1_x*BlockDistanceL1*(16384/BlockDistanceRef1)))+8192)>>14))(3)

mvE1_y＝Clip3(-32768,32767,Sign(mvRef1_y*BlockDistanceL1*BlockDistanceRef1)*(((Abs(mvRef1_y*BlockDistanceL1*(16384/BlockDistanceRef1)))+8192)>>14))(4)

In formula (3) and formula (4), Clip3(-32768,32767) is used to limit the value range of mvE1_x and mvE1_y between -32768 and 32767, that is, when the value is less than -32768, let mvE1_x and The value of mvE1_y is -32768, and when the value is greater than 32767, set the values of mvE1_x and mvE1_y to 32767. Sign represents a sign function, Abs represents an absolute value, and >> represents a right shift. mvRef1_x represents the horizontal motion vector of the backward motion vector of the neighboring block in the time domain, and mvRef1_y represents the vertical motion vector of the backward motion vector of the neighboring block in the time domain. mvE1_x indicates the horizontal motion vector of the backward motion vector of the current block, and mvE1_y indicates the vertical motion vector of the backward motion vector of the current block. BlockDistanceL1 represents the distance between the current frame and the backward reference frame of the current block, and BlockDistanceRef1 represents the distance between the co-located frame and the backward reference frame of the temporal adjacent block. The numerical values in formula (3) and formula (4) can also be adjusted according to actual needs, and there is no limit to this. Of course, formula (3) and formula (4) are just examples, as long as it can be based on the distance between the co-located frame and the backward reference frame of the temporal adjacent block, and the distance between the current frame and the backward reference frame of the current block , stretching the first backward motion vector mvRef1 to obtain the stretched backward motion vector mvE1 of the current block. For example, based on the distance between the co-located frame and the backward reference frame of the temporal adjacent block, and the distance between the current frame and the backward reference frame of the current block, the backward motion vector of the temporal adjacent block The horizontal motion vector of the current block is stretched to obtain the horizontal motion vector in the backward motion vector of the current block; and, based on the distance between the co-located frame and the backward reference frame of the temporal adjacent block, and the backward The distance between the reference frames is used to scale the vertical motion vector among the backward motion vectors of adjacent blocks in time domain to obtain the vertical motion vector among the backward motion vectors of the current block.

To sum up, the forward motion information in the temporal motion information of the current block and the backward motion information in the temporal motion information of the current block can be obtained, and the forward motion information and the backward motion information are combined into a current Temporal motion information of a block.

To sum up, as shown in Figure 3A, if the forward motion information of adjacent blocks in time domain is available, and the backward motion information of adjacent blocks in time domain is available, then the distance index DistanceIndexCol of the co-located frame can be compared with the time domain relative The distance index DistanceIndexRef0 of the forward reference frame of the adjacent block determines the distance BlockDistanceRef0 between the co-located frame and the forward reference frame of the adjacent block in the time domain, which can be based on the distance index DistanceIndexCur of the current frame and the forward reference frame of the current block. The distance index DistanceIndexL0 determines the distance BlockDistanceL0 between the current frame and the forward reference frame of the current block, and then stretches the forward motion vector mvRef0 of the temporal adjacent block through BlockDistanceRef0 and BlockDistanceL0 to obtain the forward motion of the current block Vector mvE0. Based on the distance index DistanceIndexCol of the co-located frame and the distance index DistanceIndexRef1 of the backward reference frame of the temporal adjacent block, the distance BlockDistanceRef1 between the co-located frame and the backward reference frame of the temporal adjacent block can be determined, which can be based on the current frame The distance index DistanceIndexCur and the distance index DistanceIndexL1 of the backward reference frame of the current block determine the distance BlockDistanceL1 between the current frame and the backward reference frame of the current block, and then, through BlockDistanceRef1 and BlockDistanceL1, the backward movement of adjacent blocks in the time domain The vector mvRef1 is stretched to obtain the backward motion vector mvE1 of the current block.

Then, the forward motion vector mvE0 of the current block, the backward motion vector mvE1 of the current block, the forward reference frame index of the current block and the backward reference frame index of the current block can form the temporal motion information of the current block.

Step c3, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is not available, then the time domain motion of the current block can be determined according to the forward motion information of the neighboring block in the time domain The forward motion information in the information, and determine the backward motion information in the time domain motion information of the current block according to the forward motion information of the adjacent blocks in the time domain.

If the reference frame index of the forward motion information of the adjacent blocks in the time domain is greater than or equal to 0, the forward motion information of the adjacent blocks in the time domain is available. If the reference frame index of the backward motion information of the adjacent blocks in the temporal domain is -1, the backward motion information of the adjacent blocks in the temporal domain is unavailable.

The forward motion information of the neighboring blocks in the time domain includes the forward motion vector mvRef0(mvRef0_x, mvRef0_y) of the neighboring blocks in the time domain and the forward reference frame index of the neighboring blocks in the time domain, and the forward reference frame index corresponds to the time domain phase The forward reference frame of the adjacent block, the backward motion information of the adjacent block in the temporal domain includes the backward motion vector mvRef1(mvRef1_x, mvRef1_y) of the adjacent block in the temporal domain and the backward reference frame index of the adjacent block in the temporal domain, the backward The backward reference frame index corresponds to the backward reference frame of the neighboring block in time domain. The forward motion information in the temporal motion information of the current block includes the forward motion vector mvE0(mvE0_x, mvE0_y) of the current block and the forward reference frame index of the current block, which corresponds to the forward reference frame index of the current block frame, the backward motion information in the temporal motion information of the current block includes the backward motion vector mvE1 (mvE1_x, mvE1_y) of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the backward direction of the current block to the reference frame. The distance index of the forward reference frame of the current block is recorded as DistanceIndexL0, the distance index of the backward reference frame of the current block is recorded as DistanceIndexL1, the distance between the current frame and the forward reference frame of the current block is recorded as BlockDistanceL0, and The distance between the current frame and the backward reference frame of the current block is denoted as BlockDistanceL1. The distance index of the co-located frame is recorded as DistanceIndexCol, the distance index of the forward reference frame of the temporal adjacent block is recorded as DistanceIndexRef0, and the distance index of the backward reference frame of the temporal adjacent block is recorded as DistanceIndexRef1. The distance between the co-located frame and the forward reference frame of the adjacent block in the time domain is recorded as BlockDistanceRef0, and the distance between the co-located frame and the backward reference frame of the adjacent block in the time domain is recorded as BlockDistanceRef1.

Based on the above definition, the forward motion information in the time domain motion information of the current block can be determined according to the forward motion information of the adjacent blocks in the time domain. For this process, refer to the process of step c2, which will not be repeated here.

Based on the above definition, the backward motion information in the time domain motion information of the current block is determined according to the forward motion information of the time domain adjacent blocks, which may include but not limited to: based on the co-located frame and the forward reference frame of the time domain adjacent block The distance BlockDistanceRef0 between, and the distance BlockDistanceL1 between the current frame and the backward reference frame of the current block, stretch the forward motion vector mvRef0 (mvRef0_x, mvRef0_y) of the adjacent block in the time domain to obtain the stretched motion vector, Use the stretched motion vector as the backward motion vector mvE1(mvE1_x, mvE1_y) of the current block; determine the backward direction in the temporal motion information of the current block based on the backward motion vector of the current block and the backward reference frame index of the current block Motion information, for example, backward motion information in time-domain motion information includes mvE1 (mvE1_x, mvE1_y) and the backward reference frame index of the current block.

Exemplarily, the following formula (5) and formula (6) can be used to stretch the forward motion vector mvRef0 (mvRef0_x, mvRef0_y) of the adjacent block in the time domain to obtain the stretched motion vector mvE1 (mvE1_x, mvE1_y).

mvE1_x＝Clip3(-32768,32767,Sign(mvRef0_x*BlockDistanceL1*BlockDistanceRef0)*(((Abs(mvRef0_x*BlockDistanceL1*(16384/BlockDistanceRef0)))+8192)>>14))(5)

mvE1_y＝Clip3(-32768,32767,Sign(mvRef0_y*BlockDistanceL1*BlockDistanceRef0)*(((Abs(mvRef0_y*BlockDistanceL1*(16384/BlockDistanceRef0)))+8192)>>14))(6)

In formula (5) and formula (6), Clip3(-32768, 32767) is used to limit the value range of mvE1_x and mvE1_y between -32768 and 32767. Sign represents a sign function, Abs represents an absolute value, and >> represents a right shift. mvRef0_x represents the horizontal motion vector of the forward motion vector of the neighboring block in the temporal domain, and mvRef0_y represents the vertical motion vector of the forward motion vector of the neighboring block in the temporal domain. mvE1_x indicates the horizontal motion vector of the backward motion vector of the current block, and mvE1_y indicates the vertical motion vector of the backward motion vector of the current block. BlockDistanceL1 represents the distance between the current frame and the backward reference frame of the current block, and BlockDistanceRef0 represents the distance between the co-located frame and the forward reference frame of the temporal adjacent block.

To sum up, the forward motion information in the temporal motion information of the current block and the backward motion information in the temporal motion information of the current block can be obtained, and the forward motion information and the backward motion information are combined into a current Temporal motion information of a block.

To sum up, as shown in Figure 3B, if the forward motion information of adjacent blocks in the time domain is available, and the backward motion information of adjacent blocks in the time domain is not available, the distance index DistanceIndexCol based on the co-located frame and the time domain phase The distance index DistanceIndexRef0 of the forward reference frame of the adjacent block determines the distance BlockDistanceRef0 between the co-located frame and the forward reference frame of the adjacent block in the time domain, which can be based on the distance index DistanceIndexCur of the current frame and the forward reference frame of the current block. The distance index DistanceIndexL0 determines the distance BlockDistanceL0 between the current frame and the forward reference frame of the current block, and then stretches the forward motion vector mvRef0 of the temporal adjacent block through BlockDistanceRef0 and BlockDistanceL0 to obtain the forward motion of the current block Vector mvE0. The distance BlockDistanceL1 between the current frame and the backward reference frame of the current block can be determined based on the distance index DistanceIndexCur of the current frame and the distance index DistanceIndexL1 of the backward reference frame of the current block. The forward motion vector mvRef0 is scaled to obtain the backward motion vector mvE1 of the current block.

The forward motion vector mvE0 of the current block, the backward motion vector mvE1 of the current block, the forward reference frame index of the current block and the backward reference frame index of the current block can form the temporal motion information of the current block.

Step c4, if the forward motion information of the neighboring block in the time domain is not available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion of the current block can be determined according to the backward motion information of the neighboring block in the time domain The forward motion information in the information, and determine the backward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain.

If the reference frame index of the forward motion information of the neighboring blocks in the temporal domain is -1, the forward motion information of the neighboring blocks in the temporal domain is unavailable. If the reference frame index of the backward motion information of the adjacent blocks in the time domain is greater than or equal to 0, the backward motion information of the adjacent blocks in the time domain is available.

Based on the definition of the above embodiment, the backward motion information in the time domain motion information of the current block can be determined according to the backward motion information of the adjacent blocks in the time domain. For this process, refer to the processing process of step c2, which will not be repeated here. .

Based on the definition of the above-mentioned embodiments, determining the forward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain may include but not limited to: The distance BlockDistanceRef1 between the reference frames, and the distance BlockDistanceL0 between the current frame and the forward reference frame of the current block, stretch the backward motion vector mvRef1 (mvRef1_x, mvRef1_y) of the adjacent block in the time domain, and obtain the stretched Motion vector, and use the stretched motion vector as the forward motion vector mvE0(mvE0_x, mvE0_y) of the current block; determine the temporal motion information of the current block based on the forward motion vector of the current block and the forward reference frame index of the current block For example, the forward motion information in the temporal motion information includes mvE0 (mvE0_x, mvE0_y) and the forward reference frame index of the current block.

Exemplarily, the following formula (7) and formula (8) can be used to stretch the backward motion vector mvRef1 (mvRef1_x, mvRef1_y) of the adjacent block in the time domain to obtain the stretched motion vector mvE0 (mvE0_x, mvE0_y).

mvE0_x＝Clip3(-32768,32767,Sign(mvRef1_x*BlockDistanceL0*BlockDistanceRef1)*(((Abs(mvRef1_x*BlockDistanceL0*(16384/BlockDistanceRef1)))+8192)>>14))(7)

mvE0_y＝Clip3(-32768,32767,Sign(mvRef1_y*BlockDistanceL0*BlockDistanceRef1)*(((Abs(mvRef1_y*BlockDistanceL0*(16384/BlockDistanceRef1)))+8192)>>14))(8)

In formula (7) and formula (8), Clip3(-32768, 32767) is used to limit the value range of mvE0_x and mvE0_y between -32768 and 32767. Sign represents a sign function, Abs represents an absolute value, and >> represents a right shift. mvRef1_x represents the horizontal motion vector of the backward motion vector of the neighboring block in the time domain, and mvRef1_y represents the vertical motion vector of the backward motion vector of the neighboring block in the time domain. mvE0_x indicates the horizontal motion vector of the forward motion vector of the current block, and mvE0_y indicates the vertical motion vector of the forward motion vector of the current block. BlockDistanceL0 represents the distance between the current frame and the forward reference frame of the current block, and BlockDistanceRef1 represents the distance between the co-located frame and the backward reference frame of the temporal adjacent block.

To sum up, the forward motion information in the temporal motion information of the current block and the backward motion information in the temporal motion information of the current block can be obtained, and the forward motion information and the backward motion information are combined into a current Temporal motion information of a block.

To sum up, as shown in Figure 3C, if the forward motion information of adjacent blocks in the time domain is not available, and the backward motion information of adjacent blocks in the time domain is available, the distance index DistanceIndexCol based on the co-located frame and the time domain phase The distance index DistanceIndexRef1 of the backward reference frame of the adjacent block determines the distance BlockDistanceRef1 between the co-located frame and the backward reference frame of the adjacent block in the time domain, which can be based on the distance index DistanceIndexCur of the current frame and the forward reference frame of the current block. The distance index DistanceIndexL0 determines the distance BlockDistanceL0 between the current frame and the forward reference frame of the current block, and then stretches the backward motion vector mvRef1 of the temporal adjacent block through BlockDistanceRef1 and BlockDistanceL0 to obtain the forward motion of the current block Vector mvE0. The distance BlockDistanceL1 between the current frame and the backward reference frame of the current block can be determined based on the distance index DistanceIndexCur of the current frame and the distance index DistanceIndexL1 of the backward reference frame of the current block. The backward motion vector mvRef1 is scaled to obtain the backward motion vector mvE1 of the current block.

The forward motion vector mvE0 of the current block, the backward motion vector mvE1 of the current block, the forward reference frame index of the current block and the backward reference frame index of the current block can form the temporal motion information of the current block.

The above process of Embodiment 3 is described below in combination with a specific application scenario:

If the L0 reference index and the L1 reference frame index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 are both -1, Then the L0 reference index and the L1 reference index of the current block are both equal to 0. The size and position of the coding unit where the current block is located are used as the size and position of the current block, and then the L0 motion vector predictor and the L1 motion vector predictor obtained according to the MVP are respectively used as the L0 motion vector MvE0 and the L1 motion vector MvE1 of the current block, And make the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block both equal to 0, and end the derivation process of the time-domain motion information of the current block.

If the reference index value of the reference image queue 1 is 0, the L0 reference index and the L1 reference frame index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left luminance sample position of the current block is not both -1 , then the L0 reference index and the L1 reference index of the current block are both equal to 0. The distance index of the image corresponding to the L0 reference index of the current block is recorded as DistanceIndexL0, the distance index of the image corresponding to the L1 reference index of the current block is recorded as DistanceIndexL1; the BlockDistance of the image corresponding to the L0 reference index of the current block is recorded as BlockDistanceL0, that is, the current block The distance between the image corresponding to the L0 reference index of the current frame and the current frame; the BlockDistance of the image corresponding to the L1 reference index of the current block is recorded as BlockDistanceL1, that is, the distance between the image corresponding to the L1 reference index of the current block and the current frame.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block Forward motion information is available), then the L0 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef0( mvRef0_x, mvRef0_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L0 motion vector is recorded as DistanceIndexRef0. Exemplarily, BlockDistanceRef0 may be determined by the following formula: BlockDistanceRef0=DistanceIndexCol−DistanceIndexRef0.

If the L1 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block backward motion information is available), then the L1 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef1( mvRef1_x, mvRef1_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L1 motion vector is recorded as DistanceIndexRef1. Exemplarily, BlockDistanceRef1 may be determined by the following formula: BlockDistanceRef1=DistanceIndexCol−DistanceIndexRef1.

Then, the L0 reference index RefIdxL0 of the current block is equal to 0, the L0 motion vector of the current block is mvE0 (mvE0_x, mvE0_y), the L1 reference index RefIdxL1 of the current block is equal to 0, and the L1 motion vector of the current block is mvE1 (mvE1_x, mvE1_y). The L0 reference index RefIdxL0 and the L0 motion vector mvE0 (mvE0_x, mvE0_y) form the forward motion information, and the L1 reference index RefIdxL1 and the L1 motion vector mvE1 (mvE1_x, mvE1_y) form the backward motion information.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index value is 0 in the reference image queue 1 is greater than or equal to 0, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, and the L0 motion vector is calculated using formulas (1) and (2) mvE0(mvE0_x, mvE0_y), using formulas (3) and (4) to calculate the L1 motion vector mvE1(mvE1_x, mvE1_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index value is 0 in the reference image queue 1 is greater than or equal to 0, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left luminance sample position of the current block is located is -1, then formulas (1) and (2) are used to calculate the L0 motion Vector mvE0 (mvE0_x, mvE0_y), using the formulas (5) and (6) to calculate the L1 motion vector mvE1 (mvE1_x, mvE1_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is -1, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, then formulas (7) and (8) are used to calculate the L0 motion Vector mvE0 (mvE0_x, mvE0_y), using the formulas (3) and (4) to calculate the L1 motion vector mvE1 (mvE1_x, mvE1_y).

In Embodiment 3, if the current frame is a B frame, it is L0 of the time domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index is 0 in the reference image queue 1 The motion information is used as the reference motion information, and the L1 motion information of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is used as the reference motion information .

Embodiment 4: In order to obtain the temporal motion information of the current block, if the current frame where the current block is located is a P frame, that is, the current image is a P image, then the determination method of the temporal motion information refers to Embodiment 2, which will not be repeated here. If the current frame where the current block is located is a B frame, that is, the current image is a B image, then the following steps are used to determine the temporal motion information of the current block:

Step d1, based on the preset position of the current block, determine from the co-located frame the temporal adjacent block corresponding to the preset position, the motion information of the temporal adjacent block includes forward motion information and backward motion information. If the forward motion information of the neighboring blocks in the time domain is not available, and the backward motion information of the neighboring blocks in the time domain is not available, the time domain motion information of the current block is obtained according to the MVP method.

Exemplarily, the implementation process of step d1 is similar to that of step c1, and will not be repeated here.

Step d2, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion information of the current block can be determined according to the forward motion information of the neighboring block in the time domain The forward motion information in the time domain, and determine the backward motion information in the time domain motion information of the current block according to the forward motion information of the adjacent blocks in the time domain.

Exemplarily, compared with step c2, if the forward motion information of adjacent blocks in time domain is available, and the backward motion information of adjacent blocks in time domain is available, then determine the current Instead of determining the backward motion information in the time domain motion information of the current block according to the backward motion information of adjacent blocks in the time domain.

For the process of determining the backward motion information in the time domain motion information of the current block according to the forward motion information of adjacent blocks in the time domain, please refer to step c3, which will not be repeated here. For the process of determining the forward motion information in the time domain motion information of the current block according to the forward motion information of adjacent blocks in the time domain, refer to step c2, which will not be repeated here.

In another possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is available and the backward motion information of the neighboring blocks in the time domain is available, then the current The forward motion information in the time domain motion information of the block is determined, and the backward motion information in the time domain motion information of the current block is determined according to the backward motion information of the adjacent blocks in the time domain.

Exemplarily, compared with step c2, if the forward motion information of adjacent blocks in time domain is available, and the backward motion information of adjacent blocks in time domain is available, then determine the current Instead of determining the forward motion information in the time domain motion information of the current block according to the forward motion information of adjacent blocks in the time domain.

For the process of determining the forward motion information in the time domain motion information of the current block according to the backward motion information of adjacent blocks in the time domain, refer to step c4, which will not be repeated here. For the process of determining the backward motion information in the time domain motion information of the current block according to the backward motion information of adjacent blocks in the time domain, refer to step c2, which will not be repeated here.

Step d3, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is not available, then the time domain motion of the current block can be determined according to the forward motion information of the neighboring block in the time domain The forward motion information in the information, and determine the backward motion information in the time domain motion information of the current block according to the forward motion information of the adjacent blocks in the time domain.

Exemplarily, the implementation process of step d3 is similar to that of step c3, and will not be repeated here.

Step d4, if the forward motion information of the neighboring block in the time domain is not available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion of the current block can be determined according to the backward motion information of the neighboring block in the time domain The forward motion information in the information, and determine the backward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain.

Exemplarily, the implementation process of step d4 is similar to that of step c4, and will not be repeated here.

The above process of Embodiment 4 is described below in conjunction with a specific application scenario:

If the L0 reference index and the L1 reference frame index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 are both -1, Then the L0 reference index and the L1 reference index of the current block are both equal to 0. The size and position of the coding unit where the current block is located are used as the size and position of the current block, and then the L0 motion vector predictor and the L1 motion vector predictor obtained according to the MVP are respectively used as the L0 motion vector MvE0 and the L1 motion vector MvE1 of the current block, And make the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block both equal to 0, and end the derivation process of the time-domain motion information of the current block.

If the reference index value of the reference image queue 1 is 0, the L0 reference index and the L1 reference frame index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left luminance sample position of the current block is not both -1 , then the L0 reference index and the L1 reference index of the current block are both equal to 0. The distance index of the image corresponding to the L0 reference index of the current block is recorded as DistanceIndexL0, the distance index of the image corresponding to the L1 reference index of the current block is recorded as DistanceIndexL1; the BlockDistance of the image corresponding to the L0 reference index of the current block is recorded as BlockDistanceL0, that is, the current block The distance between the image corresponding to the L0 reference index of the current frame and the current frame; the BlockDistance of the image corresponding to the L1 reference index of the current block is recorded as BlockDistanceL1, that is, the distance between the image corresponding to the L1 reference index of the current block and the current frame.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block Forward motion information is available), then the L0 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef0( mvRef0_x, mvRef0_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L0 motion vector is recorded as DistanceIndexRef0. Exemplarily, BlockDistanceRef0 may be determined by the following formula: BlockDistanceRef0=DistanceIndexCol−DistanceIndexRef0.

If the L1 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block backward motion information is available), then the L1 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef1( mvRef1_x, mvRef1_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L1 motion vector is recorded as DistanceIndexRef1. Exemplarily, BlockDistanceRef1 may be determined by the following formula: BlockDistanceRef1=DistanceIndexCol−DistanceIndexRef1.

Then, the L0 reference index RefIdxL0 of the current block is equal to 0, the L0 motion vector of the current block is mvE0 (mvE0_x, mvE0_y), the L1 reference index RefIdxL1 of the current block is equal to 0, and the L1 motion vector of the current block is mvE1 (mvE1_x, mvE1_y). The L0 reference index RefIdxL0 and the L0 motion vector mvE0 (mvE0_x, mvE0_y) form the forward motion information, and the L1 reference index RefIdxL1 and the L1 motion vector mvE1 (mvE1_x, mvE1_y) form the backward motion information.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0, formula (1) and (2) Calculate the L0 motion vector mvE0 (mvE0_x, mvE0_y), and use formulas (5) and (6) to calculate the L1 motion vector mvE1 (mvE1_x, mvE1_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is -1, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, then formulas (7) and (8) are used to calculate the L0 motion Vector mvE0 (mvE0_x, mvE0_y), using the formulas (3) and (4) to calculate the L1 motion vector mvE1 (mvE1_x, mvE1_y).

In Embodiment 4, if the current frame is a B frame, it is L0 of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index is 0 in the reference image queue 1 The motion information is used as the reference motion information, and the L1 motion information of the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is used as the reference motion information .

Embodiment 5: In order to obtain the temporal motion information of the current block, if the current frame where the current block is located is a P frame, that is, the current image is a P image, then the determination method of the temporal motion information refers to Embodiment 2, which will not be repeated here. If the current frame where the current block is located is a B frame, that is, the current image is a B image, then the following steps are used to determine the temporal motion information of the current block:

Step s1 , based on the preset position of the current block, determine the adjacent temporal block corresponding to the preset position from the co-located frame, and the motion information of the adjacent temporal block includes forward motion information and backward motion information. If the forward motion information of the neighboring blocks in the time domain is not available, and the backward motion information of the neighboring blocks in the time domain is not available, the time domain motion information of the current block is obtained according to the MVP method.

Step s2, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion information of the current block can be determined according to the forward motion information of the neighboring block in the time domain and determine the backward motion information in the time domain motion information of the current block according to the backward motion information of the adjacent blocks in the time domain.

Exemplarily, the implementation process of step s2 is similar to that of step c2, and will not be repeated here.

Step s3, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is not available, then the time domain motion of the current block can be determined according to the forward motion information of the neighboring block in the time domain Forward motion information in the message.

Exemplarily, for the process of determining the forward motion information in the temporal motion information of the current block according to the forward motion information of adjacent blocks in the temporal domain, please refer to the step c2 of determining the forward motion information in the temporal motion information process, no more details.

Step s4, if the forward motion information of the neighboring blocks in the time domain is not available, and the backward motion information of the neighboring blocks in the time domain is available, then the time domain motion of the current block can be determined according to the backward motion information of the neighboring blocks in the time domain Backward motion information in the message.

Exemplarily, for the process of determining the backward motion information in the time domain motion information of the current block according to the backward motion information of adjacent blocks in the time domain, please refer to the step c2 of determining the backward motion information in the time domain motion information process, no more details.

The above-mentioned process of Embodiment 5 is described below in conjunction with a specific application scenario:

If the L0 reference index and the L1 reference frame index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 are both -1, Then the L0 reference index and the L1 reference index of the current block are both equal to 0. The size and position of the coding unit where the current block is located are used as the size and position of the current block, and then the L0 motion vector predictor and the L1 motion vector predictor obtained according to the MVP are respectively used as the L0 motion vector MvE0 and the L1 motion vector MvE1 of the current block, And make the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block both equal to 0, and end the derivation process of the time-domain motion information of the current block.

If the reference index value of the reference image queue 1 is 0, the L0 reference index and the L1 reference frame index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left luminance sample position of the current block is not both -1 , then the L0 reference index and the L1 reference index of the current block are both equal to 0. The distance index of the image corresponding to the L0 reference index of the current block is recorded as DistanceIndexL0, the distance index of the image corresponding to the L1 reference index of the current block is recorded as DistanceIndexL1; the BlockDistance of the image corresponding to the L0 reference index of the current block is recorded as BlockDistanceL0, that is, the current block The distance between the image corresponding to the L0 reference index of the current frame and the current frame; the BlockDistance of the image corresponding to the L1 reference index of the current block is recorded as BlockDistanceL1, that is, the distance between the image corresponding to the L1 reference index of the current block and the current frame.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block Forward motion information is available), then the L0 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef0( mvRef0_x, mvRef0_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L0 motion vector is recorded as DistanceIndexRef0. Exemplarily, BlockDistanceRef0 may be determined by the following formula: BlockDistanceRef0=DistanceIndexCol−DistanceIndexRef0.

If the L1 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block backward motion information is available), then the L1 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef1( mvRef1_x, mvRef1_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L1 motion vector is recorded as DistanceIndexRef1. Exemplarily, BlockDistanceRef1 may be determined by the following formula: BlockDistanceRef1=DistanceIndexCol−DistanceIndexRef1.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index value is 0 in the reference image queue 1 is greater than or equal to 0, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, so that the L0 reference index RefIdxL0 of the current block is equal to 0, and the L0 motion The vector is mvE0(mvE0_x, mvE0_y), the L1 reference index RefIdxL1 is equal to 0, and the L1 motion vector mvE1(mvE1_x, mvE1_y). The L0 reference index RefIdxL0 and the L0 motion vector mvE0 (mvE0_x, mvE0_y) form the forward motion information, and the L1 reference index RefIdxL1 and the L1 motion vector mvE1 (mvE1_x, mvE1_y) form the backward motion information.

On this basis, formula (1) and formula (2) are used to calculate L0 motion vector mvE0(mvE0_x, mvE0_y), and formula (3) and formula (4) are used to calculate L1 motion vector mvE1(mvE1_x, mvE1_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index value is 0 in the reference image queue 1 is greater than or equal to 0, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is -1, then the L0 reference index RefIdxL0 of the current block is equal to 0, and the current The L0 motion vector of the block is mvE0(mvE0_x, mvE0_y), let the L1 reference index RefIdxL1 of the current block be equal to -1, and let the L1 motion vector of the current block be mvE1(0,0). The L0 reference index RefIdxL0 and the L0 motion vector mvE0 (mvE0_x, mvE0_y) constitute the forward motion information, and the L1 reference index RefIdxL1 being -1 indicates that the backward motion information is not available.

On this basis, formula (1) and formula (2) are used to calculate the L0 motion vector mvE0 (mvE0_x, mvE0_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is -1, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, then the L0 reference index RefIdxL0 of the current block is equal to -1, The L0 motion vector of the current block is mvE0(0,0), the L1 reference index RefIdxL1 of the current block is equal to 0, and the L1 motion vector of the current block is mvE1(mvE1_x, mvE1_y). The L1 reference index RefIdxL1 and the L1 motion vector mvE1 (mvE1_x, mvE1_y) constitute the backward motion information, and the L0 reference index RefIdxL0 being -1 indicates that the forward motion information is not available.

On this basis, formula (3) and formula (4) are used to calculate the L1 motion vector mvE1(mvE1_x, mvE1_y).

Embodiment 6: In order to obtain the time domain motion information of the current block, if the current frame where the current block is located is a P frame, that is, the current image is a P image, then the determination method of the time domain motion information refers to Embodiment 2, which will not be repeated here. If the current frame where the current block is located is a B frame, that is, the current image is a B image, then the following steps are used to determine the temporal motion information of the current block:

Step w1, based on the preset position of the current block, determine the adjacent block in time domain corresponding to the preset position from the co-located frame, the motion information of the adjacent block in time domain includes forward motion information and backward motion information. If the forward motion information of the neighboring blocks in the time domain is not available, and the backward motion information of the neighboring blocks in the time domain is not available, the time domain motion information of the current block is obtained according to the MVP method.

Step w2, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion information of the current block can be determined according to the forward motion information of the neighboring block in the time domain The forward motion information in the time domain, and determine the backward motion information in the time domain motion information of the current block according to the forward motion information of the adjacent blocks in the time domain.

Exemplarily, compared with step c2, the backward motion information in the time domain motion information of the current block is determined according to the forward motion information of the time domain adjacent blocks, instead of the current Backward motion information in the block's temporal motion information. For the process of determining the backward motion information in the time domain motion information of the current block according to the forward motion information of adjacent blocks in the time domain, please refer to step c3, which will not be repeated here. For the process of determining the forward motion information in the time domain motion information of the current block according to the forward motion information of adjacent blocks in the time domain, please refer to step c2, which will not be repeated here.

Step w3, if the forward motion information of the neighboring block in the time domain is available, and the backward motion information of the neighboring block in the time domain is not available, then the time domain motion of the current block can be determined according to the forward motion information of the neighboring block in the time domain Forward motion information in the message.

Step w4. If the forward motion information of the neighboring block in the time domain is not available, and the backward motion information of the neighboring block in the time domain is available, then the time domain motion of the current block can be determined according to the backward motion information of the neighboring block in the time domain Backward motion information in the message.

The above process of Embodiment 6 is described below in conjunction with a specific application scenario:

If the L0 reference index and the L1 reference frame index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 are both -1, Then the L0 reference index and the L1 reference index of the current block are both equal to 0. The size and position of the coding unit where the current block is located are used as the size and position of the current block, and then the L0 motion vector predictor and the L1 motion vector predictor obtained according to the MVP are respectively used as the L0 motion vector MvE0 and the L1 motion vector MvE1 of the current block, And make the L0 reference index RefIdxL0 and the L1 reference index RefIdxL1 of the current block both equal to 0, and end the derivation process of the time-domain motion information of the current block.

If the reference index value of the reference image queue 1 is 0, the L0 reference index and the L1 reference frame index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left luminance sample position of the current block is not both -1 , then the L0 reference index and the L1 reference index of the current block are both equal to 0. The distance index of the image corresponding to the L0 reference index of the current block is recorded as DistanceIndexL0, the distance index of the image corresponding to the L1 reference index of the current block is recorded as DistanceIndexL1; the BlockDistance of the image corresponding to the L0 reference index of the current block is recorded as BlockDistanceL0, that is, the current block The distance between the image corresponding to the L0 reference index of the current frame and the current frame; the BlockDistance of the image corresponding to the L1 reference index of the current block is recorded as BlockDistanceL1, that is, the distance between the image corresponding to the L1 reference index of the current block and the current frame.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block Forward motion information is available), then the L0 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef0( mvRef0_x, mvRef0_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L0 motion vector is recorded as DistanceIndexRef0. Exemplarily, BlockDistanceRef0 may be determined by the following formula: BlockDistanceRef0=DistanceIndexCol−DistanceIndexRef0.

If the L1 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the upper-left corner luminance sample position of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is greater than or equal to 0 (the time-domain adjacent block backward motion information is available), then the L1 motion vector of the time domain motion information storage unit where the luminance sample corresponding to the luminance sample position of the upper left corner of the current block in the image whose reference index is 0 in the reference image queue 1 is recorded as mvRef1( mvRef1_x, mvRef1_y), the distance index of the image where the motion information storage unit is located is recorded as DistanceIndexCol, and the distance index of the image where the reference unit pointed to by the L1 motion vector is recorded as DistanceIndexRef1. Exemplarily, BlockDistanceRef1 may be determined by the following formula: BlockDistanceRef1=DistanceIndexCol−DistanceIndexRef1.

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index value is 0 in the reference image queue 1 is greater than or equal to 0, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, so that the L0 reference index RefIdxL0 of the current block is equal to 0, and the L0 motion The vector is mvE0(mvE0_x, mvE0_y), the L1 reference index RefIdxL1 is equal to 0, and the L1 motion vector mvE1(mvE1_x, mvE1_y). The L0 reference index RefIdxL0 and the L0 motion vector mvE0 (mvE0_x, mvE0_y) form the forward motion information, and the L1 reference index RefIdxL1 and the L1 motion vector mvE1 (mvE1_x, mvE1_y) form the backward motion information.

On this basis, formula (1) and formula (2) are used to calculate L0 motion vector mvE0(mvE0_x, mvE0_y), and formula (5) and formula (6) are used to calculate L1 motion vector mvE1(mvE1_x, mvE1_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image whose reference index value is 0 in the reference image queue 1 is greater than or equal to 0, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is -1, then the L0 reference index RefIdxL0 of the current block is equal to 0, and the current The L0 motion vector of the block is mvE0(mvE0_x, mvE0_y), let the L1 reference index RefIdxL1 of the current block be equal to -1, and let the L1 motion vector of the current block be mvE1(0,0). The L0 reference index RefIdxL0 and the L0 motion vector mvE0 (mvE0_x, mvE0_y) constitute the forward motion information, and the L1 reference index RefIdxL1 being -1 indicates that the backward motion information is not available.

On this basis, formula (1) and formula (2) are used to calculate the L0 motion vector mvE0 (mvE0_x, mvE0_y).

If the L0 reference index stored in the time-domain motion information storage unit where the luminance sample corresponding to the luminance sample position in the upper left corner of the current block is located in the image with a reference index value of 0 in the reference image queue 1 is -1, the reference image in the reference image queue 1 In an image with an index value of 0, the L1 reference index stored in the time domain motion information storage unit where the luminance sample corresponding to the upper left corner luminance sample position of the current block is located is greater than or equal to 0, then the L0 reference index RefIdxL0 of the current block is equal to -1, The L0 motion vector of the current block is mvE0(0,0), the L1 reference index RefIdxL1 of the current block is equal to 0, and the L1 motion vector of the current block is mvE1(mvE1_x, mvE1_y). The L1 reference index RefIdxL1 and the L1 motion vector mvE1 (mvE1_x, mvE1_y) constitute the backward motion information, and the L0 reference index RefIdxL0 being -1 indicates that the forward motion information is not available.

On this basis, formula (3) and formula (4) are used to calculate the L1 motion vector mvE1(mvE1_x, mvE1_y).

In the foregoing embodiments, the current block may also be referred to as the current prediction unit.

In a possible implementation manner, if the control information allows the current block to enable the time-domain motion information derivation mode, any embodiment of Embodiment 1 to Embodiment 6 may be used to derive the time-domain motion information of the current block. Exemplarily, the control information may include but not limited to: sequence level control information, frame level control information, Slice (piece) level control information, Tile (piece) level control information, Patch (piece) level control information, CTU (Coding Tee Unit, coding tree unit) level control information, LCU (Largest Coding Unit, maximum coding unit) level control information, block level control information, CU (Coding Unit, coding unit) level control information, PU (Prediction Unit, prediction unit) level Control information, etc., without limitation. For example, for the encoding end and the decoding end, if it is known that the control information allows the current block to enable the time-domain motion information derivation mode, any embodiment of Embodiment 1-Embodiment 6 is used to derive the time-domain motion information of the current block.

Embodiment 7: Based on the time-domain motion information derivation method of Embodiment 1-Embodiment 6, this embodiment of the present application proposes a codec method, which can be applied to the encoding end or the decoding end. For the encoding side, a motion information candidate list of the current block can be constructed, and the motion information candidate list includes the time-domain motion information of the current block. For the derivation method of the time-domain motion information of the current block, refer to Embodiment 1-Embodiment 6 Any embodiment of the present invention will not be repeated here. After obtaining the motion information candidate list, the encoding end sequentially traverses each motion information in the motion information candidate list, determines the rate-distortion cost value corresponding to the motion information, and selects the motion information corresponding to the smallest rate-distortion cost value. The position index information of the motion information corresponding to the minimum rate-distortion cost value in the motion information candidate list is carried into the coded bit stream of the current block, and transmitted to the decoding end.

For the decoding end, a motion information candidate list of the current block can also be constructed, the motion information candidate list includes the time-domain motion information of the current block, and the derivation method of the time-domain motion information of the current block can refer to Embodiment 1-Embodiment 6 Any embodiment in the above will not be repeated here. The motion information candidate list constructed by the decoder is the same as the motion information candidate list constructed by the encoder. The decoding end can analyze the position index information of the motion information in the motion information candidate list from the coded bit stream of the current block, and select the motion information corresponding to the position index information from the motion information candidate list, based on the selected motion information to Motion compensation is performed on the current block to obtain a prediction value of the current block, and there is no limitation on this motion compensation process.

Embodiment 8: The implementation process of Embodiment 1-Embodiment 7 can be applied in the AWP (Angular Weighted Prediction, angle weighted prediction) mode, that is, the motion information candidate list can be the motion information candidate list of the AWP mode, and the AWP mode uses the frame The idea of inner angle prediction is to first set the reference weight value of the peripheral position of the current block (integral pixel position and sub-pixel position), and then use the angle to obtain the corresponding weight value of each pixel position, and realize two different frames through the finally obtained weight array The weighting of the inter-prediction value, the motion information candidate list will be described below in conjunction with the AWP mode.

Step e1, when it is determined to start angle-weighted prediction for the current block, obtain the weight prediction angle of the current block, the weight prediction position of the current block, and the weight conversion rate of the current block. Based on at least one of the weight prediction angle of the current block, the weight prediction position of the current block, and the size of the current block, the starting position of the weight transformation of the current block is determined. Exemplarily, the weight transformation rate of the current block and the starting position of the weight transformation of the current block may be collectively referred to as weight configuration parameters of the current block.

Exemplarily, the weight prediction angle indicates the angle direction pointed to by the pixel position inside the current block. Referring to FIG. 4A , based on a certain weight prediction angle, it shows the pixel position inside the current block (such as pixel position 1, pixel The angular direction pointed by position 2 and pixel position 3) points to some peripheral position outside the current block. Referring to FIG. 4B , based on another weight prediction angle, it shows the angular direction pointed by the pixel positions (such as pixel position 2, pixel position 3 and pixel position 4) inside the current block, and the angular direction points to the outside of the current block a surrounding location.

The weight prediction angle may be any angle, such as any angle within 180 degrees, or any angle within 360 degrees, and there is no limitation on the weight prediction angle. For example, the weight prediction angle may be a horizontal angle or a vertical angle, or the absolute value of the slope of the weight prediction angle may be 2 to the nth power, where n is an integer, such as a positive integer, 0, a negative integer, and the like. For example, the absolute value of the slope of the weight prediction angle may be 1 (ie, 2 to the 0th power), 2 (ie, 2 to the 1st power), 1/2 (ie, 2 to the -1 power), etc. Referring to FIG. 4C , there are 8 kinds of weight prediction angles, and the absolute value of the slope is 2 to the nth power.

The weight prediction position is used to configure the reference weight value of the peripheral position outside the current block. For example, according to the weight prediction angle of the current block, the size of the current block, etc., the range of peripheral positions outside the current block (that is, the number of peripheral positions outside the current block) is determined, and the range of peripheral positions is divided into N equal parts. The value can be configured arbitrarily, such as 4, 6, 8, etc., and 8 is used as an example for illustration. The weight prediction position is used to indicate which peripheral position outside the current block is used as the starting position of the weight transformation of the current block, so that according to the starting position of the weight transformation Position configures the reference weight value for surrounding positions outside the current block. Referring to FIG. 4D , after dividing all surrounding positions into 8 equal parts, 7 weighted prediction positions can be obtained. On this basis, when the weight prediction position is 0, it may indicate that the surrounding position a0 is used as the starting position of the weight transformation of the surrounding positions outside the current block, and so on.

The weight conversion rate indicates the conversion rate of the reference weight value of the peripheral position outside the current block, and is used to indicate the change speed of the reference weight value. The weight conversion rate can be any number other than 0. For example, the weight conversion rate can be -4, - 2, -1, 1, 2, 4, 0.5, 0.75, 1.5, etc. When the absolute value of the weight conversion rate is 1, that is, the weight conversion rate is -1 or 1, which is used to indicate that the change speed of the reference weight value is 1. Based on this, the reference weight value needs to go through 0, 1, 2 from 0 to 8 , 3, 4, 5, 6, 7, 8 and other values, the reference weight value needs to go through 8, 7, 6, 5, 4, 3, 2, 1, 0 and other values from 8 to 0.

Step e2. Configure reference weight values for peripheral positions outside the current block according to the weight configuration parameters of the current block.

Exemplarily, the number of peripheral positions outside the current block may be determined based on the size of the current block and/or the weighted prediction angle of the current block, for example, the number of peripheral positions outside the current block may be determined based on the size of the current block and/or the weighted prediction angle of the current block. The number M of the surrounding positions of the block, and configure reference weight values for the M surrounding positions according to the weight configuration parameters of the current block.

Exemplarily, the peripheral positions outside the current block may include integer pixel positions, or sub-pixel positions, or both integer pixel positions and sub-pixel positions. The peripheral position outside the current block may include but not limited to: the peripheral position of the upper row outside the current block, or the peripheral position of the left column outside the current block, or the peripheral position of the lower row outside the current block, or the current block The perimeter position of the outer right column. Of course, the above are only examples of peripheral locations, and are not limited thereto.

In step e2, it is necessary to obtain an effective number (the number of peripheral positions outside the current block is an effective number) reference weight values. The effective number can be determined based on the size of the current block and/or the weight prediction angle of the current block, for example , the effective number can be determined in the following manner: ValidLenth=(N+(M>>X))<<1, N and M are the height and width of the current block respectively, and X is the absolute value of the slope of the weight prediction angle of the current block The log2 logarithmic value, such as 0 or 1.

A valid number of reference weight values may be monotonically increasing or monotonically decreasing. For example, the valid number of reference weight values may be [88...88765432100...00], that is, monotonically decreasing. For another example, the effective number of reference weight values may be [00...00123456788...88], that is, monotonically increasing. Of course, the above is just an example, without limitation.

Referring to FIG. 4E , there are several examples of effective number of reference weight values, which are not limited.

In a possible implementation manner, an effective number of reference weight values may be determined according to the weight conversion rate and the starting position of the weight conversion. For example, the reference weight value can be determined in the following manner: y=Clip3 (minimum value, maximum value, a*(x-s)), x represents the index of the surrounding position, that is, the value range of x can be 1-effective quantity value, such as When x is 1, it represents the first peripheral position, y represents the reference weight value of the first peripheral position, x is 2, it represents the second peripheral position, and y represents the reference weight value of the second peripheral position. a is used to indicate the weight transformation rate, and s is used to indicate the starting position of the weight transformation.

Clip3 is used to limit the reference weight value between the minimum value and the maximum value. Both the minimum value and the maximum value can be configured based on experience. For the convenience of description, in the following process, the minimum value is 0 and the maximum value is 8. .

s represents the starting position of the weight transformation, and s can be determined by the weight prediction position, for example, s=f (weight prediction position), that is, s is a function related to the weight prediction position. For example, after the range of surrounding positions outside the current block is determined, the effective number of surrounding positions can be determined, and all surrounding positions are divided into N equal parts. The value of N can be configured arbitrarily, such as 4, 6, 8, etc., and the weight The predicted position is used to indicate which surrounding position outside the current block is used as the target surrounding area of the current block, and the surrounding position corresponding to the weighted predicted position is the starting position of the weight transformation. Or, s may be determined by weight prediction angle and weight prediction position, for example, s=f(weight prediction angle, weight prediction position), that is, s is a function related to weight prediction angle and weight prediction position. For example, the range of surrounding positions outside the current block can be determined according to the weight prediction angle. After the range of surrounding positions outside the current block is determined, the effective number of surrounding positions can be determined, and all surrounding positions are divided into N equal parts, and the weight prediction position It is used to indicate which surrounding position outside the current block is used as the target surrounding area of the current block, and the surrounding position corresponding to the weight prediction position can be the starting position of weight transformation.

To sum up, an effective number of reference weight values of the current block can be obtained, and these reference weight values monotonically increase or monotonically decrease. Then, according to the valid number of reference weight values, the reference weight values of peripheral positions outside the current block are configured. Exemplarily, the number of peripheral positions outside the current block is an effective number, and the number of reference weight values is an effective number, therefore, an effective number of reference weight values may be configured as reference weight values for peripheral positions outside the current block.

Step e3, for each pixel position of the current block, determine the surrounding matching position pointed to by the pixel position from the surrounding positions outside the current block according to the weight prediction angle of the current block; determine the pixel according to the reference weight value associated with the surrounding matching position The target weight value of the position, and determine the associated weight value of the pixel position according to the target weight value of the pixel position.

Exemplarily, since the weighted prediction angle represents the angular direction pointed to by the pixel position inside the current block, therefore, for each pixel position of the current block, the angular direction pointed to by the pixel position is determined based on the weighted predicted angle, and then according to the The angular direction determines the surrounding matching location that the pixel location points to from the surrounding locations outside the current block.

For each pixel position of the current block, after determining the surrounding matching position pointed to by the pixel position, determine the target weight value of the pixel position based on the reference weight value associated with the surrounding matching position, for example, the reference weight value associated with the surrounding matching position The weight value is determined as the target weight value for the pixel location. Then, determine the associated weight value of the pixel position according to the target weight value of the pixel position. For example, the sum of the target weight value and the associated weight value of each pixel position can be a fixed preset value. Therefore, the associated weight value It can be the difference between the preset value and the target weight value. Suppose the preset value is 8, and the target weight value of the pixel position is 0, then the associated weight value of the pixel position is 8; if the target weight value of the pixel position is 1, the associated weight value of the pixel position is 7, so that By analogy, as long as the sum of the target weight value and the associated weight value is 8.

Step e4, obtaining a motion information candidate list, the motion information candidate list including at least one motion information candidate; obtaining the first target motion information and the second target motion information of the current block based on the motion information candidate list.

For the encoding side, based on the rate-distortion principle, one candidate motion information can be selected from the motion information candidate list as the first target motion information of the current block, and another candidate motion information can be selected from the motion information candidate list as the first target motion information of the current block. Two target motion information, the first target motion information is different from the second target motion information, which is not limited.

When the encoding end sends the encoded bit stream to the decoding end, the encoded bit stream carries indication information a and indication information b, the indication information a is used to indicate the index value 1 of the first target motion information of the current block, and the index value 1 indicates the first target motion The information is the number of candidate motion information in the motion information candidate list. The indication information b is used to indicate the index value 2 of the second target motion information of the current block, and the index value 2 indicates which candidate motion information the second target motion information is in the motion information candidate list.

For the decoding end, after receiving the coded bit stream, the indication information a and the indication information b are parsed from the coded bit stream. Based on the indication information a, the decoding end selects the candidate motion information corresponding to the index value 1 from the motion information candidate list, and the candidate motion information is used as the first target motion information of the current block. Based on the indication information b, the decoding end selects the candidate motion information corresponding to the index value 2 from the motion information candidate list, and the candidate motion information is used as the second target motion information of the current block.

Step e5, for each pixel position of the current block, determine the first predicted value of the pixel position according to the first target motion information of the current block, and determine the second predicted value of the pixel position according to the second target motion information of the current block; A weighted predictor for the pixel location is determined based on the first predictor, the target weight value, the second predictor and the associated weight value.

Exemplarily, assuming that the target weight value is the weight value corresponding to the first target motion information, and the associated weight value is the weight value corresponding to the second target motion information, then the weighted prediction value of the pixel position may be: (the pixel position of the first A predicted value*target weight value of the pixel position+second predicted value of the pixel position*associated weight value of the pixel position)/fixed preset value. Or, assuming that the target weight value is the weight value corresponding to the second target motion information, and the associated weight value is the weight value corresponding to the first target motion information, then the weighted prediction value of the pixel position can be: (the second prediction value of the pixel position value*the target weight value of the pixel position+the first predicted value of the pixel position*the associated weight value of the pixel position)/fixed preset value.

Step e6, determine the weighted prediction value of the current block according to the weighted prediction values of all pixel positions of the current block.

For example, the weighted prediction values of all pixel positions of the current block are combined to form the weighted prediction value of the current block.

To sum up, for the encoding end and the decoding end, in step e4, the motion information candidate list needs to be obtained, that is, the motion information candidate list of the current block needs to be constructed, and in this embodiment, it can be based on the time domain of the current block Motion information, build a candidate list of motion information for the current block. For example, based on the available motion information to be added to the motion information candidate list, the motion information candidate list of the current block is constructed, the available motion information includes the time domain motion information of the current block, and the derivation method of the time domain motion information of the current block can refer to the implementation Any embodiment in Example 1-Embodiment 6 will not be repeated here. In addition to the temporal motion information, the available motion information also includes at least one of the following motion information: spatial motion information; HMVP (History-based Motion Vector Prediction, history-based motion vector prediction) motion information; preset motion information.

Exemplarily, if the available motion information includes spatial motion information, before constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list, for the spatial adjacent blocks of the current block, if the spatial Adjacent blocks exist, and the spatial adjacent block adopts the inter prediction mode, then the motion information of the spatial adjacent block can be determined as the available motion information; and/or, for the first spatial adjacent block of the current block, if the The first spatial adjacent block exists, and the first spatial adjacent block adopts inter-frame prediction mode, and the motion information of the first spatial adjacent block is the same as the second spatial adjacent block of the current block (which may be at least one second If the motion information of adjacent spatial blocks) is different (the motion information of the second adjacent spatial block has been determined as available motion information), then the motion information of the first spatial adjacent block may be determined as available motion information.

Exemplarily, if the available motion information includes preset motion information, the preset motion information is determined as the available motion information before constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list ; The preset motion information may include but not limited to at least one of the following motion information: zero motion information; default motion information derived based on existing motion information in the motion information candidate list; existing motion information in the motion information candidate list Sports information.

In a possible implementation manner, based on the available motion information to be added to the motion information candidate list, the motion information candidate list of the current block is constructed, including but not limited to: for the available motion information to be added to the motion information candidate list,

If the available motion information is one-way motion information, the one-way motion information can be added to the motion information candidate list; if the available motion information is two-way motion information, then the two-way motion information can be clipped into the first one-way motion information information and second unidirectional motion information, and add the first unidirectional motion information to the motion information candidate list; or,

If the available motion information is unidirectional motion information, and the unidirectional motion information does not overlap with the existing motion information in the motion information candidate list, then add the unidirectional motion information to the motion information candidate list; if the available motion information is two-way motion information, then clip the two-way motion information into first one-way motion information and second one-way motion information, and add some one-way motion information to the motion information candidate list; for example, if the first one-way motion information The motion information does not overlap with the existing motion information in the motion information candidate list, then add the first unidirectional motion information to the motion information candidate list; or,

If the available motion information is unidirectional motion information, and the unidirectional motion information does not overlap with the existing motion information in the motion information candidate list, then add the unidirectional motion information to the motion information candidate list; if the available motion information is two-way motion information, then clip the two-way motion information into first one-way motion information and second one-way motion information, and add some one-way motion information to the motion information candidate list; for example, if the first one-way motion information The motion information does not overlap with the existing motion information in the motion information candidate list, then add the first one-way motion information to the motion information candidate list; if the first one-way motion information is the same as the existing motion information in the motion information candidate list If the information is repeated, and the second unidirectional motion information does not overlap with existing motion information in the motion information candidate list, then add the second unidirectional motion information to the motion information candidate list.

In the above embodiment, the first unidirectional motion information may be unidirectional motion information pointing to the reference frame in the first reference frame list; the second unidirectional motion information may be unidirectional motion pointing to the reference frame in the second reference frame list information.

Exemplarily, if the total number of candidate motion information existing in the motion information candidate list (that is, the total number currently present in the motion information candidate list) is an even number, the first reference frame list is the forward reference frame list of the current block , the second reference frame list is the backward reference frame list of the current block; if the total number of candidate motion information existing in the motion information candidate list is an odd number, the first reference frame list is the backward reference frame list of the current block, The second reference frame list is a forward reference frame list of the current block. Or, if the total number of candidate motion information existing in the motion information candidate list is an odd number, the first reference frame list is the forward reference frame list of the current block, and the second reference frame list is the backward reference frame list of the current block ; If the total number of candidate motion information existing in the motion information candidate list is an even number, the first reference frame list is the backward reference frame list of the current block, and the second reference frame list is the forward reference frame list of the current block.

Exemplarily, the first reference frame list is a forward reference frame list of the current block, and the second reference frame list is a backward reference frame list of the current block. Alternatively, the first reference frame list is a backward reference frame list of the current block, and the second reference frame list is a forward reference frame list of the current block.

The following describes the construction process of the motion information candidate list in the AWP mode in conjunction with several application scenarios. In each application scenario, the motion information candidate list is called AwpCandArray. Assume that the length of AwpCandArray is X, that is, X pieces of available motion information need to be added, and X can be configured according to experience. For example, the length of AwpCandArray is 5.

Application scenario 1: The motion information candidate list of AWP mode can be constructed through the following steps:

Step f1, as shown in FIG. 4F, F, G, C, A, B, and D are spatial adjacent blocks of the current block, and the "availability" of the motion information of F, G, C, A, B, and D can be determined . Exemplarily, if F exists and adopts the inter-frame prediction mode, the motion information of F is available motion information; otherwise, the motion information of F is unavailable motion information. If G exists and adopts the inter prediction mode, the motion information of G is available motion information; otherwise, the motion information of G is unavailable motion information. If C exists and adopts the inter prediction mode, the motion information of C is available motion information; otherwise, the motion information of C is unavailable motion information. If A exists and adopts the inter prediction mode, the motion information of A is available motion information; otherwise, the motion information of A is unavailable motion information. If B exists and adopts the inter prediction mode, the motion information of B is available motion information; otherwise, the motion information of B is unavailable motion information. If D exists and adopts inter-frame prediction mode, the motion information of D is available motion information; otherwise, the motion information of D is unavailable motion information.

Step f2, according to the order of F, G, C, A, B and D (the order is variable), add unidirectional available motion information to AwpCandArray and check for duplicates until the length of AwpCandArray is Y or the traversal ends. The value of Y can be smaller than the value of X, and can be configured according to experience. For example, the value of Y is 4, indicating that at least X-Y positions are reserved for time domain motion information.

Step f3, if the length of AwpCandArray is less than Y, according to the order of F, G, C, A, B and D (the order is variable), the bidirectional available motion information is split into a forward reference frame list (ie reference frame list The one-way motion information of List0) and the one-way motion information pointing to the backward reference frame list (that is, the reference frame list List1), first perform the duplication check operation of the one-way motion information, if not repeated, put it into AwpCandArray and check the duplicates until The length of AwpCandArray is Y or the end of traversal.

Step f4, split the time-domain motion information of the current block into unidirectional motion information pointing to the forward reference frame list and unidirectional motion information pointing to the backward reference frame list, and first perform a duplication check operation on the unidirectional motion information, if not Repeatedly put into AwpCandArray until the length of AwpCandArray is X or the traversal ends.

Exemplarily, regarding the manner of deriving the temporal motion information of the current block, reference may be made to Embodiment 1-Embodiment 6.

Step f5, if the length of AwpCandArray is less than X, then repeatedly fill the last unidirectional motion information in AwpCandArray until the length of AwpCandArray is X, and obtain the AwpCandArray of the current block.

Application scenario 2: The motion information candidate list of AWP mode can be constructed through the following steps:

Steps g1, F, G, C, A, B, D are adjacent blocks in the current block, if F exists and adopts inter-frame prediction mode, then the motion information of F is the available motion information; otherwise, the motion information of F is Sports information is not available. If G exists and the inter prediction mode is used, and the motion information of G is different from that of F, then the motion information of G is available; otherwise, the motion information of G is unavailable. If C exists and the inter prediction mode is used, and the motion information of C is different from that of G, the motion information of C is available; otherwise, the motion information of C is unavailable. If A exists and adopts the inter prediction mode, and the motion information of A is different from that of F, the motion information of A is available; otherwise, the motion information of A is unavailable. If B exists and adopts the inter prediction mode, the motion information of B is available motion information; otherwise, the motion information of B is unavailable motion information. If D exists and adopts the inter prediction mode, the motion information of D is different from the motion information of A, and the motion information of D is different from the motion information of G, then the motion information of D is the available motion information; otherwise, the motion information of D is Information is unavailable athletic information.

Step g2, according to the order of F, G, C, A, B and D (the order is variable), put the available motion information (one-way motion information or two-way motion information) into AwpCandArray until the length of AwpCandArray is Y or the traversal ends . The value of Y can be smaller than the value of X, and can be configured according to experience. For example, the value of Y is 4, indicating that at least X-Y positions are reserved for time domain motion information.

Step g3. Put the time-domain motion information of the current block into AwpCandArray until the length of AwpCandArray is X or the traversal ends.

Exemplarily, regarding the manner of deriving the temporal motion information of the current block, reference may be made to Embodiment 1-Embodiment 6.

Step g4, if the length of AwpCandArray is less than X, add bidirectional zero motion information to AwpCandArray until the length of AwpCandArray is X, and obtain the AwpCandArray of the current block. In bidirectional zero motion information, the motion vector value can be zero, the L0 reference picture index can be equal to 0, and the L1 reference picture index can be equal to 0.

Step g5. Assign a certain one-way motion information of the AwpCandIdx0+1th motion information in AwpCandArray to mvAwp0L0, mvAwp0L1, RefIdxAwp0L0 and RefIdxAwp0L1. For example, the following method is used for assignment:

Case 1. If AwpCandIdx0 is an even number, the one-way motion information pointing to the forward reference frame list is assigned to mvAwp0L0 and RefIdxAwp0L0, mvAwp0L1 is equal to zero, and RefIdxAwp0L1 is equal to -1. If the unidirectional motion information pointing to the forward reference frame list does not exist, assign the unidirectional motion information pointing to the backward reference frame list to mvAwp0L1 and RefIdxAwp0L1, mvAwp0L0 is equal to zero, and RefIdxAwp0L0 is equal to -1.

Case 2. If AwpCandIdx0 is an odd number, assign the one-way motion information pointing to the backward reference frame list to mvAwp0L1 and RefIdxAwp0L1, mvAwp0L0 is equal to zero, and RefIdxAwp0L0 is equal to -1. If the unidirectional motion information pointing to the backward reference frame list does not exist, assign the unidirectional motion information pointing to the forward reference frame list to mvAwp0L0 and RefIdxAwp0L0, mvAwp0L1 is equal to zero, and RefIdxAwp0L1 is equal to -1.

Step g6. Assign a certain one-way motion information of the AwpCandIdx1+1th motion information in AwpCandArray to mvAwp1L0, mvAwp1L1, RefIdxAwp1L0 and RefIdxAwp1L1. For example, the following method is used for assignment:

Case 1. If AwpCandIdx1 is an even number, assign the one-way motion information pointing to the forward reference frame list to mvAwp1L0 and RefIdxAwp1L0, mvAwp1L1 is equal to zero, and RefIdxAwp1L1 is equal to -1. If the unidirectional motion information pointing to the forward reference frame list does not exist, assign the unidirectional motion information pointing to the backward reference frame list to mvAwp1L1 and RefIdxAwp1L1, mvAwp1L0 is equal to zero, and RefIdxAwp1L0 is equal to -1.

Case 2. If AwpCandIdx1 is an odd number, assign the unidirectional motion information pointing to the backward reference frame list to mvAwp1L1 and RefIdxAwp1L1, mvAwp1L0 is equal to zero, and RefIdxAwp1L0 is equal to -1. If the unidirectional motion information pointing to the backward reference frame list does not exist, the unidirectional motion information pointing to the forward reference frame list is assigned to mvAwp1L0 and RefIdxAwp1L0, mvAwp1L1 is equal to zero, and RefIdxAwp1L1 is equal to -1.

Application scenario 3: The motion information candidate list of AWP mode can be constructed through the following steps:

Step h1-step h4 and the implementation process are the same as step f1-step f4, and will not be repeated here.

Step h5, if the length of AwpCandArray is less than X, generate four one-way motion information in turn, record the motion vector corresponding to the first one-way motion information in AwpCandArray as (x, y), and generate four one-way motion information for reference The frame information is the same as the reference frame information of the first unidirectional motion information in AwpCandArray, and the motion vectors corresponding to the four unidirectional motion information are (x0, y0), (x1, y1), (x2, y2), ( x3, y3), the generation of these motion vectors is shown in the following way. Add the newly generated motion information to AwpCandArray sequentially until the length of AwpCandArray is X.

x0=abs(x)<8? 8:(abs(x)<=64?((abs(x)*5+2)>>2):((abs(x)*9+4)>>3 ))

x0=x<0? -x0:x0

y0=y

x1=x

y1=abs(y)<8? 8:(abs(y)<=64?((abs(y)*5+2)>>2):((abs(y)*9+4)>>3 ))

y1=y<0? -y1:y1

x2=abs(x)<8? 8:(abs(x)<=64?((abs(x)*3+2)>>2):((abs(x)*7+4)>>3 ))

x2=x<=0? -x2:x2

y2=y

x3=x

y3=abs(y)<8? 8:(abs(y)<=64?((abs(y)*3+2)>>2):((abs(y)*7+4)>>3 ))

y3=y<=0? -y3:y3

Step h6. If the length of AwpCandArray is less than X, repeat the padding operation for the last unidirectional motion information in AwpCandArray until the length of AwpCandArray is X, and obtain the AwpCandArray of the current block.

Application scenario 4: The motion information candidate list of AWP mode can be constructed through the following steps:

Step k1-step k3, the implementation process is the same as step g1-step g3, and will not be repeated here.

Step k4, if the length of AwpCandArray is less than X, generate four unidirectional motion information in sequence, and add the newly generated motion information into AwpCandArray in sequence until the length of AwpCandArray is X. If the first valid motion information in AwpCandArray is unidirectional motion information, and the motion vector is recorded as (x, y), the reference frame information of the generated four unidirectional motion information is the same as the reference frame information of the first valid motion information , the generated motion vectors of the four unidirectional motion information are (x0, y0), (x1, y1), (x2, y2), (x3, y3), respectively. For the specific method, refer to step h5, which will not be repeated here.

If the first effective motion information in AwpCandArray is two-way motion information, and the motion vectors are recorded as (x, y) and (x', y'), the reference frame information of the four generated two-way motion information and the first effective motion information The same reference frame information. The motion vectors of the four generated bidirectional motion information are (x0, y0) and (x0', y0'), (x1, y1) and (x1', y1'), (x2, y2) and (x2', y2'), (x3,y3) and (x3',y3'). (x0, y0), (x1, y1), (x2, y2), (x3, y3) are determined in step h5. (x0',y0'), (x1',y1'), (x2',y2'), and (x3',y3') are determined as follows:

x0'=abs(x')<8? 8:(abs(x')<=64?((abs(x')*5+2)>>2):((abs(x')*9+ 4)>>3))

x0=x'<0? -x0':x0'

y0'=y'

x1'=x'

y1'=abs(y')<8? 8:(abs(y')<=64?((abs(y')*5+2)>>2):((abs(y')*9+ 4)>>3))

y1'=y'<0? -y1':y1'

x2'=abs(x')<8? 8:(abs(x')<=64?((abs(x')*3+2)>>2):((abs(x')*7+ 4)>>3))

x2'=x'<=0? -x2':x2'

y2'=y'

x3'=x'

y3'=abs(y')<8? 8:(abs(y')<=64?((abs(y')*3+2)>>2):((abs(y')*7+ 4)>>3))

y3'=y'<=0? -y3':y3'

Step k5, if the length of AwpCandArray is less than X, add bidirectional zero motion information to AwpCandArray until the length of AwpCandArray is X, and obtain the AwpCandArray of the current block. In bidirectional zero motion information, the motion vector value can be zero, the L0 reference picture index can be equal to 0, and the L1 reference picture index can be equal to 0.

Step k6. Assign a certain one-way motion information of the AwpCandIdx0+1th motion information in AwpCandArray to mvAwp0L0, mvAwp0L1, RefIdxAwp0L0 and RefIdxAwp0L1. For details, refer to step g5.

Step k7. Assign a certain one-way motion information of the AwpCandIdx1+1th motion information in AwpCandArray to mvAwp1L0, mvAwp1L1, RefIdxAwp1L0 and RefIdxAwp1L1. For details, refer to step g6.

Application scenario 5: For any application scenario in the above application scenario 1-application scenario 4, when the available motion information is put into AwpCandArray in the order of F, G, C, A, B, and D, the length of AwpCandArray can be X or end of traversal. Exemplarily, if the length of the AwpCandArray is X, the time-domain motion information of the current block may not be put into the AwpCandArray, that is, the AwpCandArray may not include the time-domain motion information of the current block. If the length of AwpCandArray is less than X at the end of the traversal, put the time domain motion information of the current block into AwpCandArray.

Application scenario 6: Assume that the minimum block size supported by AWP mode is 8, the maximum block size is 64, and a total of 8 weight prediction angles are supported, as shown in Figure 4C. Each weight prediction angle supports 7 weight prediction positions, see Figure 4D As shown, therefore, for each block, there are 56 patterns in total. Referring to FIG. 4E , it is a schematic diagram of reference weights for weight prediction positions supported by the AWP mode. As shown in Figure 4G, the weight prediction angle is divided into 4 partitions. According to the different areas where the weight prediction angle is located, the formula derived from the pixel-by-pixel weight is slightly different. Note that the block size of the current block is MxN, and M is the width. N is high, X is log2 (the absolute value of the slope of the weight prediction angle), and Y is the weight prediction position.

Taking weight prediction angle 0 and weight prediction angle 1 in angle partition 0 as an example, the derivation process is as follows:

The valid length ValidLenth is calculated using the following formula: ValidLenth=(N+(M>>X))<<1.

Set the reference weight value ReferenceWeights[x], the value range of x is 0~ValidLength-1; FirstPos=(ValidLength>>1)-6+Y*((ValidLength-1)>>3); ReferenceWeights[x]= Clip3(0,8,x-FirstPos); Exemplarily, FirstPos represents the starting position of the weight transformation in the above embodiment. Export weight SampleWeight[x][y] pixel by pixel: SampleWeight[x][y]=ReferenceWeights[(y<<1)+((x<<1)>>X)]

The chroma weight is exported as: directly take the position of the upper left corner corresponding to the 2*2 brightness weight, SampleWeightChroma[x][y]=SampleWeight[x>>1][y>>1], and the value range of x is 0～( M/2-1), the value range of y is 0~(N/2-1).

Embodiment 9: The implementation process of Embodiment 1-Embodiment 7 can be applied in UMVE (Ultimate MotionVector Expression, advanced motion information expression) mode, that is, the motion information candidate list can be the motion information candidate list of UMVE mode, and the UMVE mode is Offset is performed on the basis of the generated motion information, so as to obtain better motion information. The motion information candidate list will be described below in conjunction with the motion information candidate list in the UMVE mode.

Exemplarily, for the encoding end and the decoding end, it is necessary to obtain the motion information candidate list, that is, to construct the motion information candidate list of the current block, and in this embodiment, the current block can be constructed based on the time-domain motion information of the current block Candidate list of motion information. For example, a motion information candidate list for the current block is constructed based on available motion information to be added to the motion information candidate list. Exemplarily, the available motion information includes time-domain motion information of the current block, and the derivation manner of the time-domain motion information of the current block can refer to any embodiment in Embodiment 1-Embodiment 6, which will not be repeated here. In addition to the time-domain motion information, the available motion information also includes at least one of the following motion information: spatial-domain motion information; preset motion information.

Exemplarily, if the available motion information includes spatial motion information, before constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list, for the spatial adjacent blocks of the current block, if the spatial Adjacent blocks exist, and the spatial adjacent block adopts the inter prediction mode, then the motion information of the spatial adjacent block can be determined as the available motion information; and/or, for the first spatial adjacent block of the current block, if the The first spatial adjacent block exists, and the first spatial adjacent block adopts inter-frame prediction mode, and the motion information of the first spatial adjacent block is the same as the second spatial adjacent block of the current block (which may be at least one second If the motion information of adjacent spatial blocks) is different (the motion information of the second adjacent spatial block has been determined as available motion information), then the motion information of the first spatial adjacent block may be determined as available motion information.

Exemplarily, if the available motion information includes preset motion information, the preset motion information is determined as the available motion information before constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list ; The preset motion information may include but not limited to at least one of the following motion information: zero motion information; default motion information derived based on existing motion information in the motion information candidate list; existing motion information in the motion information candidate list Sports information.

In a possible implementation manner, based on the available motion information to be added to the motion information candidate list, constructing the motion information candidate list of the current block may include but not limited to: for the available motion information to be added to the motion information candidate list, If the available motion information is unidirectional motion information, add the unidirectional motion information to the motion information candidate list; if the available motion information is bidirectional motion information, add the bidirectional motion information to the motion information candidate list.

To sum up, for the construction process of the motion information candidate list in UMVE mode, the motion information of the adjacent blocks in the spatial domain of the current block can be put in first (need to be available motion information); and then the time domain motion information of the current block can be put in; Then put in zero motion information, and the following describes the construction process of the motion information candidate list in UMVE mode in combination with specific application scenarios. The motion information candidate list of the UMVE mode is called UmveList, assuming that the length of UmveList is X, that is, X pieces of available motion information need to be added, X can be configured according to experience, for example, the length of UmveList can be 2.

Step t1, as shown in FIG. 4F, F, G, C, A, and D are adjacent blocks in the space of the current block, and F, G, and C are sequentially determined in the order of F->G->C->A->D , the "availability" of A and D. If F exists and adopts the inter prediction mode, the motion information of F is available motion information; otherwise, the motion information of F is unavailable motion information. If G exists and the inter prediction mode is used, and the motion information of G is different from that of F, then the motion information of G is available; otherwise, the motion information of G is unavailable. If C exists and the inter prediction mode is used, and the motion information of C is different from that of G, the motion information of C is available; otherwise, the motion information of C is unavailable. If A exists and adopts the inter prediction mode, and the motion information of A is different from that of F, the motion information of A is available; otherwise, the motion information of A is unavailable. If D exists and adopts the inter prediction mode, the motion information of D is different from the motion information of A, and the motion information of D is different from the motion information of G, then the motion information of D is the available motion information; otherwise, the motion information of D is Information is unavailable athletic information.

Step t2: According to the order of F->G->C->A->D (the order can be changed according to the actual situation), the available airspace motion information is added to the UmveList in sequence until the length of the UmveList is X or the traversal ends.

Step t3. If the length of UmveList is less than X, add the time-domain motion information of the current block (refer to Embodiment 1 to Embodiment 6 for the derivation method) to UmveList until the length of UmveList is X or the traversal ends.

Step t4. If the length of the UmveList is less than X, add zero motion information into the UmveList until the length is X.

Exemplarily, based on the motion information candidate list of the UMVE mode, the prediction process of the UMVE mode includes:

As shown in Figure 5A, each initial motion information constructed through the motion information candidate list is used as a starting point to search from four directions, up, down, left, and right, and the offsets are 1/4, 1/2, 1, 2, and 4 pixels, respectively. , there are two starting points (2base), four offset directions (4direction) and 5 offsets, and a total of 40 candidate motion vectors are generated.

The encoding end performs motion compensation on the generated candidate motion vector through RDO to obtain the optimal prediction block, and encodes the motion information corresponding to the optimal prediction block, including umve_mv_idx (indicating the initial motion information), umve_step_idx (indicating the offset value ) and umve_dir_idx (indicating the offset direction). The decoding end obtains the motion vector of the current block by constructing the motion information candidate list, and parsing out umve_mv_idx (indicating the initial motion information), umve_step_idx (indicating the offset value) and umve_dir_idx (indicating the offset direction), and performing motion compensation to obtain the motion vector of the current block Predictive value.

Embodiment 10: The implementation process of Embodiment 1-Embodiment 7 can be applied in AFFINE (affine motion compensation) mode, that is, the motion information candidate list can be the motion information candidate list in AFFINE mode. Exemplarily, the motion information of the AFFINE mode is calculated from the motion information of the upper left corner and the upper right corner of the current block, and the motion information of the current block is calculated through a multi-parameter model. The AFFINE mode has a good prediction for scenes with rotation and scaling Effect.

Exemplarily, for the encoding end and the decoding end, it is necessary to obtain the motion information candidate list of the AFFINE mode, that is, it is necessary to construct the motion information candidate list of the current block, and in this embodiment, based on the time domain motion information of the current block, Construct the motion information candidate list of the current block. For example, a multi-parameter model is obtained based on available motion information, the multi-parameter model may include at least four parameters, the available motion information includes temporal motion information of the current block and motion information of spatial adjacent blocks of the current block; then, according to This multi-parameter model builds a candidate list of motion information for the current block. For the manner of deriving the time-domain motion information of the current block, reference may be made to any of the embodiments in Embodiment 1 to Embodiment 6, which will not be repeated here.

In a possible implementation manner, the multi-parameter model may include a four-parameter model and/or a six-parameter model; if the multi-parameter model includes a four-parameter model, it may be based on the temporal motion information of the current block and the A four-parameter model for motion information acquisition. Of course, a four-parameter model can also be obtained based on the motion information of two spatial adjacent blocks. If the multi-parameter model includes a six-parameter model, the six-parameter model can be obtained based on the temporal motion information of the current block and the motion information of two spatial adjacent blocks. Of course, a four-parameter model can also be obtained based on the motion information of three spatial adjacent blocks.

Exemplarily, the AFFINE mode adopts a multi-parameter model, such as a four-parameter model and a six-parameter model, as shown in FIG. 5B , and the motion information of each sub-block is obtained through the motion information of two or three control points. The AFFINE mode can be divided into two parts: AFFINE skip/direct and AFFINE inter. AFFINE skip/direct uses some information around the current block to derive a multi-parameter model, and builds a motion information candidate list based on the multi-parameter model. The motion information candidate list can support 5 multi-parameter models. The construction process of the AFFINE skip/direct motion information candidate list includes: copying the multi-parameter model from the adjacent AFFINE coding unit; using the motion information of the corner points around the current block to generate a multi-parameter model; if the number of multi-parameter models in the motion information candidate list has not reached Five, just pad with zeros. AFFINE inter adds MVD on the basis of AFFINE skip/direct. Each control point needs to transmit MVD, and the construction process of its motion information candidate list is similar.

The following describes the construction process of the motion information candidate list in the AFFINE mode in conjunction with specific application scenarios. Suppose the length of the motion information candidate list is X, that is, X multi-parameter models need to be added, and X can be configured according to experience, such as 5.

See Fig. 4F, F, G, C, A, B, D are adjacent blocks in the space of the current block, if at least one of A, B, and D is "available", scan in the order of A, B, D Get the first "available" neighboring block X0; otherwise, X0 "does not exist". Exemplarily, at least one of A, B, and D is "available", indicating that the motion information of at least one spatial adjacent block is available motion information. Assuming that the adjacent block X0 is A, it means that the motion information of A is available motion information. information.

If at least one of G and C is "available", scan in the order of G and C to get the first available adjacent block X1; otherwise, X1 "does not exist". At least one of G and C is "available", indicating that the motion information of at least one spatial adjacent block is available. Assuming that the adjacent block X1 is C, it means that the motion information of C is available.

If F is "available", record the adjacent block F as X2; otherwise, X2 "does not exist".

If H "exists", record H as X3; otherwise, X3 "does not exist", and the H block is the collocated position, that is, the adjacent block in the time domain of the current block. If H "exists", then the motion information of H is the current block For the time-domain motion information of the current block, reference may be made to any of the embodiments in Embodiment 1 to Embodiment 6 for the derivation manner of the time-domain motion information of the current block, which will not be repeated here.

Exemplarily, if X (X is X0, X1 or X2) "exists" and the L0 reference index of X is not equal to -1, then the L0 motion vector of X is MVX_L0; otherwise, MVX_L0 "does not exist". If X (X is X0, X1 or X2) is "present" and X's L1 reference index is not equal to -1, then X's L1 motion vector is MVX_L1; otherwise MVX_L1 is "absent". If X3 "exists" and the current image is a B image, then derive the L0 motion vector MVX3_L0 and the L1 motion vector MVX3_L1 according to the derivation method of temporal motion information, and the reference indexes of L0 and L1 of X3 are both equal to 0; if X3 "exists" and the current If the picture is a P picture, then the L0 motion vector MVX3_L0 is derived according to the derivation method of temporal motion information, the L0 reference index of X3 is equal to 0, and the L1 reference index is equal to -1; otherwise, MVX3_L0 and MVX3_L1 "do not exist".

To sum up, the motion information of X0, the motion information of X1, the motion information of X2, and the motion information of X3 can be obtained.

Exemplarily, based on the above motion information, a motion information candidate list of AFFINE mode can be constructed, and the motion information candidate list includes the following multi-parameter models in sequence: {CPMV0, CPMV1, CPMV2}, {CPMV0, CPMV1, CPMV3}, {CPMV0, CPMV2, CPMV3}, {CPMV1, CPMV2, CPMV3}, {CPMV0, CPMV1}, {CPMV0, CPMV2}, CPMV0-CPMV3 are the motion information generated by X0-X3 in turn.

Exemplarily, if the motion information candidate list does not reach five, zero motion information is filled.

Exemplarily, based on the motion information candidate list of the AFFINE mode, the prediction process of the AFFINE mode includes:

By constructing the motion information candidate list in AFFINE mode, the multi-parameter model in the motion information candidate list is used to calculate the mv(v _x , v _y ) of each small block, and perform motion compensation on the small block. The calculation formula is:

For the four-parameter model, the mv(v _x , v _y ) of each small block is calculated using the following formula:

For the six-parameter model, the mv(v _x , v _y ) of each small block is calculated using the following formula:

Affine motion compensation supports 4-parameter and 6-parameter motion models. If the current block uses two-way reference, then generate sub-block mv according to the 8*8 small block and perform motion compensation for each sub-block; otherwise, the current block uses unidirectional motion information (forward reference or backward reference), generate sub-block mv according to 4*4 or 8*8 small blocks, and perform motion compensation on each sub-block.

Exemplarily, Embodiment 1-Embodiment 10 may be implemented individually or in combination. For example, embodiment 1 and any embodiment of embodiment 2-embodiment 6 are realized in combination, embodiment 7 and embodiment 1-embodiment 6 are realized in combination, embodiment 8 and embodiment 1-embodiment Any embodiment of 6 is implemented in combination, embodiment 9 is implemented in combination with any embodiment of embodiment 1-embodiment 6, and embodiment 10 is implemented in combination with any embodiment of embodiment 1-embodiment 6.

Of course, the above are just a few examples, and the combinations among these embodiments are not limited.

Embodiment 11: Based on the same application concept as the above method, the embodiment of the present application also proposes a device for constructing a motion information candidate list, which is applied to the encoding end or the decoding end, as shown in FIG. 6A, which is a structural diagram of the device ,include:

The determination module 611 is configured to determine the temporal adjacent blocks of the current block from the co-located frame of the current frame if the current frame where the current block is located is a B frame; and determine the current block based on the motion information of the temporal adjacent blocks. Time-domain motion information of the block, the time-domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time-domain adjacent blocks is not available, the time-domain adjacent blocks If the backward motion information is available, the determination module 611 determines the backward motion information in the time domain motion information according to the backward motion information of the time domain adjacent blocks; the construction module 612 is used to determine the backward motion information based on the current block The time-domain motion information of the current block is constructed to construct a motion information candidate list of the current block.

In a possible implementation manner, if the forward motion information of the adjacent blocks in the time domain is available but the backward motion information of the adjacent blocks in the time domain is not available, the determination module 611 The forward motion information of the neighboring blocks determines the forward motion information in the time domain motion information, and the backward motion information in the time domain motion information is determined according to the forward motion information of the time domain neighboring blocks.

In a possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is available and the backward motion information of the neighboring blocks in the time domain is available, the determination module 611 The forward motion information of the adjacent blocks determines the forward motion information in the time domain motion information, and determines the backward motion information in the time domain motion information according to the backward motion information of the time domain adjacent blocks.

In a possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is available and the backward motion information of the neighboring blocks in the time domain is available, the determination module 611 The forward motion information of the adjacent blocks determines the forward motion information in the time domain motion information, and determines the backward motion information in the time domain motion information according to the forward motion information of the time domain adjacent blocks.

In a possible implementation manner, if the forward motion information of the neighboring blocks in the time domain is available, but the backward motion information of the neighboring blocks in the time domain is not available, the determination module 611 The forward motion information of the adjacent blocks determines the forward motion information in the time domain motion information.

In a possible implementation manner, if the forward motion information of the adjacent blocks in the time domain is not available, but the backward motion information of the adjacent blocks in the time domain is available, the determination module 611 The backward motion information of the neighboring blocks determines the forward motion information in the time domain motion information, and the backward motion information in the time domain motion information is determined according to the backward motion information of the time domain neighboring blocks.

In a possible implementation manner, the motion information candidate list includes an angle weighted prediction mode motion information candidate list; the construction module 612 is specifically configured to: construct the motion information candidate list based on the available motion information to be added to the motion information candidate list The motion information candidate list of the current block; wherein, the available motion information includes temporal motion information of the current block; the available motion information also includes at least one of the following motion information: spatial motion information; HMVP motion information; Preset exercise information.

In a possible implementation manner, the motion information candidate list includes a motion information candidate list of an advanced motion information expression mode; the construction module 612 is specifically configured to: construct The motion information candidate list of the current block; wherein, the available motion information includes temporal motion information of the current block; the available motion information also includes at least one of the following motion information: spatial motion information; preset motion information.

In a possible implementation manner, the motion information candidate list includes a motion information candidate list of an affine motion compensation mode; the construction module 612 is specifically configured to: acquire a multi-parameter model based on available motion information; wherein, the multiple The parameter model includes at least four parameters, the available motion information includes temporal motion information of the current block and motion information of spatial adjacent blocks of the current block; constructing the motion of the current block according to the multi-parameter model Information Candidate List.

Based on the same application concept as the above-mentioned method, the decoding end device (also called a video decoder) provided by the embodiment of the present application, from the hardware level, its hardware architecture schematic diagram can be specifically shown in FIG. 6B . Including: a processor 621 and a machine-readable storage medium 622, wherein: the machine-readable storage medium 622 stores machine-executable instructions that can be executed by the processor 621; the processor 621 is used to execute machine-executable instructions Instructions to implement the methods disclosed in the above examples of the present application. For example, the processor 621 is configured to execute machine-executable instructions to implement the following steps:

If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame;

Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks If it is not available, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain; based on the current block The time-domain motion information of the current block is constructed to construct a motion information candidate list of the current block.

Based on the same application concept as the above-mentioned method, the encoding end device (also called a video encoder) provided in the embodiment of the present application, from the hardware level, its hardware architecture schematic diagram can be specifically referred to in FIG. 6C . Including: a processor 631 and a machine-readable storage medium 632, wherein: the machine-readable storage medium 632 stores machine-executable instructions that can be executed by the processor 631; the processor 631 is used to execute machine-executable instructions Instructions to implement the methods disclosed in the above examples of the present application. For example, the processor 631 is configured to execute machine-executable instructions to implement the following steps:

If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame;

Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks If it is not available, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the time domain motion information according to the backward motion information of the adjacent blocks in the time domain; based on the current block The time-domain motion information of the current block is constructed to construct a motion information candidate list of the current block.

Based on the same application idea as the above-mentioned method, the embodiment of the present application also provides a machine-readable storage medium, on which several computer instructions are stored, and when the computer instructions are executed by a processor, the present invention can be realized. The method disclosed in the above examples is applied, for example, the method for constructing the motion information candidate list in the above embodiments.

The systems, devices, modules, or units described in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementing device is a computer, which may take the form of a personal computer, laptop computer, cellular phone, camera phone, smart phone, personal digital assistant, media player, navigation device, e-mail device, game control device, etc. desktops, tablets, wearables, or any combination of these. For the convenience of description, when describing the above devices, functions are divided into various units and described separately. Of course, when implementing the present application, the functions of each unit can be implemented in one or more pieces of software and/or hardware.

Those skilled in the art should understand that the embodiments of the present application may be provided as methods, systems, or computer program products. This application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Embodiments of the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The above descriptions are only examples of the present application, and are not intended to limit the present application. For those skilled in the art, various modifications and changes may occur in this application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims (12)

1. A construction method of motion information candidate list, is characterized in that, described method comprises: If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame; Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks Unavailable, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the motion information in the time domain according to the backward motion information of the adjacent blocks in the time domain; Constructing a motion information candidate list of the current block based on the time-domain motion information of the current block; Wherein, the backward motion information of the neighboring block in time domain includes the backward motion vector of the neighboring block in time domain and the backward reference frame index of the neighboring block in time domain, and the backward reference frame index corresponds to The backward reference frame of the adjacent block, the backward motion information in the time domain motion information includes the backward motion vector of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the current the backward reference frame of the block; Determining the backward motion information in the time domain motion information according to the backward motion information of the temporal domain adjacent block includes: based on the distance between the co-located frame and the backward reference frame of the temporal domain adjacent block distance, and the distance between the current frame and the backward reference frame of the current block, the backward motion vector of the adjacent block in the time domain is stretched to obtain the stretched motion vector, and the stretched Determine the motion vector of the current block as the backward motion vector of the current block; determine the backward motion information in the time domain motion information based on the backward motion vector of the current block and the backward reference frame index of the current block. 2. The method of claim 1, wherein, The determining the time-domain motion information of the current block based on the motion information of the adjacent blocks in the time domain includes: If the forward motion information of the neighboring blocks in the time domain is available and the backward motion information of the neighboring blocks in the time domain is available, then determine the time domain motion information according to the forward motion information of the neighboring blocks in the time domain The forward motion information in the time domain, and determine the backward motion information in the time domain motion information according to the backward motion information of the time domain adjacent blocks; Wherein, the forward motion information of the neighboring block in the time domain includes the forward motion vector of the neighboring block in the time domain and the forward reference frame index of the neighboring block in the time domain, and the forward reference frame index corresponds to The forward reference frame of the adjacent block, the forward motion information in the time domain motion information includes the forward motion vector of the current block and the forward reference frame index of the current block, the forward reference frame index The forward reference frame corresponding to the current block; Determining the forward motion information in the time domain motion information according to the forward motion information of the time domain adjacent block includes: based on the distance between the co-located frame and the forward reference frame of the time domain adjacent block distance, and the distance between the current frame and the forward reference frame of the current block, stretch the forward motion vector of the neighboring block in the time domain to obtain the stretched motion vector, and stretch the stretched motion vector Determine the motion vector of the current block as the forward motion vector of the current block; determine the forward motion information in the time domain motion information based on the forward motion vector of the current block and the forward reference frame index of the current block. 3. The method of claim 1, wherein, The determining the time-domain motion information of the current block based on the motion information of the adjacent blocks in the time domain includes: If the forward motion information of the neighboring blocks in the time domain is available but the backward motion information of the neighboring blocks in the time domain is not available, then determine the time domain motion according to the forward motion information of the neighboring blocks in the time domain Forward motion information in the message; Wherein, the forward motion information of the neighboring block in the time domain includes the forward motion vector of the neighboring block in the time domain and the forward reference frame index of the neighboring block in the time domain, and the forward reference frame index corresponds to The forward reference frame of the adjacent block, the forward motion information in the time domain motion information includes the forward motion vector of the current block and the forward reference frame index of the current block, the forward reference frame index The forward reference frame corresponding to the current block; Determining the forward motion information in the time domain motion information according to the forward motion information of the time domain adjacent block includes: based on the distance between the co-located frame and the forward reference frame of the time domain adjacent block distance, and the distance between the current frame and the forward reference frame of the current block, stretch the forward motion vector of the neighboring block in the time domain to obtain the stretched motion vector, and stretch the stretched motion vector Determine the motion vector of the current block as the forward motion vector of the current block; determine the forward motion information in the time domain motion information based on the forward motion vector of the current block and the forward reference frame index of the current block. 4. The method according to any one of claims 1-3, characterized in that, The determining the time-domain adjacent blocks of the current block from the co-located frame of the current frame includes: Based on the preset position of the current block, determine a temporal adjacent block corresponding to the preset position from co-located frames of the current frame; wherein the co-located frame is a reference image queue of the current frame The corresponding frame whose reference frame index in 1 is 0; the preset position is the upper left pixel position of the current block, or the upper right pixel position of the current block, or the lower left pixel position of the current block, or The pixel position of the lower right corner of the current block, or the center pixel position of the current block. 5. The method according to any one of claims 1-3, characterized in that, The motion information candidate list includes a motion information candidate list for an angle-weighted prediction mode; The constructing the motion information candidate list of the current block based on the time-domain motion information of the current block includes: Constructing a motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list; Wherein, the available motion information includes time domain motion information of the current block; the available motion information further includes at least one of the following motion information: spatial domain motion information; HMVP motion information; preset motion information. 6. The method of claim 5, wherein, If the available motion information includes spatial motion information, before constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list, the method further includes: For the spatial adjacent block of the current block, if the spatial adjacent block exists and the spatial adjacent block adopts an inter-frame prediction mode, then determine the motion information of the spatial adjacent block as available motion information; and / or, For the first spatial neighboring block of the current block, if the first spatial neighboring block exists, the first spatial neighboring block adopts inter-frame prediction mode, and the motion information of the first spatial neighboring block Different from the motion information of the second spatial adjacent block of the current block, the motion information of the first spatial adjacent block is determined as available motion information. 7. The method of claim 5, wherein, If the available motion information includes preset motion information, before constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list, the method further includes: Determining preset exercise information as available exercise information; Wherein, the preset motion information includes at least one of the following motion information: zero motion information; default motion information derived based on motion information existing in the motion information candidate list; motion information existing in the motion information candidate list . 8. The method according to claim 5, wherein said constructing the motion information candidate list of the current block based on the available motion information to be added to the motion information candidate list comprises: For the available motion information to be added to the motion information candidate list, If the available motion information is one-way motion information, then add the one-way motion information to the motion information candidate list; if the available motion information is two-way motion information, then clip the two-way motion information into the first single Adding the first one-way motion information to the motion information candidate list for the motion information and the second one-way motion information; Or, if the available motion information is unidirectional motion information, and the unidirectional motion information does not overlap with existing motion information in the motion information candidate list, then add the unidirectional motion information to the motion information candidate list; If the available motion information is bidirectional motion information, clip the bidirectional motion information into first unidirectional motion information and second unidirectional motion information; if the first unidirectional motion information and the motion information candidate list have The existing motion information is not repeated, then adding the first unidirectional motion information to the motion information candidate list; Or, if the available motion information is unidirectional motion information, and the unidirectional motion information does not overlap with existing motion information in the motion information candidate list, then add the unidirectional motion information to the motion information candidate list; If the available motion information is bidirectional motion information, clip the bidirectional motion information into first unidirectional motion information and second unidirectional motion information; If the motion information is not repeated, the first unidirectional motion information is added to the motion information candidate list; if the first unidirectional motion information is duplicated with the existing motion information in the motion information candidate list, and the second unidirectional motion information and If the existing motion information in the motion information candidate list is not repeated, the second unidirectional motion information is added to the motion information candidate list. 9. The method of claim 8, wherein The first unidirectional motion information is unidirectional motion information pointing to reference frames in the first reference frame list; The second unidirectional motion information is unidirectional motion information pointing to reference frames in the second reference frame list; Wherein, if the total number of candidate motion information existing in the motion information candidate list is an even number, the first reference frame list is the forward reference frame list of the current block, and the second reference frame list is the subsequent reference frame list of the current block. to the list of reference frames; If the total number of candidate motion information existing in the motion information candidate list is an odd number, the first reference frame list is the backward reference frame list of the current block, and the second reference frame list is the forward reference frame of the current block a list of frames; or, If the total number of candidate motion information existing in the motion information candidate list is an odd number, the first reference frame list is the forward reference frame list of the current block, and the second reference frame list is the backward reference frame of the current block frame list; If the total number of candidate motion information existing in the motion information candidate list is an even number, the first reference frame list is a backward reference frame list of the current block, and the second reference frame list is a forward reference frame list of the current block a list of frames; or, The first reference frame list is a forward reference frame list of the current block, and the second reference frame list is a backward reference frame list of the current block; or, The first reference frame list is a backward reference frame list of the current block, and the second reference frame list is a forward reference frame list of the current block. 10. A device for constructing a motion information candidate list, characterized in that the device comprises: A determining module, configured to determine the temporal adjacent blocks of the current block from the co-located frame of the current frame if the current frame where the current block is located is a B frame; and determine the current block based on the motion information of the temporal adjacent blocks time domain motion information, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent block is not available, the time domain adjacent block after If the backward motion information is available, the determination module determines the backward motion information in the time domain motion information according to the backward motion information of the time domain adjacent blocks; A construction module, configured to construct a motion information candidate list of the current block based on the temporal motion information of the current block; Wherein, the backward motion information of the neighboring block in time domain includes the backward motion vector of the neighboring block in time domain and the backward reference frame index of the neighboring block in time domain, and the backward reference frame index corresponds to The backward reference frame of the adjacent block, the backward motion information in the time domain motion information includes the backward motion vector of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the current A backward reference frame of a block; wherein, when the determining module determines the backward motion information in the time domain motion information according to the backward motion information of the time domain adjacent block, it is specifically used to: based on the co-located frame and The distance between the backward reference frames of the adjacent blocks in the time domain, and the distance between the current frame and the backward reference frame of the current block, the backward motion vector of the adjacent blocks in the time domain Perform stretching to obtain a stretched motion vector, and determine the stretched motion vector as the backward motion vector of the current block; based on the backward motion vector of the current block and the backward reference frame of the current block An index determines backward motion information in the temporal motion information. 11. A decoding device, comprising: a processor and a machine-readable storage medium, the machine-readable storage medium storing machine-executable instructions that can be executed by the processor; The processor is used to execute machine-executable instructions to implement the following steps: If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame; Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks Unavailable, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the motion information in the time domain according to the backward motion information of the adjacent blocks in the time domain; Constructing a motion information candidate list of the current block based on the time-domain motion information of the current block; Wherein, the backward motion information of the neighboring block in time domain includes the backward motion vector of the neighboring block in time domain and the backward reference frame index of the neighboring block in time domain, and the backward reference frame index corresponds to The backward reference frame of the adjacent block, the backward motion information in the time domain motion information includes the backward motion vector of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the current the backward reference frame of the block; Determining the backward motion information in the time domain motion information according to the backward motion information of the temporal domain adjacent block includes: based on the distance between the co-located frame and the backward reference frame of the temporal domain adjacent block distance, and the distance between the current frame and the backward reference frame of the current block, the backward motion vector of the adjacent block in the time domain is stretched to obtain the stretched motion vector, and the stretched Determine the motion vector of the current block as the backward motion vector of the current block; determine the backward motion information in the time domain motion information based on the backward motion vector of the current block and the backward reference frame index of the current block. 12. An encoding device, comprising: a processor and a machine-readable storage medium, the machine-readable storage medium storing machine-executable instructions that can be executed by the processor; The processor is used to execute machine-executable instructions to implement the following steps: If the current frame where the current block is located is a B frame, then determine the time-domain adjacent blocks of the current block from the co-located frame of the current frame; Determine the time domain motion information of the current block based on the motion information of the time domain adjacent blocks, the time domain motion information includes forward motion information and backward motion information; wherein, if the forward motion information of the time domain adjacent blocks Unavailable, the backward motion information of the adjacent blocks in the time domain is available, then determine the backward motion information in the motion information in the time domain according to the backward motion information of the adjacent blocks in the time domain; Constructing a motion information candidate list of the current block based on the time-domain motion information of the current block; Wherein, the backward motion information of the neighboring block in time domain includes the backward motion vector of the neighboring block in time domain and the backward reference frame index of the neighboring block in time domain, and the backward reference frame index corresponds to The backward reference frame of the adjacent block, the backward motion information in the time domain motion information includes the backward motion vector of the current block and the backward reference frame index of the current block, and the backward reference frame index corresponds to the current the backward reference frame of the block; Determining the backward motion information in the time domain motion information according to the backward motion information of the temporal domain adjacent block includes: based on the distance between the co-located frame and the backward reference frame of the temporal domain adjacent block distance, and the distance between the current frame and the backward reference frame of the current block, the backward motion vector of the adjacent block in the time domain is stretched to obtain the stretched motion vector, and the stretched Determine the motion vector of the current block as the backward motion vector of the current block; determine the backward motion information in the time domain motion information based on the backward motion vector of the current block and the backward reference frame index of the current block.

Images