EA045634B1 - VIDEO ENCODING AND VIDEO DECODING - Google Patents

Description

Technical field

The invention relates to video encoding and video decoding.

Prior Art

Video compression provides opportunities to reduce the payload in the transmission channel. Known video encoding standards allow bitstream data defining a video to be transmitted such that a bitstream receiver is capable of decoding the bitstream so as to generate a decoded video that substantially accurately represents the source video from which the encoded bitstream was derived.

Early video encoding standards were developed with video playback in mind on hardware where relatively low-to-medium quality restoration was acceptable. This includes portable devices or personal computing devices. To a large extent, the acceptability of specific quality levels is largely determined by the needs of the user depending on the capabilities of the playback equipment.

As receiving equipment improves in quality and functionality, user demands for higher-quality playback of the original video also increase. Thus, the technical challenge is to be able to play video on a player at a higher quality than has been achieved so far.

Description of drawings

Fig. 1 is an illustrative example of a frame to be encoded, transmitted, and decoded by the embodiments disclosed herein.

Fig. 2 is an illustrative example of a reference frame based on which the frame shown in FIG. 1.

Fig. 3 is an illustrative example of a frame block according to FIG. 1 to be encoded, transmitted and decoded by the embodiments disclosed herein.

Fig. 4 is an illustrative example of an inter-prediction candidate for the block in FIG. 3.

Fig. 5 is an illustrative example of an intra-prediction candidate for the block in FIG. 3.

Fig. 6 is a schematic diagram of a communication network according to an embodiment.

Fig. 7 is a schematic representation of the emitter (transmitter) of the communication network according to FIG. 6.

Fig. 8 is a diagram illustrating an encoder implemented in the emitter according to FIG. 7.

Fig. 9 is a flowchart of a prediction process performed in the encoder prediction unit of FIG. 8.

Fig. 10 is a schematic diagram of a communication network receiver according to FIG. 6.

Fig. 11 is a diagram illustrating a decoder implemented in the receiver according to FIG. 10; And

Fig. 12 is a flowchart of a prediction process performed by the decoder prediction module of FIG. 8.

Description of Embodiments

In general terms, embodiments disclosed herein provide efficient prediction of samples in a block at a video encoder and decoder by computing both an inter-predicted prediction hypothesis and an intra-predicted prediction hypothesis. The two prediction hypotheses are then combined together through a parametric transformation to form a new prediction hypothesis that can be more accurate in predicting the contents of the original block. The information may be provided to the decoder to enable it to determine a parametric transform to use.

According to one embodiment, a method for encoding a block of pixels in an image of a sequential plurality of images comprises:

calculating a set of motion-compensated inter-prediction candidates for a block of pixels based on a reference image in a sequential set of images;

calculating a set of intra-prediction candidates for a block of pixels based on a predetermined set of intra-prediction modes;

arranging a set of combinations of said motion compensated inter-prediction candidates and intra-prediction candidates;

determining, for a plurality of parameter sets, each parameter set configuring a parametric transformation of one of the motion-compensated inter-prediction candidates and one of the intra-prediction candidates, a combined prediction for each combination of said motion-compensated inter-prediction candidates and intra-prediction candidates;

evaluating each combined prediction relative to the dimension;

selecting one of the combined measurement-based predictions as the encoding of the pixel block.

Another embodiment provides an encoder configured to perform the above-described encoding process. The encoder may be included in a suitable signal processing device.

According to another embodiment, a method for decoding a bitstream comprises:

extracting motion predictor configuration information from the bit stream and, based on the motion predictor configuration information, calculating one or more motion compensated interprediction candidates for the block of samples;

extracting intra-prediction configuration information from the bitstream and, based on the intra-prediction configuration information, calculating one or more intra-prediction candidates, wherein the intra-prediction candidates are obtained by an intra-prediction process for a block of samples;

extracting from the bitstream information a parametric transform configuration enabling parameters to be determined for the parametric transform of one or more motion compensated inter-prediction candidates and one or more intra-prediction candidates;

determining parametric transformation parameters based on the parametric transformation configuration information; and calculating a combined transformed prediction for said block of samples by parametric transforming one or more motion compensated inter-prediction candidates and one or more intra-prediction candidates, wherein the parametric transform is controlled by parametric transform parameters.

Another embodiment provides a decoder configured to perform the above-described decoding process. The decoder may be included in a suitable signal receiving device.

Another embodiment provides a computer program product containing computer-executable instructions operative to configure a computer device to perform a method as described above.

As will be appreciated by the reader, a video representation typically contains a plurality of frames for sequential display on playback equipment. Various strategies are used to reduce the amount of data required to describe each frame in turn in the bit stream transmitted over the communication channel from the emitter (transmitter) to the receiver. As will be appreciated, the emitter will include an encoder for encoding frame data into a bitstream, and the receiver will include a decoder for generating frame data based on the information contained in the bitstream.

In embodiments of the present disclosure, each frame of the video presentation is divided into blocks. In the encoder, the content of a block is predicted based on the previously compressed content. This block prediction is subtracted from the actual block, resulting in a set of residual differences (residuals). In one embodiment, the residual data may be encoded using frequency domain conversion. However, it should be understood that the conversion of data from time domain to frequency domain may be specific to some implementations and is not essential to the performance of the disclosed embodiments.

Residues, whether converted or not, can then be scaled up to accommodate further processing.

The quantized residuals are then encoded using an entropy encoder into a set of coefficients that are placed into the block predictive bitstream. Various technical standards typically define the signal structure for the bitstream and the syntax for organizing the information along with block prediction identification so that the decoder can reliably extract the coefficients for decoding.

The decoder is configured to receive the bitstream. In general, a decoder is configured to expect a bit stream organized according to a predefined syntactic structure so that data can be retrieved predictably and reliably. Such a syntactic structure may be jointly agreed upon, for example, in accordance with a technical standard. Alternatively, the syntactic structure may be specified by a specific manufacturer or other party.

The decoder according to an embodiment extracts coefficients from the bit stream. These coefficients are decoded to obtain decoded data. The decoded data is then back-quantized. An option may exist to specify whether or not a region transform should be used on the received bitstream. If a domain transformation is specified, then the inversely quantized data is inversely transformed to produce the reconstructed residuals. Information about how to calculate the prediction block is also extracted from the bitstream and decoded; it is then used to form a prediction block using the previously reconstructed data. The reconstructed residuals are then added to the generated prediction block to obtain a reconstructed block.

It should be understood that, at the start of a transfer, it may be necessary to transfer a complete block of data,

- 2 045634 representing the full frame of the video presentation, without reference to any other frame. All subsequent prediction-based data can then be derived from the initial conditions established by such a reference frame. The reference frames can be constant for encoding and decoding the entire video representation, or can be updated. For example, a scene change may require updating a reference frame because frames of the new scene may have little or no resemblance to a previously established reference frame. Even within a scene, frames may vary sufficiently to reduce the efficiency of encoding relative to a reference frame compared to encoding and transmitting an updated reference frame.

It should be noted that the reference frame, for a particular frame in question, does not have to be before the frame in question in the video presentation sequence. The reference frame may be in the future relative to the frame in question.

The reference frames may be compressed and transmitted separately from the encoded parameters from which predictions may be formed. The reference frames may be indexed so that a reference frame to be used in generating a particular prediction block may be indicated in the bitstream.

The embodiments described herein relate to a prediction block generation process. Traditional video coding schemes use intra-prediction or inter-prediction. Intra-prediction methods involve computing a prediction using previously compressed content extracted from the same frame as the current block being predicted. In contrast, inter-prediction methods involve predicting the contents of a block of original samples using previously compressed portions of a video signal extracted from previously compressed frames.

The notation used to describe frame splits is varied. For example, in HEVC, the term slice is used to describe the division of a frame - a slice can contain many blocks. The terminology used to describe frame splits in the present disclosure is not intended to imply specific applicability of the present embodiments to a particular standardized technology, nor is it intended to be excluded from applicability to another standardized technology by using different terms for similar concepts.

In general, aspects of the present disclosure relate to a method for efficiently predicting samples within a block in a video encoder and decoder by computing both an inter-predicted prediction hypothesis and an intra-predicted prediction hypothesis. The two prediction hypotheses are then combined together to form a new prediction hypothesis that can be more accurate in predicting the contents of the original block.

An embodiment provides a method for combining two prediction hypotheses based on a combination algorithm. The combining algorithm may be preset in the receiver. The combining algorithm may be specified using information sent to the receiver. The receiver may store a library of combining algorithms and may select and use one of these stored algorithms based on information extracted from the received bit stream.

One approach to combining using a combining algorithm is to use a weighted average. This may be a weighted arithmetic average of two prediction hypotheses. The weights to be used in the weighted average may be preset, or parameters may be transmitted in a bit stream to the receiver to enable calculation, determination, or selection of the weights to be used.

In an embodiment, the information in the bitstream may contain weight values to be used in the weighted average.

In one embodiment, the information in the bitstream may include a weight selection parameter based on which the receiver may respond to select a particular one of a plurality of pre-stored weights. It should be clear that, for a weighted arithmetic average of two prediction hypotheses, one weight value is sufficient to describe the combination algorithm, with one of the two hypotheses being multiplied by the given weight w, and the other of the hypotheses being multiplied by the complementary weight (1-w) . Other algorithms may be implemented. In the case where more than two prediction hypotheses are used and they need to be combined, more weights can be specified to allow for weighted combination.

Combination algorithms based on another factor may be specified. For example, a function other than weighted average may be specified for which coefficients may be specified in the bit stream and received and retrieved at the receiver. For example, a linear or quadratic function can be used to interpolate values in a set of prediction hypotheses. The coefficients of a linear or quadratic function can be calculated, selected, or determined based on information extracted from the bitstream.

A parameter may be provided in the bitstream for extraction and use at the receiver specifying whether or not a particular algorithm should be used. In the case where two algorithms

- 3 045634 combinations are specified as available for use in the receiver, such a parameter may be a boolean.

In particular, the embodiment provides a mechanism used to compute motion compensated prediction. Motion compensated prediction is a technique used so far in numerous technologies such as MPEG, which provides certain motion information to the decoder. In general terms, motion compensated prediction involves determining the difference between the current frame and a reference frame, in terms of the transformation from the reference frame to the current frame. If the current frame and the reference frame were, for example, captured by a stationary camera in a scene in which no objects were moving (but, for example, lighting effects might have changed over time), simply subtracting the reference frame from the current frame provides an effective basis for compression. However, if the viewpoint at which the camera is mounted moves between the reference frame and the current frame, and/or objects in the viewed scene have moved, then simple subtraction is not sufficient to effectively describe the difference. In such a case, motion vectors must be used to describe the transformation between the current frame and the reference frame.

In the case when only the object moved in the scene, but the background of the scene remained unchanged, it is advisable to consider a division, for example, a block or frame. In this case, the transformation between the current frame block and the reference frame block can be described simply as a translation (translational movement), i.e. a shear that can be expressed as a motion vector.

This motion information appears as motion vectors and a reference frame index. Such information is compressed and included in the bit stream. Various techniques can be used to compress motion information. For example, when using fusion prediction, the decoder computes a set of candidates, namely a set of hypotheses constructed using motion information extracted from neighboring blocks. The index is then signaled to determine the correct hypothesis for the current block. This allows the decoder to perform motion compensation while requiring very few bits to compress the necessary motion information.

Embodiments described herein relate to a method for decoding a bit stream carrying information for reconstructing a plurality of frames of video content, each frame consisting of one or more blocks, the method comprising calculating a prediction of the block by calculating a plurality of different calculated prediction candidates and combining candidates for obtaining a combined prediction.

In one embodiment, the plurality of prediction candidates may comprise one candidate computed by inter-prediction and one candidate computed by intra-prediction. In one embodiment, the candidates may be combined through a parametric transformation. In one embodiment, the parametric transformation may comprise a weighted prediction characterized by one or more weights. In one embodiment, the weight or weights may be extracted from the bitstream. In one embodiment, the weight or weights may be calculated, determined, or selected based on coefficients or parameters extracted from the bitstream.

It should be clear that both intra- and inter-prediction implementations achieve satisfactory results in video compression schemes. However, circumstances may arise in which using a single prediction process to predict all samples in a block may not be ideal.

For example, we will now examine the case of an object moving against a static background. From frame to frame, the object will be presented in successively different positions against the background. In the case where an object cannot be segmented using a partitioning scheme when used in a video codec, existing prediction methods will not be able to accurately predict the contents of the current block. Using inter-prediction will result in a prediction candidate that contains an accurate prediction of the moving object, but not the background. In contrast, using intra-prediction methods can lead to satisfactory prediction of the background, but not the object details in the block.

It should be noted that what constitutes a satisfactory decoding result depends on the user's expectations. Objective standards measuring the fidelity of encoded images can be established, for example, based on an objective distortion metric. However, in conjunction with this, attention may be paid to the user's overall subjective impression that a particular decoding produces visually satisfactory results.

An example of this scenario is shown in FIG. 1 and 2. In FIG. 1 shows the current frame to be encoded, while FIG. Figure 2 shows a reference frame. Fig. 3 illustrates the current frame block from FIG. 1 for which the prediction is to be generated. In fig. 4 shows an intraprediction candidate, while FIG. Figure 5 shows an inter-prediction candidate. Obviously, no prediction candidate is an exact representation of the contents of the current block, as shown in FIG. 4.

Thus, the embodiment described herein generates a new prediction obtained by combining the two hypotheses as potentially better suited to predict the original content.

However, it should be noted that simply averaging each sample may not be ideal either. Depending on the accuracy of either the intra- or inter-prediction, there may be many cases in which one of the two predictions may be better suited to represent the contents of the current block, but the other prediction may still provide some missing information that would improve the prediction process. In such cases, performing a weighted prediction of the two candidates has a potential advantage. By appropriately selecting the weights to perform the weighted prediction, the best combined prediction can be obtained.

The embodiment described here uses a fixed set of weights. Weights are available in lookup tables, where different weights can be used for blocks with different characteristics, such as block size, or depending on other information available. Lookup tables can be provided in advance for suitably configured encoders and decoders. Alternatively, lookup tables may be transmitted from the encoder to the decoder either as part of the encoded video transmission or as a prior configuration transmission.

An index is then decoded from the bitstream to retrieve the correct entry from the lookup table so that the correct set of weights can be selected and used in making the prediction.

In another embodiment, rather than passing an index for the LUT entry, the weight or weights can be directly extracted from the bitstream.

In fig. 6 illustrates a structure containing a schematically illustrated video transmission network 10 in which an emitter (transmitter) 20 and a receiver 30 communicate over a communication channel 40. In practice, communications link 40 may comprise a satellite link, a cable network, a terrestrial broadcast network, a POTS link such as that used to provide Internet services to homes and small businesses, fiber optic communications systems, or combinations any of the above and any other possible communication media.

In addition, the disclosure also extends to communicating, through physical transfer, a storage medium on which a machine-readable record of an encoded bit stream is stored for passage to a suitably configured receiver capable of reading the medium and obtaining the bit stream from it. An example of this is the provision of a Digital Versatile Disc (DVD) or equivalent. The following description focuses on signal transmission, for example using an electronic or electromagnetic signal carrier, but should not be interpreted as excluding the above-mentioned approach using storage media.

As shown in FIG. 7, the emitter 20 is a computer device in its structure and function. It may share, with a general purpose computing device, certain features, but some features may be implementation specific given the specialized function for which the emitter 20 is to be installed. It should be clear which features may be of the general purpose type and which may need to be configured specifically for use in a video emitter (transmitter).

Thus, the emitter 20 includes a graphics processing unit 202 configured for specific use in graphics processing and the like. Emitter 20 also includes one or more other processors 204 that are either generally provided or configured for other purposes, such as mathematical operations, audio processing, communications link control, and so on.

Input interface 206 provides a means for receiving user input actions. Such user input actions may, for example, be caused by the user interacting with a particular input device, including one or more control buttons and/or switches, a keyboard, mouse or other pointing device, a speech recognition unit operated to receive and convert speech into control commands, a signal processor configured to receive and control processes from another device, such as a tablet or smartphone, or a remote control receiver. This list should be understood to be non-exhaustive and other forms of input, whether user-initiated or automated, may also be contemplated.

Likewise, output interface 214 is configured to provide a means for outputting signals to a user or other device. Such a pin may include a display signal for driving a local video display unit (VDU) or any other device.

Communication interface 208 implements a communication channel, whether broadcast or point-to-point, with one or more signal recipients. In the context of the present embodiment, the communication interface is configured to cause emission of a signal carrying a bit stream defining a video signal encoded by the emitter 20.

The processors 204, and especially for purposes of the present disclosure, the GPU 202 are driven to execute computer programs when the encoder is operating. In this case, they resort to storage means

- 5,045,634 data provided by mass storage device 208, which is implemented to provide large-scale data storage, albeit on a relatively slow access basis, and, in practice, will store computer programs and, in the current context, video presentation data in preparation for process execution coding.

Read-only memory (ROM) 210 is preconfigured with executable programs designed to provide the core functionality of emitter 20, and random access memory 212 is provided for quickly accessing and storing data and program instructions during execution of a computer program.

The operation of the emitter 20 will be described next with reference to FIG. 8. Fig. 8 shows a processing pipeline executed by the encoder implemented in the emitter 20, through executable instructions, on a data file representing a video representation containing a plurality of frames for sequential display as a sequence of pictures (frames).

The data file may also contain audio playback information to accompany the video presentation, and other additional information such as electronic program guide information, subtitles, or metadata to enable cataloging of the presentation. The handling of these aspects of the data file is not relevant to the present invention.

As shown in FIG. 8, the current picture or frame in a sequence of pictures is sent to the partitioning module 230, where it is divided into blocks of a certain size for sequential processing by the encoder. Each block is then input to prediction module 232, which seeks to discard temporal and spatial redundancies present in the sequence and obtain a prediction signal using the previously encoded content. The information that allows such a prediction to be calculated is encoded in the bit stream. This information may contain sufficient information to enable computation, including the ability to output to the receiver other information necessary to complete the prediction.

The prediction signal is subtracted from the original signal to obtain the residual signal. It is then input to transform module 234, which attempts to further reduce spatial redundancies in the block by using a more appropriate data representation. It should be noted that, in some embodiments, region transformation may be an optional step and may be omitted altogether. The use or non-use of a region transform may be signaled in the bitstream.

The resulting signal is then typically quantized by a quantization module 236, and finally the resulting data generated from the coefficients and the information needed to calculate the prediction for the current block is input to an entropy encoding module 238, which uses statistical redundancy to represent the signal in a compact form through short binary codes. Again, it should be noted that entropy encoding, in some embodiments, may be an optional feature, and in some cases may not be required at all. The use of entropy encoding may be signaled in the bitstream along with information to enable decoding, such as an index for the entropy encoding mode (eg, Huffman encoding) and/or codebook.

By replicating the operation of the emitter encoding means 20, the bit stream of block information elements can be constructed for transmission to a receiver or multiple receivers, as appropriate. A bitstream may also contain information elements that apply across multiple block information elements and are thus stored in the bitstream syntax independent of the block information elements. Examples of such information elements include configuration options, parameters applicable to a sequence of frames, and parameters related to the presentation of the video as a whole.

The prediction module 232 will now be described in more detail with reference to FIG. 9. As will be appreciated, this is an example only, and other approaches may be considered within the scope of this disclosure and the appended claims.

The following process is performed for each motion-compensated block in the interpredicted frame.

In step S1, a motion-compensated inter-prediction candidate list is calculated.

In step S2, a list of intra-prediction candidates is calculated. Several intra-prediction modes can be implemented. Intra-prediction modes can be determined based on the degree to which and the manner in which neighboring blocks are used to construct a block prediction. Various modes are defined in established technical standards such as H.264/AVC and H.265/HEVC. In this embodiment, a similar approach can be used so that DC, planar, pure vertical and pure horizontal intra-prediction modes can be defined. More modes than these may be defined, for example in H.264, in which an index with a value between 0 and 8 can be used to indicate one of nine available modes. Not all modes must be used. It is possible to consider a more limited set in which fewer or even one intraprediction candidate is considered.

- 6 045634

In step S3, a set of combinations of motion-compensated inter-prediction candidates and considered intra-prediction candidates is created. Loop L1 is then executed on each identified combination.

In this loop, step S4 considers a look-up table (LUT) containing weights that should be used to perform a weighted prediction between two available candidates. The LUT may depend on characteristics of the current block (such as its width and height) or characteristics of two prediction candidates (such as the magnitude of the motion vector of an inter-predicted candidate or the intra-prediction mode of an intra-predicted candidate). In one embodiment, multiple LUTs may be considered, and the encoder may select one LUT and transmit in bitstream information to enable the correct LUT to be identified or retrieved for use among the plurality of available LUTs.

Additionally, inside loop L1, loop L2 is executed over all elements in the LUT. For each element k, the following steps S6 and S7 are performed:

S6: The corresponding set of weights is retrieved at position k in the LUT.

S7: The combined prediction is computed by performing a weighted prediction using the two prediction candidates and the weights extracted at position k in the LUT.

After the L2 loop has been carried out over all the elements in the LUT, then, in step S8, the element k0 is selected from among the elements in the LUT. The selection process may depend, for example, on calculating the cost obtained for each element in the LUT, which may depend on the distortion and bit rate required to encode a block using that element. Other methods may be used to make this selection. The weights corresponding to the k0 entry in the LUT are then used to perform the final prediction for the current block. This completes the L1 cycle for the identified combination.

The selected index k0 is encoded in the bitstream along with the rest of the prediction information, including information to enable the motion compensated prediction to be computed. This may contain a motion vector and a reference frame index or a fusion index (to identify a valid fusion candidate from a predefined list) together with an index to identify the intra-prediction mode to use. In the case where only one intra-prediction mode is considered, there is no need to signal the intra-prediction mode in the bitstream.

The structural architecture of the receiver is shown in FIG. 10. It has elements that represent a computer-implemented device. Thus, receiver 30 includes a graphics processing unit 302 configured for specific use in graphics processing and the like. Receiver 30 also includes one or more other processors 304, either provided generally or configured for other purposes, such as math operations, audio processing, link control, and so on.

It should be understood that receiver 30 may be implemented in the form of a set-top box, portable personal electronic device, personal computer, or any other device suitable for reproducing video presentations.

Input interface 306 provides a means for receiving user input actions. Such user input actions may, for example, be caused by the user interacting with a particular input device, including one or more control buttons and/or switches, a keyboard, mouse or other pointing device, a speech recognition unit operated to receive and convert speech into control commands, a signal processor configured to receive and control processes from another device, such as a tablet or smartphone, or a remote control receiver. This list should be understood to be non-exhaustive and other forms of input, whether user initiated or automated, are possible.

Likewise, output interface 314 is configured to provide a means for outputting signals to a user or other device. Such output may include a television signal in a suitable format for driving a local television device.

Communication interface 308 implements a communication channel, whether broadcast or point-to-point, with one or more signal recipients. In the context of the present embodiment, the communication interface is configured to cause emission of a signal carrying a bitstream defining a video signal encoded by the receiver 30.

The processors 304, and especially for purposes of the present disclosure, the GPU 302 are driven to execute computer programs when the receiver is operating. This resorts to the data storage facilities provided by the mass memory device 308, which is implemented to provide large-scale data storage, albeit on a relatively slow access basis, and, in practice, will store computer programs and, in the current context, video presentation data being the result of the acceptance process.

Read-only memory (ROM) 310 is preconfigured with executable programs designed to provide the core functionality of receiver 30, and memory 312 with random

- 7 045634 access is provided for quick access and storage of data and program instructions during the execution of a computer program.

The function of the receiver 30 will now be described with reference to FIG. 11. In FIG. 11 shows a processing pipeline executed by a decoder implemented at receiver 20 through executable instructions on a bitstream received at receiver 30 containing structured information from which a video representation containing a reconstruction of frames encoded through the encoder functionality of emitter 20 can be derived.

The decoding process illustrated in FIG. 11 aims to reverse the process performed in the encoder. It should be clear that this does not mean that the decoding process is the exact inverse of the encoding process.

The received bitstream contains a sequence of encoded information elements, each element belonging to a block. The block information element is decoded in entropy decoding module 330 to obtain a block of coefficients and information necessary to compute a prediction for the current block. The block of coefficients is typically dequantized in dequantization module 332 and typically converted back to the spatial domain via transform module 334 .

As noted above, it should be understood that entropy decoding, dequantization, and inverse conversion need to be used at the receiver only if entropy encoding, quantization, and conversion, respectively, have been used at the transmitter.

The prediction signal is generated, as before, from pre-decoded samples from the current or previous frames and using information decoded from the bitstream by prediction module 336. A reconstruction of the original picture block is then obtained from the decoded residual signal and the calculated prediction block in reconstruction block 338.

By repeating the operation of decoding functionality on sequentially received block information elements, picture blocks can be reconstructed into frames, which can then be assembled to produce a video representation for playback.

An exemplary decoder algorithm complementary to the encoder algorithm described previously is shown in FIG. 12.

As noted above, receiver decoder functionality 30 extracts from the bit stream a sequence of block information elements as encoded by means of emitter encoder 20, determining block information and corresponding configuration information.

In general terms, the decoder uses information from previous predictions when constructing a prediction for the current block. In this case, the decoder can combine knowledge from interprediction, i.e. from the previous frame, and intra-prediction, i.e. from another block in the same frame.

That is, for an inter-predicted motion-compensated block in an inter-predicted frame, in step S21, information enabling generation of a motion-compensated inter-prediction candidate is extracted from the bit stream. Motion information enabling motion compensation can be obtained, for example, by directly decoding the motion vector and reference frame index from the bitstream, or through a fusion prediction process in which a list of fusion candidates is generated and the fusion index is extracted from the bitstream to identify the correct fusion traffic information.

After the information has been extracted, in step S22, a motion-compensated inter-prediction candidate is generated based on the extracted information.

Then, in step S23, a list of intra-prediction candidates is calculated, for example, a set of DC, planar, pure vertical and pure horizontal intra-prediction modes; a more limited set may be considered, where fewer or even just one intraprediction candidate is considered.

In step S24, an index is extracted from the bit stream to identify an intra prediction mode to use; In the case where only one intra-prediction mode is considered, there is no need to extract such mode identification index from the bit stream and therefore this step can be dispensed with. This step selects one of the candidates calculated in step S23 to use.

In step S25, an index k0 is extracted from the bit stream to identify an element to use from a previously stored LUT corresponding to the LUT used in the encoder process.

Next, in step S26, a look-up table (LUT) containing weights to be used to perform a weighted prediction between the two available candidates is considered. A plurality of LUTs may be implemented, and the LUT to be used in a particular case may depend on characteristics of the current block (such as its width and height) or characteristics of two prediction candidates (such as the magnitude of the inter-predicted candidate's motion vector or the intra-prediction mode -preproven candidate). In one embodiment, information may be extracted from the bitstream to enable selection of the appropriate

-

Claims (5)

LUT among many available LUTs. If one LUT is used for all circumstances, then there is no need to signal which LUT should be used in a particular case. In step S27, the weights at position k0 are extracted from the LUT identified in step S26. Then, in step S28, the weights are used to calculate a weighted average of the intra- and inter-predicted candidates and obtain the final prediction for the current block. The process described above is performed for all received block information elements for the entire transmission. On this basis, the construction of decoded blocks can be reconstructed into frames, and the frames into a reconstructed video representation. It should be understood that the invention is not limited to the embodiments described above, and various modifications and improvements may be made without deviating from the concepts described herein. Except where features are mutually exclusive, any of the features may be used alone or in combination with any other features, and the disclosure extends to and includes all combinations and subcombinations of one or more features described herein. CLAIM 1. A computer-implemented method for encoding a block of pixels in an image of a sequential set of images, containing: calculating a set of motion compensated inter-prediction candidates for a block of pixels based on the reference image in the sequential plurality of images; calculating a set of intra-prediction candidates for a block of pixels based on a predefined set of intra-prediction modes; arranging a set of combinations of said motion compensated inter-prediction candidates and intra-prediction candidates; a definition for a plurality of parameter sets, each parameter set configuring a parametric transformation of one of the motion-compensated inter-prediction candidates and one of the intra-prediction candidates, a combined prediction for each combination of said motion-compensated inter-prediction candidates and intra-prediction candidates, wherein at least one set of said parameters configures a parametric transformation comprising a weighted average of one or more motion-compensated inter-prediction candidates and one or more intra-prediction candidates, controlled by a weight or combination of weights, wherein the weight or combination of weights contains an entry in a lookup table with the corresponding weight index; evaluating each combined prediction against an objective distortion metric; selecting one of the combined predictions based on the result of said estimation as an encoding of the block of pixels. 2. An encoder operating to carry out the method according to claim 1. 3. A signal generation device containing an encoder according to claim 2. 4. A signal generated by modulating a bit stream onto a carrier, wherein the bit stream is a product of the encoding method of claim 1. 5. A computer-implemented method for decoding a bitstream, comprising: extracting motion predictor configuration information from the bit stream and, based on the motion predictor configuration information, calculating one or more motion compensated interprediction candidates for the block of samples; extracting intra-prediction configuration information from the bit stream and, based on the intra-prediction configuration information, calculating one or more intra-prediction candidates, wherein the intra-prediction candidates are obtained by an intra-prediction process for a block of samples; extracting from the bitstream information a parametric transform configuration enabling parameters to be determined for the parametric transform of one or more motion compensated inter-prediction candidates and one or more intra-prediction candidates, wherein the determined parameters comprise a weight index; retrieving from a lookup table mapping weight indexes to weights, weights or combinations of weights corresponding to the weight index; and calculating a combined transformed prediction for said block of samples by parametrically transforming one or more motion-compensated inter-prediction candidates and one or more intra-prediction candidates, wherein the parametric transformation comprises a weighted average of the one or more motion-compensated inter-prediction candidates and one or more more intra-prediction candidates, driven by a weight or combination of weights extracted/extracted from the lookup table. -