Encoders, decoders, methods and video bit streams, and computer programs, for hybrid video coding
The encoder and decoder system addresses inefficiencies in video coding by adapting interpolation filters based on video sequence characteristics, enhancing encoding efficiency and quality through reduced bitstream size and complexity.
Patent Information
- Application Number
- JP2024146506
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2019-03-12
- Filing Date
- 2024-08-28
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-03-12
AI Technical Summary
Existing video coding standards face challenges in achieving efficient trade-offs between bit rate, complexity, and quality due to fixed interpolation filters for fractional sample positions, which limits adaptability and encoding efficiency.
An encoder and decoder system that selects and adapts interpolation filters based on video sequence characteristics, using syntax elements to determine optimal processing schemes for different parts of the video, thereby improving encoding efficiency and quality by reducing bitstream size and computational effort.
The system enhances encoding efficiency by adapting interpolation filters to video sequence characteristics, reducing bitstream size and computational complexity while maintaining high quality, thus improving the overall coding performance.
Smart Images

Figure 2024161603000004 
Figure 2024161603000005 
Figure 2024161603000006
Abstract
Description
[Technical field]
[0001] An embodiment according to the invention relates to an encoder and a decoder for hybrid video coding. Further embodiments according to the invention relate to a method for hybrid video coding. Further embodiments according to the invention relate to a video bitstream for hybrid video coding. [Background technology]
[0002] All modern video coding standards use motion compensated prediction (MCP) with fractional sample accuracy of the motion vector (MV). Interpolation filters are used to find sample values of the reference picture at fractional sample positions.
[0003] In the High Efficiency Video Coding (HEVC) standard, which uses a quarter luma sample motion vector precision, there are 15 fractional luma sample positions that are calculated using various one-dimensional finite impulse response (FIR) filters. However, for each fractional sample position, the interpolation filter is fixed.
[0004] It is also possible to select an interpolation filter for each fractional sample position from a set of interpolation filters instead of having a fixed interpolation filter for each fractional sample position. Techniques that allow switching between different interpolation filters at the slice level include Non-separable Adaptive Interpolation Filter (AIF) [1], Separable Adaptive Interpolation Filter (SAIF) [2], Switched Interpolation Filtering with Offset (SIFO) [3], and Enhanced Adaptive Interpolation Filter (EAIF) [4]. The interpolation filter coefficients are either explicitly signaled for each slice (e.g., AIF, SAIF, EAIF), or it is indicated which of a set of predefined interpolation filters is to be used (as in the case of SIFO).
[0005] The current draft of the latest Versatile Video Coding (VVC) standard [5] supports the so-called Adaptive Motion Vector Resolution (AMVR). With AMVR, the precision of both the Motion Vector Predictor (MVP) and the Motion Vector Differential (MVD) can be selected jointly among the following possibilities: -Quarter Sample (QPEL): Quarter luma sample motion vector precision, - Full Sample (FPEL): Luma sample motion vector accuracy, -4 Samples (4PEL): Motion vector precision of 4 luma samples.
[0006] An additional syntax element is sent in the bitstream to select the desired precision. In the bi-predictive case, the indicated precision applies to both list0 and list1. AMVR is not available in skip and merge modes. Furthermore, AMVR is not available if all MVDs are equal to 0. If a precision other than QPEL is used, the MVP is first rounded to the specified precision before adding the MVDs to obtain the MV.
[0007] Some embodiments of the present disclosure may implement or be usable in the context of one or more of the concepts described above. For example, embodiments may implement the HEVC or VVC standards and may use AMVR and the MV precision described above.
[0008] A concept for encoding, decoding, and transmitting video data that offers an improved tradeoff between bit rate, complexity, and achievable quality of hybrid video coding is desired. Summary of the Invention
[0009] Aspects of the present disclosure relate to an encoder for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), for providing an encoded representation of a video sequence based on input video content, where the encoder determines one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., a block) of the video sequence, and determines (e.g., by a decoder) a processing scheme (e.g., interpolation filters) to be applied to the portion of the video sequence based on characteristics described by the one or more syntax elements (e.g., decoder settings signaled by the one or more syntax elements, or characteristics of the portion of the video sequence signaled by the one or more syntax elements). a processing scheme for obtaining samples (e.g., samples, e.g., fractional samples, associated with the video sequence) for motion compensated prediction at integer and / or fractional positions within a portion of the video sequence, encoding an index indicating the selected processing scheme (in dependence on the one or more syntax elements) such that a given encoded index value represents (e.g., is associated with) a different processing scheme depending on characteristics described by the one or more syntax elements (e.g., depending on the one or more syntax elements), and providing (e.g., transmitting, uploading, storing as a file) a bitstream including the one or more syntax elements and the encoded index as a coded representation of the video sequence.
[0010] By selecting a processing scheme to be applied to a part based on the characteristics of the part, it becomes possible to adapt the applied processing scheme to the part, for example according to the characteristics of the part. For example, different processing schemes can be selected for different parts of a video sequence. For example, a particular processing scheme may be particularly beneficial for encoding a part with given characteristics, since the particular processing scheme may be adapted to facilitate a particularly efficient encoding and / or to facilitate a particularly high quality of the encoded representation of the part with the given characteristics. Thus, by selecting a processing scheme depending on the characteristics of the part (for example, a set of possible processing schemes can be adapted to the characteristics of the part), the overall efficiency and quality of the encoding of the video sequence can be simultaneously improved. The efficiency of the encoding can refer to a low computational effort of the encoding or to a high compression rate of the encoded information.
[0011] Providing an encoded index indicating the selected processing scheme may ensure proper decoding of the part. Since a given value of the encoded index may represent different processing schemes depending on the characteristics of the part, the encoder may adapt the encoding of the index according to the characteristics. For example, the number of possible values of the encoded index may be limited to the number of processing schemes applicable to the part with that characteristic. Thus, even if the total number of selectable processing schemes (e.g., related to different characteristics) is large, the encoded index may require a small number of bits (which may, for example, be adapted to the number of selectable processing schemes in view of the actual characteristics of the currently considered part). Thus, the encoder may exploit knowledge about the characteristics of the part available to the encoder by determining one or more syntax elements in order to efficiently encode the index. Thus, the combination of selecting a processing scheme according to the characteristics (the characteristics may reduce the number of possible processing schemes for the currently considered part when compared to the total number of different processing schemes available) and encoding the index according to the characteristics may increase the overall efficiency of the encoding, although the selected processing scheme may be indicated in the bitstream.
[0012] According to one embodiment, the encoder is configured to adapt a mapping rule that maps index values to encoded index values depending on the characteristics described by one or more syntax elements. By adapting the mapping rule depending on the characteristics, it can be ensured that the indexes can be encoded such that a given encoded index value represents different processing schemes depending on the characteristics.
[0013] According to one embodiment, the encoder is configured to adapt the number of bins used to provide the coded index indicating the selected processing scheme depending on the characteristic described by one or more syntax elements. Thus, the encoder can adapt the number of bins to the number of processing schemes acceptable for the characteristic, so that the encoder can use a small number of bins when a small number of processing schemes are acceptable for the characteristic, thus reducing the size of the bitstream.
[0014] According to one embodiment, the encoder is configured to determine a set of permissible processing schemes in terms of characteristics described by one or more syntax elements and select a processing scheme from the determined set. By determining a set of permissible processing schemes in terms of characteristics, the encoder can infer information about selectable processing schemes from one or more syntax elements. Thus, the encoder uses information already available from determining one or more syntax elements to select a processing scheme, thus increasing the efficiency of encoding. Furthermore, relying on a set of processing schemes to select a processing scheme can limit the number of permissible processing schemes to the set of processing schemes, which allows the index to be coded to a small size of the coded index.
[0015] According to one embodiment, the encoder is configured to determine whether a single processing scheme is acceptable for a portion of a video sequence, and selectively omit inclusion of the coded index in response to finding that only a single processing scheme (e.g., a spatial filter or an interpolation filter) is acceptable for a portion of the video sequence. Thus, the encoder may, for example, infer from characteristics of the portion that a single processing scheme is acceptable for the portion and select the single processing scheme as the processing scheme to be applied to the portion. Omitting inclusion of the coded index in the bitstream may reduce the size of the bitstream.
[0016] According to one embodiment, the encoder is configured to select a mapping rule that maps index values to encoded index values such that the number of bins representing the encoded index value is adapted to the number of processing schemes acceptable for a portion of the video sequence. Thus, the encoder can adapt the number of bins to the number of processing schemes acceptable for a portion, so that the encoded index can have a small number of bins if a small number of processing schemes are acceptable for the portion, thus reducing the size of the bitstream.
[0017] According to one embodiment, at the encoder, the processing scheme is a filter (eg, an interpolation filter) or a set of filters (eg, interpolation filters) (eg, parameterized by the fractional part of the motion vector).
[0018] According to one embodiment, in the encoder, the filters in the set of filters are separably applicable to the portions of the video sequence (e.g., one-dimensional filters that can be applied sequentially in spatial directions, e.g., in the x-direction and / or y-direction). Thus, the filters can be selected individually, providing more flexibility for selecting filters and, as a result, allowing the set of filters to be adapted very precisely to the characteristics of the portions.
[0019] According to one embodiment, the encoder is configured to select different processing schemes (e.g., different interpolation filters) for different blocks (e.g., different prediction units PU and / or coding units CU and / or coding tree units CTU) within a video frame or picture (different interpolation filters for the same fractional sample positions in different parts of a video sequence or different parts of a single video frame or picture). Thus, the selection of the processing scheme can take into account individual characteristics of the different blocks, thus facilitating improved coding of the blocks, or coding efficiency, or reduction in the size of the bitstream.
[0020] According to one embodiment, the encoder is configured to code an index indicating the selected processing scheme (e.g., depending on one or more syntax elements), such that a given coded index value represents (e.g., is associated with) a different processing scheme depending on the motion vector precision (which can, for example, be signaled using a syntax element in the bitstream and can, for example, take one of the following values: QPEL, HPEL, FPEL, 4PEL). Thus, the coding of the index may depend on the MV precision, e.g., the number of processing schemes acceptable for the MV precision. For some settings of the MV precision, the number of acceptable processing schemes may be smaller than others, which is an efficient way to reduce the size of the bitstream.
[0021] According to one embodiment, in the encoder, the one or more syntax elements include at least one of: motion vector precision (e.g., quarter sample, half sample, full sample), fractional sample position, block size, block shape, number of prediction hypotheses (e.g., one hypothesis for uni-prediction, two hypotheses for bi-prediction), prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra), availability of the coded residual signal (e.g., coded block flag), spectral characteristics of the coded residual signal, reference picture signal (e.g., reference block defined by motion data, block edges from co-located prediction units PU within the reference block, high or low frequency characteristics of the reference signal), loop filter data (e.g., edge offset or band offset classification from sample adaptive offset filter SAO, deblocking filter decision and boundary strength), motion vector length (e.g., enable only long motion vectors or additional smoothing filters for specific directions), or adaptive motion vector resolution mode.
[0022] According to one embodiment, the encoder is configured to code an index indicating the selected processing scheme (e.g., depending on one or more syntax elements) such that a given coded index value represents (e.g., is associated with) different processing schemes depending on the fractional part (or fractional parts) of the motion vector (e.g., different coding schemes are used depending on whether all motion vectors of the considered part point to integer sample positions, e.g., luma sample positions, or not). Thus, the encoder can adapt the coding of the index to the number of processing schemes permissible for the fractional part of the motion vector. This is an efficient way to reduce the size of the bitstream, since the fractional part of the motion vector can be determined from the part.
[0023] According to one embodiment, the encoder is configured to selectively determine a set of processing schemes (e.g., "multiple possibilities coding") for a motion vector precision (e.g., a sub-sample motion vector resolution finer than the aforementioned motion vector precision, e.g., QPEL) and a minimum motion vector precision (e.g., 4PEL) (e.g., sub-sample motion vector precision, e.g., HPEL) or a motion vector precision (e.g., HPEL) between the maximum motion vector precision (e.g., QPEL) and a full-sample motion vector precision (e.g., FPEL) and select a processing scheme from the determined set. For example, at the maximum MV precision, it may be beneficial not to determine a set of processing schemes, since there may be one particularly efficient processing scheme, whereas at the maximum or full-sample MV precision, selection of a processing scheme may be unnecessary, since none of the selectable processing schemes may be applicable. In contrast, for MV precisions between the maximum MV precision and the full-sample or minimum MV precision, it may be particularly beneficial to adapt the processing scheme to the characteristics of the part.
[0024] According to one embodiment, the encoder is configured to encode an index (a processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering), a second FIR filtering, and a third FIR filtering.
[0025] According to one embodiment, the encoder is configured to encode an index (a processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering) and a second FIR filtering.
[0026] According to one embodiment, the encoder is configured to select (e.g., switch) between processing schemes (e.g., interpolation filters) having different characteristics (e.g., stronger vs. weaker low-pass characteristics). For example, a processing scheme or filter with stronger low-pass characteristics can attenuate high-frequency noise components, thereby improving the quality of the encoded representation of the video sequence.
[0027] According to one embodiment, the encoder is configured to select a mapping rule depending on the motion vector precision and the fractional sample position or fractional part (or multiple fractional parts) of the motion vector. Thus, the encoder can adapt the encoding of the index to the MV precision and the number of processing schemes acceptable for the fractional part of the motion vector. Since the MV precision and the fractional part of the motion vector can be determined from one or more syntax elements, it can be very efficient to use these characteristics to determine the number of acceptable processing schemes and select the mapping rule accordingly. According to one embodiment, the encoder is configured to determine a set of available processing schemes (eg, interpolation filters) depending on the motion vector precision.
[0028] According to one embodiment, the encoder is configured to select between quarter sample motion vector resolution, half sample motion vector resolution, full sample motion vector resolution, and four sample motion vector resolution (e.g., index elements describing the processing scheme or interpolation filters are selectively included when half sample motion vector resolution is selected, and index elements describing the processing scheme are, for example, omitted in other cases).
[0029] According to one embodiment, in the encoder, the processing scheme comprises one or more FIR filters, which are particularly stable filters for processing video sequences.
[0030] Another aspect of the disclosure relates to an encoder for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), for providing an encoded representation of a video sequence based on input video content, wherein the encoder determines one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., a block) of the video sequence, and selects (e.g., by a decoder) a processing scheme (e.g., an interpolation filter) to be applied to the portion of the video sequence based on characteristics described by the one or more syntax elements (e.g., decoder settings signaled by the one or more syntax elements, or characteristics of the portion of the video sequence signaled by the one or more syntax elements), the processing scheme comprising: the encoding scheme is for obtaining samples (e.g., samples, e.g., fractional samples, associated with a video sequence) for compensated prediction, and is configured to select an entropy coding scheme to be used for providing an encoded index indicative of a selected processing scheme according to characteristics described by the one or more syntax elements (e.g., according to the one or more syntax elements) (e.g., the number of bins described by the entropy coding scheme may be equal to zero if only one processing scheme is acceptable from the point of view of the one or more considered syntax elements), and to provide (e.g., transmit, upload, store as a file) a bitstream including the one or more syntax elements and including the encoded index (e.g., if the selected number of bins is greater than zero) as a coded representation of the video sequence.
[0031] Selecting a processing scheme to be applied to a part based on the characteristics of the part provides equivalent functions and advantages as described with respect to the previous aspect. Moreover, selecting an entropy coding scheme according to the characteristics allows the provision of the coded index to be adapted to the characteristics. For example, the number of possible values of the coded index may be limited to the number of processing schemes applicable to a part having the characteristics. Thus, the coded index may require a low number of bits or no bits in the bitstream, even if the total number of selectable processing schemes for different parts is large. Thus, selecting an entropy coding scheme based on knowledge of the characteristics of the part available to the encoder by determining one or more syntax elements may provide an efficient provision of the coded index. Thus, selecting both the process scheme and the entropy coding scheme according to the characteristics may improve the overall efficiency of the coding, even if the selected processing scheme is indicated in the bitstream.
[0032] According to one embodiment, in the encoder, the entropy coding scheme comprises a binarization scheme for providing the coded index. Since the entropy coding scheme is selected according to the characteristics, the binarization scheme can be adapted according to the characteristics, so that the binarization scheme can provide a particularly short representation of the coded index.
[0033] According to one embodiment, in the encoder, the binarization scheme includes the number of bins used to provide the encoded index. Thus, the number of bins used to provide the encoded index can be adapted according to the characteristics, so that the encoder can select a small number of bins or no bins to provide the encoded index, so as to provide efficient encoding.
[0034] According to one embodiment, the encoder is configured to adapt a mapping rule that maps index values to encoded index values depending on characteristics described by one or more syntax elements.
[0035] According to one embodiment, the encoder is configured to adapt the number of bins used to provide an encoded index indicative of a selected processing scheme depending on characteristics described by one or more syntax elements.
[0036] According to one embodiment, the encoder is configured to determine a set of processing schemes that are permissible in terms of the characteristics described by the one or more syntax elements, and to select a processing scheme from the determined set.
[0037] According to one embodiment, the encoder is configured to determine whether a single processing scheme is acceptable for a portion of the video sequence, and, in response to a finding that only a single processing scheme (e.g., a spatial filter or an interpolation filter) is acceptable for a portion of the video sequence, selectively omit inclusion of the encoded index.
[0038] According to one embodiment, the encoder is configured to select a mapping rule that maps index values to encoded index values such that the number of bins representing the encoded index values is adapted to the number of processing schemes permissible for a portion of the video sequence.
[0039] According to one embodiment, at the encoder, the processing scheme is a filter (eg, an interpolation filter) or a set of filters (eg, interpolation filters) (eg, parameterized by the fractional part of the motion vector).
[0040] According to one embodiment, in an encoder, the filters in the set of filters are separably applicable to portions of a video sequence (eg, one-dimensional filters that can be applied sequentially in a spatial direction, eg, the x-direction and / or the y-direction).
[0041] According to one embodiment, the encoder is configured to select different processing schemes (e.g., different interpolation filters) for different blocks (e.g., different prediction units PU and / or coding units CU and / or coding tree units CTU) within a video frame or picture (different interpolation filters for the same fractional sample positions in different parts of a video sequence or different parts of a single video frame or “picture”).
[0042] According to one embodiment, the encoder is configured to encode an index indicating a selected processing scheme (e.g., in response to one or more syntax elements), such that a given encoded index value represents (e.g., is associated with) a different processing scheme depending on the motion vector precision (which can be signaled, for example, using a syntax element in the bitstream and can, for example, take one of the following values: QPEL, HPEL, FPEL, 4PEL).
[0043] According to one embodiment, in the encoder, the one or more syntax elements include at least one of: motion vector precision (e.g., quarter sample, half sample, full sample), fractional sample position, block size, block shape, number of prediction hypotheses (e.g., one hypothesis for uni-prediction, two hypotheses for bi-prediction), prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra), availability of the coded residual signal (e.g., coded block flag), spectral characteristics of the coded residual signal, reference picture signal (e.g., reference block defined by motion data, block edges from co-located prediction units PU within the reference block, high or low frequency characteristics of the reference signal), loop filter data (e.g., edge offset or band offset classification from sample adaptive offset filter SAO, deblocking filter decision and boundary strength), motion vector length (e.g., enable only long motion vectors or additional smoothing filters for specific directions), or adaptive motion vector resolution mode.
[0044] According to one embodiment, the encoder is configured to code (e.g., in response to one or more syntax elements) an index indicating the selected processing scheme, such that a given coded index value represents (e.g., is associated with) a different processing scheme depending on the fractional part (or multiple fractional parts) of the motion vector (e.g., a different coding scheme is used depending on whether all motion vectors of the considered part point to integer sample positions, e.g., luma sample positions, or not).
[0045] According to one embodiment, the encoder is configured to selectively determine a set of processing schemes (e.g., "multiple possibility encoding") for motion vector precision (e.g., sub-sample motion vector precision, e.g., HPEL) between a maximum motion vector precision (e.g., a sub-sample motion vector resolution finer than the aforementioned motion vector precision, e.g., QPEL) and a minimum motion vector precision (e.g., 4PEL), or for motion vector precision (e.g., HPEL) between the maximum motion vector precision (e.g., QPEL) and a full-sample motion vector precision (e.g., FPEL), and select a processing scheme from the determined set.
[0046] According to one embodiment, the encoder is configured to encode an index (a processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering), a second FIR filtering, and a third FIR filtering.
[0047] According to one embodiment, the encoder is configured to encode an index (a processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering) and a second FIR filtering.
[0048] According to one embodiment, the encoder is configured to select (eg, switch) between processing schemes (eg, interpolation filters) having different characteristics (eg, stronger vs. weaker low-pass characteristics).
[0049] According to one embodiment, the encoder is configured to select a mapping rule depending on the motion vector precision and the fractional sample position or fractional part (or multiple fractional parts) of the motion vector. According to one embodiment, the encoder is configured to determine a set of available processing schemes (eg, interpolation filters) depending on the motion vector precision.
[0050] According to one embodiment, the encoder is configured to select between quarter sample motion vector resolution, half sample motion vector resolution, full sample motion vector resolution, and four sample motion vector resolution (e.g., index elements describing the processing scheme or interpolation filters are selectively included when half sample motion vector resolution is selected, and index elements describing the processing scheme are, for example, omitted in other cases). According to one embodiment, in the encoder, the processing scheme comprises one or more FIR filters.
[0051] Another aspect of the disclosure relates to a decoder for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), for providing output video content based on a coded representation of a video sequence, where the decoder obtains (e.g., receives, downloads, opens from a file) a bitstream as the coded representation of the video sequence, identifies from the bitstream one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., block) (and coded index) of the video sequence, and identifies a processing scheme (e.g., an interpolation filter selected by an encoder) for the portion of the video sequence based on the one or more syntax elements and the index (which may be included in the coded form of the bitstream), where the processing scheme is for obtaining samples for motion compensated prediction (e.g., samples, e.g., fractional samples, with respect to the video sequence) at integer and / or fractional positions within the portion of the video sequence, where the decoder is configured to assign different processing schemes to a given coded index value depending on the one or more syntax elements (e.g., depending on the value of the one or more syntax elements), and apply the identified processing scheme to the portion of the video sequence.
[0052] The decoder is based on the idea of identifying a processing scheme for a portion based on a syntax element and an index, whereas a given coded index value representing the value of the index may be associated with various processing schemes. Identification of the processing scheme may be achieved by assigning a processing scheme to the coded index value according to the syntax element. Thus, the decoder can use the information provided by the syntax element to identify a processing scheme from different processing schemes that may be associated with the coded index value. Since a given coded index value can be assigned to several processing schemes, the information density of the bitstream may be particularly high and still the decoder can identify the processing scheme.
[0053] According to one embodiment, at the decoder, the processing scheme is a filter (eg, an interpolation filter) or a set of filters (eg, interpolation filters) (eg, parameterized by the fractional part of the motion vector).
[0054] According to one embodiment, at the decoder, the filters in the set of filters are separably applicable to portions of the video sequence (eg, one-dimensional filters that can be applied sequentially in spatial directions, such as the x-direction and / or the y-direction).
[0055] According to one embodiment, the decoder is configured to select (e.g., apply) a filter (e.g., an interpolation filter) depending on a fractional portion (or multiple fractional portions) of a motion vector (e.g., depending on the fractional portions of one or more motion vectors) (e.g., by selecting a first filter if the motion vector end point is at a full sample position or at a full sample position in the direction under consideration, and by selecting a second filter different from the first filter if the motion vector end point is between full samples or between full samples in the direction under consideration).
[0056] According to one embodiment, the decoder is configured to select (e.g., apply) different filters (e.g., interpolation filters) for filtering in a first direction (e.g., x) and for filtering in a second direction (e.g., y) depending on a fractional portion (or multiple fractional portions) of the motion vector.
[0057] According to one embodiment, the decoder is configured to select (e.g., apply) different filters (e.g., interpolation filters) for filtering in a first direction and for filtering in a second direction, such that stronger low-pass filtering is applied to filtering in directions in which (one or more or all) fractional portions (or multiple fractional portions) of the motion vector lie on a full sample (e.g., when half-sample accurate motion vector(s) point to full sample position(s)) (compared to filtering applied in directions in which fractional portions (or multiple portions) of the motion vector lie between full samples).
[0058] According to one embodiment, the decoder is configured to select (e.g., apply, e.g., separately for each filtering direction) a filter (e.g., an interpolation filter) signaled by a syntax element (e.g., an interpolation filter index if_idx) or a predefined filter depending on one or more fractional parts (or multiple fractional parts) of the motion vector (e.g. a filter signaled by a syntax element is selected for filtering in a first direction, e.g. when considering position coordinates in a first direction, e.g. if one or more fractional parts (or multiple fractional parts) of the motion vector are between full samples, and a predefined filter is selected for filtering in a second direction different from the first direction, e.g. when considering position coordinates in a second direction, e.g. if all fractional parts (or multiple fractional parts) of the motion vector are on full samples).
[0059] According to one embodiment, the decoder is configured to select (e.g., apply, e.g., separately for each filtering direction) a first filter (e.g., an interpolation filter) signaled by a first syntax element (e.g., an interpolation filter index if_idx) or a second filter signaled by a filter syntax element depending on one or more fractional parts (or multiple fractional parts) of the motion vector (e.g., a first filter signaled by a filter syntax element is selected for filtering in a first direction, e.g. when considering position coordinates in the first direction, e.g. if one or more fractional parts (or multiple fractional parts) of the motion vector are between full samples, and a second filter signaled by a filter syntax element is selected for filtering in a second direction different from the first direction, e.g. when considering position coordinates in the second direction, e.g. if all fractional parts (or multiple fractional parts) of the motion vector are on full samples).
[0060] According to one embodiment, the decoder is configured to take over the filter selection from a neighboring block in merge mode (eg, selectively depending on whether the motion vector is identical to the motion vector of the neighboring block).
[0061] According to one embodiment, at the decoder, the selected (eg, applied) processing scheme defines (eg, spatial) filtering of the luma component and / or (eg, spatial) filtering of the chroma components. According to one embodiment, the decoder is configured to select (eg, apply) a filter depending on the prediction mode.
[0062] According to one embodiment, the decoder is configured to adapt mapping rules that map the encoded index values to instructions (eg, index values) that indicate a processing scheme according to one or more syntax elements.
[0063] According to one embodiment, the decoder is configured to adapt the number of bins used to identify the coded index indicative of the selected processing scheme depending on one or more syntax elements.
[0064] According to one embodiment, the decoder is configured to determine a set of processing schemes that are permissible in terms of the one or more syntax elements, where a processing scheme to be applied is identified based on the one or more syntax elements and an index.
[0065] According to one embodiment, the decoder is configured to determine whether a single processing scheme is acceptable for a portion of the video sequence, and, in response to a finding that only a single processing scheme (e.g., a spatial filter or an interpolation filter) is acceptable for a portion of the video sequence, selectively omit identifying the coded index.
[0066] According to one embodiment, the decoder is configured to select a mapping rule that maps the coded index value to an indication (e.g., index value) that indicates a processing scheme such that the number of bins representing the coded index value matches the number of processing schemes allowable for a portion of the video sequence.
[0067] According to one embodiment, the decoder is configured to select (e.g., apply) different processing schemes (e.g., different interpolation filters) for different blocks (e.g., different prediction units PU and / or coding units CU and / or coding tree units CTU) within a video frame or picture (different interpolation filters for the same fractional sample positions in different parts of a video sequence or different parts of a single video frame or “picture”).
[0068] According to one embodiment, the decoder is configured to decode (e.g., in response to one or more syntax elements) an index indicating a selected processing scheme, such that a given coded index value represents (e.g., is associated with) a different processing scheme depending on the motion vector precision (which may, for example, be signaled using a syntax element in the bitstream and may, for example, take one of the following values: QPEL, HPEL, FPEL, 4PEL).
[0069] According to one embodiment, the one or more syntax elements include at least one of: motion vector precision (e.g., quarter sample, half sample, full sample), fractional sample position, block size, block shape, number of prediction hypotheses (e.g., one hypothesis for uni-prediction, two hypotheses for bi-prediction), prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra), availability of the coded residual signal (e.g., coded block flag), spectral characteristics of the coded residual signal, reference picture signal (e.g., reference block defined by motion data, block edges from co-located prediction units PU within the reference block, high or low frequency characteristics of the reference signal), loop filter data (e.g., edge offset or band offset classification from sample adaptive offset filter SAO, deblocking filter decision and boundary strength), motion vector length (e.g., enable only long motion vectors or additional smoothing filters for specific directions), or adaptive motion vector resolution mode.
[0070] According to one embodiment, the decoder is configured to decode (e.g., in response to one or more syntax elements) an index indicating a selected processing scheme, such that a given coded index value represents (e.g., is associated with) different processing schemes depending on the fractional part (or multiple fractional parts) of a motion vector (e.g., different coding schemes are used depending on whether all motion vectors of the considered part point to integer sample positions, e.g., luma sample positions, or not).
[0071] According to one embodiment, the decoder is configured to selectively determine a set of processing schemes (e.g., "multiple possibility coding") for a motion vector precision (e.g., sub-sample motion vector precision, e.g., HPEL) between a maximum motion vector precision (e.g., a sub-sample motion vector resolution finer than the aforementioned motion vector precision, e.g., QPEL) and a minimum motion vector precision (e.g., 4PEL), or for a motion vector precision (e.g., HPEL) between the maximum motion vector precision (e.g., QPEL) and a full-sample motion vector precision (e.g., FPEL) for which a processing scheme to be applied is identified.
[0072] According to one embodiment, the decoder is configured to decode an index (processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering), a second FIR filtering, and a third FIR.
[0073] According to one embodiment, the decoder is configured to decode an index (a processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering) and a second FIR filtering.
[0074] According to one embodiment, the decoder is configured to apply processing schemes (eg, interpolation filters) having different characteristics (eg, stronger vs. weaker low-pass characteristics).
[0075] According to one embodiment, the decoder is configured to select a mapping rule depending on the motion vector precision and the fractional sample position or fractional part (or multiple fractional parts) of the motion vector. According to one embodiment, the decoder is configured to determine a set of available processing schemes (eg, interpolation filters) depending on the motion vector precision.
[0076] According to one embodiment, the decoder is configured to apply one of a quarter-sample motion vector resolution, a half-sample motion vector resolution, a full-sample motion vector resolution, and a four-sample motion vector resolution (e.g., index elements describing the processing scheme or interpolation filters are selectively included in the bitstream when half-sample motion vector resolution is selected, and index elements describing the processing scheme are, for example, omitted in other cases). According to one embodiment, in the decoder the processing scheme comprises one or more FIR filters.
[0077] Another aspect of the disclosure relates to a decoder for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), for providing output video content based on a coded representation of a video sequence, the decoder obtaining (e.g., receiving, downloading, opening from a file) a bitstream as the coded representation of the video sequence, identifying from the bitstream one or more syntax elements associated with a portion of the video sequence, and, depending on the one or more syntax elements (e.g., a motion vector precision syntax element), selecting an entropy filter to be used in the bitstream to provide a coded index indicating a processing scheme (e.g., an interpolation filter, e.g., selected by an encoder) for the portion (e.g., a block) of the video sequence. determine a coding scheme, the processing scheme for obtaining samples for motion compensated prediction (e.g., samples, e.g., fractional samples, for the video sequence) at integer and / or fractional positions within a portion of the video sequence, and is configured to identify an index encoded in the bitstream if the determined entropy coding scheme specifies a number of bins greater than zero, decode the index, identify (e.g., select) a processing scheme based on one or more syntax elements and based on the decoded index if the number of bins specified by the determined coding scheme is greater than zero (or identify a processing scheme based on the one or more syntax elements only if the determined number of bins is zero), and apply the identified processing scheme to the portion of the video sequence.
[0078] The method is based on the idea that the decoder identifies a processing scheme for a portion by determining an entropy coding scheme based on a syntax element. From the determined entropy coding scheme, the decoder can infer how to identify the processing scheme. By identifying, if the number of bins defined by the entropy coding scheme is greater than zero, the decoder can infer the information encoded in the bitstream by the entropy coding scheme independent of the number of bins. In other words, the decoder can use the information that the number of bins is zero to identify the processing scheme. Furthermore, since the decoder can use the syntax elements and indexes to identify the processing scheme, the decoder can identify the processing scheme even if a given coded index value can be associated with various processing schemes. Thus, the decoder can identify the processing scheme even if the number of bins in the bitstream for transmitting information about the processing scheme is small or zero. According to one embodiment, at the decoder, the entropy coding scheme includes a binarization scheme that is used to provide the coded index. According to one embodiment, at the decoder, the binarization scheme includes the number of bins used to provide the coded index.
[0079] According to one embodiment, at the decoder, the processing scheme is a filter (eg, an interpolation filter) or a set of filters (eg, interpolation filters) (eg, parameterized by the fractional part of the motion vector).
[0080] According to one embodiment, at the decoder, the filters in the set of filters are separably applicable to portions of the video sequence (eg, one-dimensional filters that can be applied sequentially in spatial directions, such as the x-direction and / or the y-direction).
[0081] According to one embodiment, the decoder is configured to select (e.g., apply) a filter (e.g., an interpolation filter) depending on a fractional portion (or multiple fractional portions) of a motion vector (e.g., depending on the fractional portions of one or more motion vectors) (e.g., by selecting a first filter if the motion vector end point is at a full sample position or at a full sample position in the direction under consideration, and by selecting a second filter different from the first filter if the motion vector end point is between full samples or between full samples in the direction under consideration).
[0082] According to one embodiment, the decoder is configured to select (e.g., apply) different filters (e.g., interpolation filters) for filtering in a first direction (e.g., x) and for filtering in a second direction (e.g., y) depending on a fractional portion (or multiple fractional portions) of the motion vector.
[0083] According to one embodiment, the decoder is configured to select (e.g., apply) different filters (e.g., interpolation filters) for filtering in a first direction and for filtering in a second direction, such that stronger low-pass filtering is applied to filtering in directions in which (one or more or all) fractional portions (or multiple fractional portions) of the motion vector lie on a full sample (e.g., when half-sample accurate motion vector(s) point to full sample position(s)) (compared to filtering applied in directions in which fractional portions (or multiple portions) of the motion vector lie between full samples).
[0084] According to one embodiment, the decoder is configured to select (e.g., apply, e.g., separately for each filtering direction) a filter (e.g., an interpolation filter) signaled by a syntax element (e.g., an interpolation filter index if_idx) or a predefined filter depending on one or more fractional parts (or multiple fractional parts) of the motion vector (e.g. a filter signaled by a syntax element is selected for filtering in a first direction, e.g. when considering position coordinates in a first direction, e.g. if one or more fractional parts (or multiple fractional parts) of the motion vector are between full samples, and a predefined filter is selected for filtering in a second direction different from the first direction, e.g. when considering position coordinates in a second direction, e.g. if all fractional parts (or multiple fractional parts) of the motion vector are on full samples).
[0085] According to one embodiment, the decoder is configured to select (e.g., apply, e.g., separately for each filtering direction) a first filter (e.g., an interpolation filter) signaled by a first syntax element (e.g., an interpolation filter index if_idx) or a second filter signaled by a filter syntax element depending on one or more fractional parts (or multiple fractional parts) of the motion vector (e.g., a first filter signaled by a filter syntax element is selected for filtering in a first direction, e.g. when considering position coordinates in the first direction, e.g. if one or more fractional parts (or multiple fractional parts) of the motion vector are between full samples, and a second filter signaled by a filter syntax element is selected for filtering in a second direction different from the first direction, e.g. when considering position coordinates in the second direction, e.g. if all fractional parts (or multiple fractional parts) of the motion vector are on full samples).
[0086] According to one embodiment, the decoder is configured to take over the filter selection from a neighboring block in merge mode (eg, selectively depending on whether the motion vector is identical to the motion vector of the neighboring block).
[0087] According to one embodiment, at the decoder, the selected (eg, applied) processing scheme defines (eg, spatial) filtering of the luma component and / or (eg, spatial) filtering of the chroma components. According to one embodiment, the decoder is configured to select (eg, apply) a filter depending on the prediction mode.
[0088] According to one embodiment, the decoder is configured to adapt mapping rules that map the encoded index values to instructions (eg, index values) that indicate a processing scheme according to one or more syntax elements.
[0089] According to one embodiment, the decoder is configured to adapt the number of bins used to identify the coded index indicative of the selected processing scheme depending on one or more syntax elements.
[0090] According to one embodiment, the decoder is configured to determine a set of processing schemes that are permissible in terms of the one or more syntax elements, where a processing scheme to be applied is identified based on the one or more syntax elements and an index.
[0091] According to one embodiment, the decoder is configured to determine whether a single processing scheme is acceptable for a portion of the video sequence, and, in response to a finding that only a single processing scheme (e.g., a spatial filter or an interpolation filter) is acceptable for a portion of the video sequence, selectively omit identifying the coded index.
[0092] According to one embodiment, the decoder is configured to select a mapping rule that maps the coded index value to an indication (e.g., index value) that indicates a processing scheme such that the number of bins representing the coded index value matches the number of processing schemes allowable for a portion of the video sequence.
[0093] According to one embodiment, the decoder is configured to select (e.g., apply) different processing schemes (e.g., different interpolation filters) for different blocks (e.g., different prediction units PU and / or coding units CU and / or coding tree units CTU) within a video frame or picture (different interpolation filters for the same fractional sample positions in different parts of a video sequence or different parts of a single video frame or “picture”).
[0094] According to one embodiment, the decoder is configured to decode (e.g., in response to one or more syntax elements) an index indicating a selected processing scheme, such that a given coded index value represents (e.g., is associated with) a different processing scheme depending on the motion vector precision (which may, for example, be signaled using a syntax element in the bitstream and may, for example, take one of the following values: QPEL, HPEL, FPEL, 4PEL).
[0095] According to one embodiment, in the decoder, the one or more syntax elements include at least one of: motion vector precision (e.g., quarter sample, half sample, full sample), fractional sample position, block size, block shape, number of prediction hypotheses (e.g., one hypothesis for uni-prediction, two hypotheses for bi-prediction), prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra), availability of the coded residual signal (e.g., coded block flag), spectral characteristics of the coded residual signal, reference picture signal (e.g., reference block defined by motion data, block edges from co-located prediction units PU within the reference block, high or low frequency characteristics of the reference signal), loop filter data (e.g., edge offset or band offset classification from sample adaptive offset filter SAO, deblocking filter decision and boundary strength), motion vector length (e.g., enable only long motion vectors or additional smoothing filters for specific directions), or adaptive motion vector resolution mode.
[0096] According to one embodiment, the decoder is configured to decode (e.g., in response to one or more syntax elements) an index indicating a selected processing scheme, such that a given coded index value represents (e.g., is associated with) different processing schemes depending on the fractional part (or multiple fractional parts) of a motion vector (e.g., different coding schemes are used depending on whether all motion vectors of the considered part point to integer sample positions, e.g., luma sample positions, or not).
[0097] According to one embodiment, the decoder is configured to selectively determine a set of processing schemes (e.g., "multiple possibility coding") for a motion vector precision (e.g., sub-sample motion vector precision, e.g., HPEL) between a maximum motion vector precision (e.g., a sub-sample motion vector resolution finer than the aforementioned motion vector precision, e.g., QPEL) and a minimum motion vector precision (e.g., 4PEL), or for a motion vector precision (e.g., HPEL) between the maximum motion vector precision (e.g., QPEL) and a full-sample motion vector precision (e.g., FPEL) for which a processing scheme to be applied is identified.
[0098] According to one embodiment, the decoder is configured to decode an index (processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering), a second FIR filtering, and a third FIR.
[0099] According to one embodiment, the decoder is configured to decode an index (a processing scheme index, if_idx) for selecting between a first FIR filtering (e.g., HEVC filtering) (e.g., HEVC 8-tap filtering) and a second FIR filtering.
[0100] According to one embodiment, the decoder is configured to apply processing schemes (eg, interpolation filters) having different characteristics (eg, stronger vs. weaker low-pass characteristics).
[0101] According to one embodiment, the decoder is configured to select a mapping rule depending on the motion vector precision and the fractional sample position or fractional part (or multiple fractional parts) of the motion vector. According to one embodiment, the decoder is configured to determine a set of available processing schemes (eg, interpolation filters) depending on the motion vector precision.
[0102] According to one embodiment, the decoder is configured to apply one of a quarter-sample motion vector resolution, a half-sample motion vector resolution, a full-sample motion vector resolution, and a four-sample motion vector resolution (e.g., index elements describing the processing scheme or interpolation filters are selectively included in the bitstream when half-sample motion vector resolution is selected, and index elements describing the processing scheme are, for example, omitted in other cases). According to one embodiment, the processing scheme comprises one or more FIR filters.
[0103] Another aspect of the disclosure relates to a method of hybrid video coding (e.g., video coding with prediction and transform coding of prediction error) for providing an encoded representation of a video sequence based on input video content, the method including (e.g., a method performed by an encoder) determining one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., a block) of the video sequence, and determining (e.g., by a decoder) a processing scheme (e.g., a motion vector precision syntax element) to be applied to the portion of the video sequence based on characteristics described by the one or more syntax elements (e.g., decoder settings signaled by the one or more syntax elements, or characteristics of the portion of the video sequence signaled by the one or more syntax elements). the processing scheme is for obtaining samples (e.g., samples, e.g., fractional samples, associated with the video sequence) for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence, encoding an index indicating the selected processing scheme (in dependence on the one or more syntax elements) such that a given coded index value represents (e.g., is associated with) a different processing scheme depending on characteristics described by the one or more syntax elements (e.g., depending on the one or more syntax elements), and providing (e.g., transmitting, uploading, storing as a file) a bitstream including the one or more syntax elements and the coded index as a coded representation of the video sequence.
[0104] Another aspect of the disclosure relates to a method of hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), the method providing an encoded representation of a video sequence based on input video content, the method (e.g., a method performed by an encoder) including: determining one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., a block) of the video sequence; and selecting (e.g., by a decoder) a processing scheme (e.g., an interpolation filter) to be applied to the portion of the video sequence based on characteristics described by the one or more syntax elements (e.g., decoder settings signaled by the one or more syntax elements, or characteristics of the portion of the video sequence signaled by the one or more syntax elements), the processing scheme selecting (e.g., by a decoder) a processing scheme (e.g., an interpolation filter) to be applied to the portion of the video sequence based on integer and / or fractional motion vectors within the portion of the video sequence. The method includes obtaining samples (e.g., samples, e.g., fractional samples) for motion compensated prediction at a few positions (e.g., associated with a video sequence), selecting an entropy coding scheme to be used to provide an encoded index indicative of the selected processing scheme in accordance with characteristics described by one or more syntax elements (e.g., in accordance with one or more syntax elements) (e.g., the number of bins described by the entropy coding scheme may be equal to zero if only one processing scheme is acceptable from the point of view of one or more considered syntax elements), and providing (e.g., transmitting, uploading, saving as a file) a bitstream including the one or more syntax elements and including the encoded index (e.g., if the selected number of bins is greater than zero) as a coded representation of the video sequence.
[0105] Another aspect of the disclosure relates to a method of hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), for providing output video content based on a coded representation of a video sequence, the method (e.g., a method performed by a decoder), comprising obtaining (e.g., receiving, downloading, opening from a file) a bitstream as the coded representation of the video sequence, identifying from the bitstream one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., block) (and coded index) of the video sequence, and The method includes identifying a processing scheme (e.g., an interpolation filter selected by an encoder) for a portion of the video sequence based on an index and an index (which may be included in a bitstream in encoded form), the processing scheme being for obtaining samples for motion compensated prediction (e.g., samples, e.g., fractional samples, for the video sequence) at integer and / or fractional positions within the portion of the video sequence, the method including assigning different processing schemes to a given coded index value in response to one or more syntax elements (e.g., in response to values of the one or more syntax elements) and applying the identified processing scheme to the portion of the video sequence.
[0106] Another aspect of the disclosure relates to a method of hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), for providing output video content based on a coded representation of a video sequence, the method (e.g., a method performed by a decoder), comprising obtaining (e.g., receiving, downloading, opening from a file) a bitstream as the coded representation of the video sequence, identifying from the bitstream one or more syntax elements (e.g., motion vector precision syntax elements) associated with a portion (e.g., a block) of the video sequence, and, responsive to the one or more syntax elements, generating an interpolation signal for the bitstream to provide a coded index indicating a processing scheme (e.g., an interpolation filter, e.g., selected by an encoder) to be used for the portion of the video sequence. determining an entropy coding scheme, the processing scheme for obtaining samples for motion compensated prediction (e.g., samples, e.g., fractional samples, for the video sequence) at integer and / or fractional positions within a portion of the video sequence, comprising: identifying an index encoded within the bitstream if the determined entropy coding scheme specifies a number of bins greater than zero; decoding the index and identifying (e.g., selecting) a processing scheme based on one or more syntax elements and, if the number of bins specified by the determined entropy coding scheme is greater than zero, based on the decoded index (or identifying a processing scheme based on the one or more syntax elements only if the determined number of bins is zero); and applying the identified processing scheme to the portion of the video sequence.
[0107] Another aspect of the disclosure relates to a video bitstream for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), where the video bitstream describes a video sequence (e.g., in the form of a coded representation of the video sequence), where the video bitstream includes one or more syntax elements (e.g., motion vector precision syntax elements) associated with portions (e.g., blocks) of the video sequence, where the video bitstream includes a processing scheme index bitstream element (e.g., if_idx, e.g., describing an interpolation filter used by a video decoder) for at least some of the portions of the video sequence, where the presence of the processing scheme index bitstream element (e.g., describing a filtering or interpolation filter performed by the video decoder) varies depending on one or more other syntax elements (e.g., the processing scheme index bitstream element is present for a first value of another syntax element or a first combination of values of a plurality of other syntax elements and e.g., the processing scheme index bitstream element is omitted for a second value of another syntax element or a second combination of values of a plurality of other syntax elements).
[0108] Another aspect of the disclosure relates to a video bitstream for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), where the video bitstream describes a video sequence (e.g., in the form of a coded representation of the video sequence), where the video bitstream includes one or more syntax elements (e.g., motion vector precision syntax elements) associated with portions (e.g., blocks) of the video sequence, where the video bitstream includes processing scheme index bitstream elements (e.g., if_idx, e.g., describing an interpolation filter used by a video decoder) for at least some of the portions of the video sequence, where a number of bins of the processing scheme index bitstream element varies depending on the one or more syntax elements associated with the portions of the video sequence (e.g., the number of bins of the processing scheme index bitstream element is adapted to a number of processing schemes allowable for the portion of the video sequence, e.g., the number of bins of the processing scheme index bitstream element is equal to zero, e.g., if only one processing scheme is allowable for the portion of the video sequence, the processing scheme index bitstream element is omitted).
[0109] Another aspect of the disclosure relates to a video bitstream for hybrid video coding (e.g., video coding with prediction and transform coding of prediction error), where the video bitstream describes a video sequence (e.g., in the form of a coded representation of the video sequence), the video bitstream includes one or more syntax elements (e.g., motion vector precision syntax elements) associated with portions (e.g., blocks) of the video sequence, the video bitstream includes processing scheme index bitstream elements (e.g., if_idx, e.g., that describe an interpolation filter used by a video decoder) for at least some of the portions of the video sequence, and the meaning of a codeword (e.g., coded index value) of the processing scheme index bitstream element varies depending on one or more syntax elements associated with the portions of the video sequence (e.g., such that different processing schemes are assigned to a given coded index value depending on one or more syntax elements associated with the portions of the video sequence).
[0110] Another aspect of the disclosure relates to a hybrid video coding decoder, the decoder for providing output video content based on a coded representation of a video sequence, the decoder being configured to obtain a bitstream as the coded representation of the video sequence, identify from the bitstream one or more syntax elements associated with a portion of the video sequence, select a processing scheme for the portion of the video sequence in response to the one or more syntax elements, the processing scheme being fully specified by characteristics of the portion and a fractional portion of a motion vector, the characteristics being described by the one or more syntax elements, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence, and apply the identified processing scheme to the portion of the video sequence.
[0111] Since the decoder can select a processing scheme based on one or more syntax elements, the decoder can select a processing scheme for a portion without requiring a syntax element specifically dedicated to the processing scheme. Thus, the decoder can select a processing scheme even if the processing scheme is not explicitly indicated in the bitstream, which can result in a particularly small bitstream. When the decoder uses one or more syntax elements to obtain a processing scheme, the decoder can select a processing scheme with the granularity provided by the one or more syntax elements. Thus, the decoder can efficiently use information already provided in the bitstream to select a processing scheme with fine granularity without requiring additional bits in the bitstream. The fractional part of the motion vector can characterize the processing scheme used to obtain samples particularly accurately, since the motion vector can be used for motion compensation prediction. Thus, a processing scheme can be quickly and reliably selected based on the fractional part of the motion vector.
[0112] According to one embodiment, the characteristics of the part that specify the processing scheme relate to one or two or all of the following: the motion vector precision, the fractional position of the sample (obtained by the processing scheme), and the prediction mode used for the part. Since the processing scheme may be for obtaining samples for motion compensation prediction, the MV precision may provide a precise hint on the appropriate processing scheme to be used for the part. Since different processing schemes can be used for full sample, half sample, and quarter sample positions, the fractional position of the sample may limit the number of allowable processing schemes or even specify the processing scheme. One or more or all of the motion vector precision, the fractional position of the sample, and the prediction mode may be indicated by one or more syntax elements for decoding the part of the video sequence. Thus, the decoder main infers the processing scheme from already available information.
[0113] According to one embodiment, the parts are prediction units (PUs) or coding units (CUs), and the decoder is configured to select a processing scheme for different parts of the video sequence separately. Selecting a processing scheme for different prediction units or coding units separately allows the processing scheme to be adapted to local characteristics of the video sequence with a particularly fine granularity. This precise adaptation may allow decoding of parts from a bitstream of particularly small size and / or allow particularly fast decoding. Since the processing scheme is selected according to one or more syntax elements, it is possible to determine the processing scheme for different prediction units or coding units separately without increasing the size of the bitstream. According to one embodiment, the decoder is configured to use different processing schemes for the same fractional sample positions in different parts of a video sequence. According to one embodiment, the decoder is configured to switch processing schemes at the PU or CU level to adapt to local image characteristics. According to one embodiment, none of the one or more syntax elements are explicitly dedicated to indicating a processing scheme.
[0114] Another aspect of the disclosure relates to a hybrid video coding decoder, for providing output video content based on a coded representation of a video sequence, the decoder being configured to obtain a bitstream as the coded representation of the video sequence, identify from the bitstream one or more syntax elements associated with a portion of the video sequence, select a processing scheme for the portion of the video sequence according to a motion vector precision, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence, and apply the identified processing scheme to the portion of the video sequence.
[0115] According to one embodiment, the decoder is configured to select the processing scheme depending on the motion vector precision and the fractional position of the sample (obtained by the processing scheme). According to one embodiment, the decoder is configured to select the processing scheme further depending on the prediction mode used for the portion.
[0116] According to one embodiment, the parts are prediction units (PUs) or coding units (CUs), and the decoder is configured to select a processing scheme for different parts of the video sequence separately. Selecting a processing scheme for different prediction units or coding units separately allows the processing scheme to be adapted to local characteristics of the video sequence with a particularly fine granularity. This precise adaptation may allow decoding of parts from a bitstream of particularly small size and / or allow particularly fast decoding. Since the processing scheme is selected according to one or more syntax elements, it is possible to determine the processing scheme for different prediction units or coding units separately without increasing the size of the bitstream. According to one embodiment, the decoder is configured to use different processing schemes for the same fractional sample positions in different parts of a video sequence. According to one embodiment, the decoder is configured to switch processing schemes at the PU or CU level to adapt to local image characteristics. According to one embodiment, none of the one or more syntax elements are explicitly dedicated to indicating a processing scheme.
[0117] Another aspect of the disclosure relates to a hybrid video coding decoder, the decoder for providing output video content based on a coded representation of a video sequence, the decoder being configured to obtain a bitstream as the coded representation of the video sequence, identify from the bitstream one or more syntax elements associated with a portion of the video sequence, determine a processing scheme for the portion of the video sequence in response to the one or more syntax elements, with PU or CU granularity, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence, and apply the identified processing scheme to the portion of the video sequence. According to one embodiment, the decoder is configured to use different processing schemes for the same fractional sample positions in different parts of a video sequence. According to one embodiment, the decoder is configured to switch processing schemes at the PU or CU level to adapt to local image characteristics. According to one embodiment, none of the one or more syntax elements are explicitly dedicated to indicating a processing scheme.
[0118] According to one embodiment, at the decoder, the processing scheme is a filter (eg, an interpolation filter) or a set of filters (eg, interpolation filters) (eg, parameterized by the fractional part of the motion vector).
[0119] According to one embodiment, at the decoder, the filters in the set of filters are separably applicable to portions of the video sequence (eg, one-dimensional filters that can be applied sequentially in spatial directions, such as the x-direction and / or the y-direction).
[0120] According to one embodiment, the decoder is configured to select (e.g., apply) different filters (e.g., interpolation filters) for filtering in a first direction (e.g., x) and for filtering in a second direction (e.g., y) depending on a fractional portion (or multiple fractional portions) of the motion vector.
[0121] According to one embodiment, the decoder is configured to select (e.g., apply) different filters (e.g., interpolation filters) for filtering in a first direction and for filtering in a second direction, such that stronger low-pass filtering is applied to filtering in directions in which (one or more or all) fractional portions (or multiple fractional portions) of the motion vector lie on a full sample (e.g., when half-sample accurate motion vector(s) point to full sample position(s)) (compared to filtering applied in directions in which fractional portions (or multiple portions) of the motion vector lie between full samples).
[0122] According to one embodiment, the decoder is configured to take over the filter selection from a neighboring block in merge mode (eg, selectively depending on whether the motion vector is identical to the motion vector of the neighboring block).
[0123] According to one embodiment, at the decoder, the selected (eg, applied) processing scheme defines (eg, spatial) filtering of the luma component and / or (eg, spatial) filtering of the chroma components.
[0124] According to one embodiment, the decoder is configured to select (e.g., apply) different processing schemes (e.g., different interpolation filters) for different blocks (e.g., different prediction units PU and / or coding units CU and / or coding tree units CTU) within a video frame or picture (different interpolation filters for the same fractional sample positions in different parts of a video sequence or different parts of a single video frame or “picture”).
[0125] According to one embodiment, the one or more syntax elements include at least one of: motion vector precision (e.g., quarter sample, half sample, full sample), fractional sample position, block size, block shape, number of prediction hypotheses (e.g., one hypothesis for uni-prediction, two hypotheses for bi-prediction), prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra), availability of the coded residual signal (e.g., coded block flag), spectral characteristics of the coded residual signal, reference picture signal (e.g., reference block defined by motion data, block edges from co-located prediction units PU within the reference block, high or low frequency characteristics of the reference signal), loop filter data (e.g., edge offset or band offset classification from sample adaptive offset filter SAO, deblocking filter decision and boundary strength), motion vector length (e.g., enable only long motion vectors or additional smoothing filters for specific directions), or adaptive motion vector resolution mode.
[0126] According to one embodiment, the decoder is configured to selectively determine a set of processing schemes (e.g., "multiple possibility coding") for a motion vector precision (e.g., sub-sample motion vector precision, e.g., HPEL) between a maximum motion vector precision (e.g., a sub-sample motion vector resolution finer than the aforementioned motion vector precision, e.g., QPEL) and a minimum motion vector precision (e.g., 4PEL), or for a motion vector precision (e.g., HPEL) between the maximum motion vector precision (e.g., QPEL) and a full-sample motion vector precision (e.g., FPEL) for which a processing scheme to be applied is identified.
[0127] According to one embodiment, the decoder is configured to apply processing schemes (eg, interpolation filters) having different characteristics (eg, stronger vs. weaker low-pass characteristics).
[0128] According to one embodiment, the decoder is configured to apply one of a quarter sample motion vector resolution, a half sample motion vector resolution, a full sample motion vector resolution, or a four sample motion vector resolution. According to one embodiment, in the decoder the processing scheme comprises one or more FIR filters.
[0129] Another aspect of the disclosure relates to an encoder for hybrid video coding, the encoder for providing a coded representation of a video sequence based on input video content, the encoder being configured to determine one or more syntax elements associated with a portion of the video sequence, select a processing scheme to apply to the portion of the video sequence based on characteristics described by the one or more syntax elements, the processing scheme being fully specified by the characteristics of the portion and a fractional portion of a motion vector, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence, and provide a bitstream including the one or more syntax elements as the coded representation of the video sequence.
[0130] According to one embodiment, the characteristics described by the one or more syntax elements relate to one or two or all of the following: the motion vector precision, the fractional position of the sample (obtained by the processing scheme), and the prediction mode used for that portion.
[0131] According to one embodiment, the portions are prediction units (PUs) or coding units (CUs), and the encoder is configured to select the processing schemes separately for different portions of the video sequence. According to one embodiment, the encoder is configured to use different processing schemes for the same fractional sample positions in different parts of a video sequence. According to one embodiment, the encoder is configured to switch processing schemes at the PU or CU level to adapt to local image characteristics. According to one embodiment, none of the one or more syntax elements are explicitly dedicated to indicating a processing scheme.
[0132] Another aspect of the disclosure relates to an encoder for hybrid video coding, for providing an encoded representation of a video sequence based on input video content, the encoder being configured to determine one or more syntax elements associated with a portion of the video sequence, select a processing scheme to apply to the portion of the video sequence based on a motion vector precision described by the one or more syntax elements, the processing scheme being specified entirely by characteristics of the portion and a fractional portion of the motion vector, and provide a bitstream including the one or more syntax elements as the encoded representation of the video sequence.
[0133] According to one embodiment, the encoder is configured to select a processing scheme based on the motion vector precision and the fractional position of the sample (obtained by the processing scheme). According to one embodiment, the encoder is configured to select the processing scheme further depending on the prediction mode used for the portion.
[0134] According to one embodiment, the portions are prediction units (PUs) or coding units (CUs), and the encoder is configured to select the processing schemes separately for different portions of the video sequence. According to one embodiment, the encoder is configured to use different processing schemes for the same fractional sample positions in different parts of a video sequence. According to one embodiment, the encoder is configured to switch processing schemes at the PU or CU level to adapt to local image characteristics. According to one embodiment, none of the one or more syntax elements are explicitly dedicated to indicating a processing scheme.
[0135] Another aspect of the disclosure relates to an encoder for hybrid video coding, for providing an encoded representation of a video sequence based on input video content, the encoder being configured to determine one or more syntax elements associated with a portion of the video sequence, select a processing scheme to be applied to the portion of the video sequence based on a motion vector precision described by the one or more syntax elements, with PU granularity or CU granularity, the processing scheme being fully specified by characteristics of the portion and fractional parts of the motion vectors, and provide a bitstream including the one or more syntax elements as the encoded representation of the video sequence. According to one embodiment, the encoder is configured to use different processing schemes for the same fractional sample positions in different parts of a video sequence. According to one embodiment, the encoder is configured to switch processing schemes at the PU or CU level to adapt to local image characteristics. According to one embodiment, none of the one or more syntax elements are explicitly dedicated to indicating a processing scheme.
[0136] According to one embodiment, the encoder is configured to determine a set of processing schemes that are permissible in terms of the characteristics described by the one or more syntax elements, and to select a processing scheme from the determined set.
[0137] According to one embodiment, at the encoder, the processing scheme is a filter (eg, an interpolation filter) or a set of filters (eg, interpolation filters) (eg, parameterized by the fractional part of the motion vector).
[0138] According to one embodiment, in an encoder, the filters in the set of filters are separably applicable to portions of a video sequence (eg, one-dimensional filters that can be applied sequentially in a spatial direction, eg, the x-direction and / or the y-direction).
[0139] According to one embodiment, the encoder is configured to select different processing schemes (e.g., different interpolation filters) for different blocks (e.g., different prediction units PU and / or coding units CU and / or coding tree units CTU) within a video frame or picture (different interpolation filters for the same fractional sample positions in different parts of a video sequence or different parts of a single video frame or picture).
[0140] According to one embodiment, in the encoder, the one or more syntax elements include at least one of: motion vector precision (e.g., quarter sample, half sample, full sample), fractional sample position, block size, block shape, number of prediction hypotheses (e.g., one hypothesis for uni-prediction, two hypotheses for bi-prediction), prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra), availability of the coded residual signal (e.g., coded block flag), spectral characteristics of the coded residual signal, reference picture signal (e.g., reference block defined by motion data, block edges from co-located prediction units PU within the reference block, high or low frequency characteristics of the reference signal), loop filter data (e.g., edge offset or band offset classification from sample adaptive offset filter SAO, deblocking filter decision and boundary strength), motion vector length (e.g., enable only long motion vectors or additional smoothing filters for specific directions), or adaptive motion vector resolution mode.
[0141] According to one embodiment, the encoder is configured to selectively determine a set of processing schemes (e.g., "multiple possibility encoding") for motion vector precision (e.g., sub-sample motion vector precision, e.g., HPEL) between a maximum motion vector precision (e.g., a sub-sample motion vector resolution finer than the aforementioned motion vector precision, e.g., QPEL) and a minimum motion vector precision (e.g., 4PEL), or for motion vector precision (e.g., HPEL) between the maximum motion vector precision (e.g., QPEL) and a full-sample motion vector precision (e.g., FPEL), and select a processing scheme from the determined set.
[0142] Another aspect of the disclosure relates to a method for hybrid video coding, for providing output video content based on a coded representation of a video sequence, the method including: obtaining a bitstream as the coded representation of the video sequence; identifying from the bitstream one or more syntax elements associated with a portion of the video sequence; selecting a processing scheme for the portion of the video sequence in response to the one or more syntax elements, the processing scheme being fully specified by characteristics of the portion and fractional portions of motion vectors, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence; and applying the identified processing scheme to the portion of the video sequence.
[0143] According to one embodiment, the encoder is configured to select between quarter-sample motion vector resolution, half-sample motion vector resolution, full-sample motion vector resolution, and four-sample motion vector resolution.
[0144] According to one embodiment, in the encoder, the processing scheme comprises one or more FIR filters, which are particularly stable filters for processing video sequences.
[0145] Another aspect of the present disclosure relates to a method for hybrid video coding, for providing output video content based on an encoded representation of a video sequence, the method including: obtaining a bitstream as the encoded representation of the video sequence; identifying from the bitstream one or more syntax elements associated with a portion of the video sequence; selecting a processing scheme for the portion of the video sequence according to a motion vector precision, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence; and applying the identified processing scheme to the portion of the video sequence.
[0146] Another aspect of the present disclosure relates to a method for hybrid video coding, for providing output video content based on an encoded representation of a video sequence, the method including: obtaining a bitstream as the encoded representation of the video sequence; identifying from the bitstream one or more syntax elements associated with a portion of the video sequence; determining a processing scheme for the portion of the video sequence in response to the one or more syntax elements, with PU or CU granularity, where the processing scheme is for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence; and applying the identified processing scheme to the portion of the video sequence.
[0147] Another aspect of the disclosure relates to a method for hybrid video coding, for providing a coded representation of a video sequence based on input video content, the method including: determining one or more syntax elements associated with a portion of the video sequence; selecting a processing scheme to be applied to the portion of the video sequence based on characteristics described by the one or more syntax elements, the processing scheme being fully specified by the characteristics of the portion and a fractional portion of a motion vector, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence; and providing a bitstream including the one or more syntax elements as the coded representation of the video sequence.
[0148] Another aspect of the disclosure relates to a method for hybrid video coding for providing an encoded representation of a video sequence based on input video content, the method including: determining one or more syntax elements associated with a portion of the video sequence; selecting a processing scheme to be applied to the portion of the video sequence based on a motion vector precision described by the one or more syntax elements, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence; and providing a bitstream including the one or more syntax elements as the encoded representation of the video sequence.
[0149] Another aspect of the disclosure relates to a method for hybrid video coding for providing an encoded representation of a video sequence based on input video content, the method including: determining one or more syntax elements associated with a portion of the video sequence; selecting, with PU or CU granularity, a processing scheme to be applied to the portion of the video sequence based on characteristics described by the one or more syntax elements, the processing scheme being for obtaining samples for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence; and providing a bitstream including the one or more syntax elements as the encoded representation of the video sequence.
[0150] Another aspect of the disclosure relates to a computer program for performing any of the described methods for hybrid video coding when the computer program runs on a computer.
[0151] The described decoder provides similar features and advantages as those described with respect to the encoder, and vice versa, and features and advantages of the encoder that refer to features of the encoder apply equally to corresponding features of the decoder, with decoding replacing encoding and vice versa. For example, if features of the encoder enable encoding of information in a particularly small size bitstream or efficient encoding, corresponding features of the decoder may enable deriving information from the small size bitstream or efficient decoding, respectively. In other words, any of the features, functions, and details disclosed herein with respect to the encoder may be optionally included in the decoder, both individually and in combination, and vice versa.
[0152] The described method for hybrid video coding relies on the same ideas as the encoders and decoders described above and provides the same or comparable functions and advantages. The method may optionally be combined (or supplemented) with any of the features, functions, and details described herein with respect to the corresponding encoders and decoders. The method for hybrid video coding may optionally be combined with the mentioned features, functions, and details, either individually or in any combination thereof. [Brief description of the drawings]
[0153] Embodiments of the present disclosure are described herein with reference to the accompanying drawings and figures. [Figure 1] FIG. 1 is a schematic diagram of an encoder capable of implementing the disclosed concepts; [Diagram 2] FIG. 1 is a schematic diagram of a decoder capable of implementing the disclosed concepts; [Diagram 3] FIG. 2 illustrates signals used by an encoder or decoder according to one embodiment. [Figure 4] FIG. 2 illustrates fractional sample positions that can be used for motion compensated prediction according to one embodiment. [Diagram 5] FIG. 2 illustrates an encoder according to one embodiment. [Figure 6] FIG. 4 illustrates an encoder according to another embodiment. [Figure 7] FIG. 2 illustrates a decoder according to one embodiment. [Figure 8] FIG. 4 shows a decoder according to another embodiment. [Figure 9] 1 is a flow chart of a method for hybrid video coding according to one embodiment. [Figure 10] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 11] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 12]4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 13] FIG. 4 illustrates an encoder according to another embodiment. [Figure 14] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 15] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 16] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 17] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 18] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. [Figure 19] 4 shows a flowchart of a method for hybrid video coding according to another embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0154] In the following, different invention embodiments and aspects are described. Further embodiments are defined by the appended claims.
[0155] It is noted that any embodiment defined by a claim can be supplemented by any of the details (features and functions) described in this specification. Furthermore, the embodiments described in this specification can be used individually or can be supplemented by any of the features contained in the claims.
[0156] It should also be noted that the individual aspects described herein can be used individually or in combination, and thus details can be added to each individual aspect without adding the details to another one of the aspects.
[0157] It should also be noted that this disclosure explicitly or implicitly describes functionality available in a video encoder (apparatus for providing an encoded representation of an input video signal) and a video decoder (apparatus for providing a decoded representation of a video signal based on the encoded representation). Thus, any of the features described herein may be used in the context of a video encoder as well as in the context of a video decoder.
[0158] Moreover, features and functions disclosed herein in relation to a method may also be used in an apparatus (configured to perform such functions). Moreover, any features and functions disclosed herein in relation to an apparatus may also be used in a corresponding method. In other words, the methods disclosed herein may be complemented by any of the features and functions described in relation to the apparatus.
[0159] Additionally, any of the features and functionality described herein can be implemented in hardware or software, or using a combination of hardware and software, as described in other sections, such as "Implementation Alternatives."
[0160] Furthermore, any of the features and syntax elements described herein may be optionally introduced into a video bitstream, both individually and in any combination.
[0161] In the following description, the embodiments are described in detail, but it should be understood that the embodiments provide many applicable concepts that can be incorporated into a wide variety of video encoders and video decoders. The specific embodiments described are merely illustrative of a particular way to implement and use the concepts, and do not limit the scope of the embodiments. In the following description of the embodiments, the same or similar elements or elements having the same functions are given the same reference numbers or identified with the same names, and repeated descriptions of elements given the same reference numbers or identified with the same names are usually omitted. Therefore, descriptions provided for elements having the same or similar reference numbers or identified with the same names may be interchangeable or may be applied to each other in different embodiments. In the following description, a number of details are presented to provide a more complete description of the embodiments of the present disclosure. However, it will be apparent to one skilled in the art that other embodiments can be implemented without these specific details. In other examples, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the examples described herein. Moreover, the features of different embodiments described herein may be combined with each other unless otherwise specified.
[0162] 1. Example of a coding framework The following description of the figures begins with the presentation of an encoder and decoder of a block-based predictive codec for coding pictures of a video to form an example of a coding framework in which embodiments of the present invention can be incorporated. The respective encoders and decoders are described with respect to Figures 1-3. Below, a description of embodiments of the inventive concepts is presented together with a description of how such concepts can be incorporated into the encoders and decoders of Figures 1 and 2, respectively, although the embodiments described in subsequent Figures 4 onwards may also be used to form encoders and decoders that do not operate according to the coding framework underlying the encoders and decoders of Figures 1 and 2.
[0163] FIG. 1 shows an apparatus for predictively coding a picture 12 into a data stream 14, exemplarily using transform-based residual coding. The apparatus or encoder is indicated using the reference number 10. FIG. 2 shows a corresponding decoder 20, i.e. an apparatus 20 configured to predictively decode a picture 12' from the data stream 14, also using transform-based residual decoding, the apostrophe being used to indicate that the picture 12' reconstructed by the decoder 20 deviates from the picture 12 originally encoded by the apparatus 10 in terms of coding losses introduced by quantization of the predictive residual signal. Although FIG. 1 and FIG. 2 exemplarily use transform-based predictive residual coding, the embodiments of the present application are not limited to this type of predictive residual coding. This also applies to other details described with respect to FIG. 1 and FIG. 2, as outlined below.
[0164] The encoder 10 is configured to spatially-to-spectral transform the prediction residual signal and to encode the prediction residual signal thus obtained into a data stream 14. Similarly, the decoder 20 is configured to decode the prediction residual signal from the data stream 14 and to spectrally-to-spatial transform the prediction residual signal thus obtained.
[0165] Internally, the encoder 10 may comprise a prediction residual signal former 22 which generates a prediction residual 24 to measure the deviation of a prediction signal 26 from an original signal, i.e. the picture 12. The prediction residual signal former 22 may for example be a subtractor which subtracts the prediction signal from the original signal, i.e. from the picture 12. The encoder 10 then further comprises a transformer 28 which performs a spatial-spectral transformation of the prediction residual signal 24 in order to obtain a spectral domain prediction residual signal 24' which is quantized by a quantizer 32 also included in the encoder 10. The prediction residual signal 24'' thus quantized is coded into the bitstream 14. For this purpose, the encoder 10 may optionally comprise an entropy coder 34 which entropy codes the prediction residual signal which is transformed and quantized into the data stream 14. The prediction signal 26 is generated by a prediction stage 36 of the encoder 10 on the basis of the prediction residual signal 24'' which is coded into the data stream 14 and decodable therefrom. For this purpose, the prediction stage 36 may internally comprise an inverse quantizer 38 for inverse quantizing the prediction residual signal 24'' so as to obtain a spectral domain prediction residual signal 24''' corresponding to the signal 24' but without the quantization losses, as shown in FIG. 1, and an inverse transformer 40 for inversely transforming, i.e. spectral-spatial transforming, the latter prediction residual signal 24''' to obtain a prediction residual signal 24'''' corresponding to the original prediction residual signal 24 but without the quantization losses. A combiner 42 of the prediction stage 36 then recombines the prediction signal 26 and the prediction residual signal 24'''', for example by addition, to obtain a reconstructed signal 46, i.e. a reconstruction of the original signal 12. The reconstructed signal 46 may correspond to the signal 12'. A prediction module 44 of the prediction stage 36 then generates a prediction signal 26 based on the signal 46, for example using a spatial prediction, i.e. an intra-picture prediction, and / or a temporal prediction, i.e. an inter-picture prediction.
[0166] Similarly, as shown in Fig. 2, the decoder 20 may be internally configured with components corresponding to the prediction stage 36 and interconnected in a manner corresponding to the prediction stage. In particular, an entropy decoder 50 of the decoder 20 may entropy decode a quantized spectral domain prediction residual signal 24'' from the data stream, with an inverse quantizer 52, an inverse transformer 54, a combiner 56 and a prediction module 58 interconnected and cooperating in the manner described above with respect to the modules of the prediction stage 36 to recover a reconstructed signal based on the prediction residual signal 24'', such that an output of the combiner 56 provides a reconstructed signal, i.e., picture 12', as shown in Fig. 2.
[0167] Although not specifically described above, it is readily apparent that the encoder 10 may set some encoding parameters, including, for example, prediction modes, motion parameters, etc., according to some optimization schemes, such as, for example, a method for optimizing some rate- and distortion-related criteria, i.e., coding cost. For example, the encoder 10 and the decoder 20 and corresponding modules 44, 58 may each support different prediction modes, such as intra-coded and inter-coded modes. The granularity at which the encoder and the decoder switch between these prediction mode types may correspond to a subdivision of the pictures 12 and 12', respectively, into coded segments or coded blocks. At the unit of these coded segments, for example, the picture may be subdivided into blocks that are intra-coded and blocks that are inter-coded. The intra-coded blocks are predicted based on the spatial already coded / decoded neighborhood of the respective blocks, as outlined in more detail below. Several intra-coding modes may exist and be selected for each intra-coded segment, including directional or angular intra-coding modes in which each segment is filled into each intra-coded segment by extrapolating neighboring sample values along a particular direction specific to each directional intra-coding mode. The intra-coding modes may also include one or more further modes, such as, for example, a DC coding mode in which the prediction of each intra-coded block assigns a DC value to all samples in the respective intra-coded segment, and / or a planar intra-coding mode in which the prediction of each block is approximated or determined based on neighboring samples to be a spatial distribution of sample values described by a two-dimensional linear function over the sample positions of the respective intra-coded block with a planar tilt drive and offset defined by the two-dimensional linear function. In comparison, inter-coded blocks may be predicted, for example, temporally.For inter-coded blocks, motion vectors may be signaled in the data stream, which indicate the spatial displacement of parts of previously coded pictures of the video to which picture 12 belongs, and which previously coded / decoded pictures are sampled to obtain a prediction signal for the respective inter-coded block. This means that in addition to the residual signal coding contained in data stream 14, such as entropy-coded transform coefficient levels representing the quantized spectral domain prediction residual signal 24'', data stream 14 may encode optional further parameters, such as coding mode parameters for assigning coding modes to the various blocks, some prediction parameters of the blocks, such as motion parameters of the inter-coded segments, and parameters for controlling and signaling the subdivision of pictures 12 and 12' into respective segments. Decoder 20 uses these parameters to subdivide the picture in the same way as the encoder did, assign the same prediction modes to the segments, and perform the same predictions resulting in the same prediction signals.
[0168] 3 shows the relationship between, on the one hand, the reconstructed signal, i.e. the reconstructed picture 12′, and, on the other hand, the combination of the prediction residual signal 24″″ and the prediction signal 26 signaled in the data stream 14. As already mentioned above, the combination may be additive. The prediction signal 26 is shown in FIG. 3 as a subdivision of the picture area into intra-coded blocks, exemplarily shown using line shading, and inter-coded blocks, exemplarily shown without line shading. The subdivision may be any subdivision, such as a regular subdivision of the picture area into rows and columns of square or non-square blocks, or a multi-tree subdivision of the picture 12 from a tree root block into several leaf blocks of various sizes, such as a quad-tree subdivision, a mixture of which is shown in FIG. 3, where the picture area is first subdivided into rows and columns of tree root blocks, which are then further subdivided into one or more leaf blocks according to a recursive multi-tree subdivision. For example, pictures 12, 12' may be subdivided into coding blocks 80, 82, which may represent, for example, coding tree blocks, and may be further subdivided into smaller coding blocks 83.
[0169] Again, data stream 14 may have an intra-coding mode coded for intra-coded blocks 80, which assigns one of several supported intra-coding modes to each intra-coded block 80. For inter-coded blocks 82, data stream 14 may have one or more motion parameters coded. Generally speaking, inter-coded blocks 82 are not limited to being temporally coded. Alternatively, inter-coded blocks 82 may be any block predicted from a previously coded portion beyond current picture 12 itself, such as a previously coded picture of the video to which picture 12 belongs, or a picture of another view or hierarchically lower layer if the encoder and decoder are scalable encoder and decoder, respectively.
[0170] The prediction residual signal 24'''' in FIG. 3 is also shown as a subdivision of the picture domain into blocks 84. These blocks may be called transform blocks to distinguish them from the coded blocks 80 and 82. In fact, FIG. 3 shows that the encoder 10 and the decoder 20 may use two different subdivisions of the picture 12 and the picture 12' into blocks, one into coded blocks 80 and 82, and the other into transform blocks 84. Although both subdivisions may be the same, i.e., each coded block 80 and 82 may simultaneously form a transform block 84, FIG. 3 shows the case where the subdivision into transform blocks 84 forms an extension of the subdivision into the coded blocks 80, 82, for example, such that any boundary between the two blocks 80 and 82 covers the boundary between the two blocks 84, or each block 80, 82 coincides with one of the transform blocks 84 or coincides with a cluster of transform blocks 84. However, these subdivisions may also be determined or selected independently of each other, such that the transformation blocks 84 may alternatively cross the block boundary between the blocks 80, 82. Thus, as far as the subdivision into transformation blocks 84 is concerned, similar statements are true as those presented with respect to the subdivision into blocks 80, 82, i.e., the blocks 84 may be the result of a regular subdivision of the picture area into blocks (with or without arrangement into rows and columns), a recursive multi-tree subdivision of the picture area, or a combination thereof, or any other type of blocking. It should be noted that the blocks 80, 82, and 84 are not limited to being square, rectangular, or any other shape.
[0171] 3 further illustrates that the combination of the prediction signal 26 and the prediction residual signal 24'''' directly results in the reconstructed signal 12'. However, it should be noted that according to alternative embodiments, multiple prediction signals 26 can be combined with the prediction residual signal 24'''' into the picture 12'.
[0172] As already outlined above, Figs. 1-3 are presented as examples in which the inventive concepts further described below can be implemented to form specific examples of encoders and decoders according to the present application. To that extent, the encoders and decoders of Figs. 1 and 2, respectively, can represent possible implementations of the encoders and decoders described later in this specification. Figs. 1 and 2 are, however, merely examples. An encoder according to an embodiment of the present application can, however, perform block-based encoding of a picture 12' from a data stream 14 using the coding concepts further outlined below, but can differ, for example, from the decoder 20 of Fig. 2 in that it subdivides the same picture 12' into blocks in a different way than described with respect to Fig. 3 and / or in that it does not derive a prediction residual from the data stream 14 in the transformed domain, but for example in the spatial domain.
[0173] 2. Fraction sample according to Fig. 4 The embodiments of the present disclosure may be part of the concept of motion compensated prediction, which may be an example of inter-picture prediction as may be applied by the prediction modules 44, 58. MCP may be performed with fractional sample precision. For example, a sample may refer to a pixel or location of a picture or video sequence, a full sample location may refer to a location signaled in the picture or video sequence, and a fractional sample location may refer to a location between full sample locations.
[0174] FIG. 4 illustrates integer samples 401 (shaded blocks in uppercase) and fractional sample positions 411 (non-shaded blocks in lowercase) of quarter-sample luma interpolation. The samples illustrated in FIG. 4 may form or be part of a portion 400 of a video sequence. Additionally, fractional sample positions 411 at quarter-sample resolution are illustrated with non-shaded blocks and lowercase. For example, portion 400 may represent a luma-coded block or may be part of a luma-coded block. Thus, FIG. 4 may represent a scheme for quarter-sample luma interpolation. However, luma-coded blocks are only one example of a coded block that may be used in the context of MCP. MCP may also be performed using chroma samples, or other schemes of color planes, for example samples of a video sequence using three color planes. For example, fractional sample positions 411 with quarter-sample resolution include half sample positions 415 and quarter sample positions 417.
[0175] The fractional sample positions 411 can be obtained from the full samples 401 by an interpolation filter. Different schemes for the interpolation filter can be possible. The scheme for the interpolation filter can provide a set of interpolation filters, each of which represents a rule for obtaining one or more specific fractional sample positions.
[0176] As an example of an interpolation filter, the scheme used in HEVC to obtain 15 fractional luma sample positions, e.g., the non-shade fractional sample positions of block 412, is described with respect to Figure 4. The HEVC scheme for the interpolation filter utilizes the FIR filters listed in Table 1.
[0177] 2.1 Interpolation Filters in HEVC In HEVC, which uses a quarter luma sample motion vector precision, there are 15 fractional luma sample positions (e.g., as shown in Figure 4) that are calculated using various one-dimensional FIR filters. However, for each fractional sample position, the interpolation filter is fixed.
[0178] [Table 1]
[0179] This scheme includes one symmetric 8-tap filter h[i] for half-sample (HPEL) positions, and one for position b 0,0 is the integer sample position A -3,0 From A 4,0 is calculated by filtering horizontally at position h 0,0 is the integer sample position A 0,-3 From A 0,4 is calculated by filtering vertically at position j 0,0 is calculated by horizontal filtering as described above. 0,-3 From b 0,4 This is calculated by first computing x, y, and then filtering these vertically using the same 8-tap filter. b 0,0 =Σ i=-3..4 A i,0 h[i],h 0,0 =Σ j=-3..4 A 0,j h[j],j 0,0 =Σ j=-3..4 b 0,j h[j]
[0180] One asymmetric 7-tap filter with coefficients q[i] for quarter-sample (QPEL) positions: Again, positions with only one fractional sample component are computed by 1D filtering of the nearest integer sample position in the vertical or horizontal direction. a0,0=Σi=-3..3 Ai,0 q[i],c0,0=Σi=-2..4 Ai,0 q[1-i] d0,0=Σj=-3..3 A0,jq[j],n0,0=Σj=-2..4 A0,jq[1-j]
[0181] All other QPEL positions are calculated by vertical filtering of the nearest fractional sample positions. For positions with a vertical HPEL component, the HPEL 8-tap filter is used.
[0182] j0,0=Σj=-3..4 a0,jh[j] or i0,0=Σj=-3..4 a0,jh[j],k0,0=Σj=-3..4 c0,jh[j] For locations with a vertical QPEL component, a QPEL 7-tap filter is used. e0,0=Σj=-3..4 a0,jh[j],f0,0=Σj=-3..4 b0,jh[j],g0,0=Σj=-3..4 c0,jh[j]
[0183] p0,0=Σj=-2..4 a0,jq[1-j],q0,0=Σj=-2..4 b0,jq[1-j],r0,0=Σj=-2..4 c0,jq[1-j] Applying this scheme, we can obtain fractional sample positions as shown in FIG.
[0184] 3. Encoders according to Fig. 5, Fig. 6 and Fig. 13 Fig. 5 illustrates an encoder 500 for hybrid video coding according to an embodiment. The encoder 500 is configured to provide an encoded representation 514 of a video sequence based on an input video content 512. The encoder 500 is configured to determine, for example by a syntax determination module 521, one or more syntax elements 520 associated with a portion 585 of the video sequence. Furthermore, the encoder 500 is configured to select a processing scheme 540 to be applied to the portion 585 of the video sequence based on characteristics described by the one or more syntax elements 520. For example, the encoder 500 comprises a processing scheme selector 541 for selecting the processing scheme 540. The processing scheme 540 is for obtaining samples, for example samples 401, 411, 415, 417, for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence. The encoder is configured to code the index 550, for example by an index encoder 561, where the index 550 indicates a selected processing scheme 540, such that a given coded index value represents different processing schemes depending on the characteristics described by the one or more syntax elements 520. Furthermore, the encoder 500 is configured to provide a bitstream including the one or more syntax elements 520 and the coded index 560 as a coded representation 514 of the video sequence. For example, the encoder 500 comprises a coder 590, for example an entropy coder, for providing the coded representation 514. The index encoder 561 may be part of the coder 590 or may be separate.
[0185] Fig. 6 illustrates an encoder 600 for hybrid video coding according to a further embodiment. The encoder 600 is configured to provide an encoded representation 614 of a video sequence based on an input video content 512. The encoder 600 is configured to determine, for example by a syntax determination module 521, one or more syntax elements 520 associated with a portion 585 of the video sequence. Furthermore, the encoder 500 is configured to select a processing scheme 540 to be applied to the portion 585 of the video sequence based on characteristics described by the one or more syntax elements 520. For example, the encoder 500 comprises a processing scheme selector 541 for selecting the processing scheme 540. The processing scheme 540 is for obtaining samples, for example samples 401, 411, 415, for motion compensated prediction at integer and / or fractional positions within the portion of the video sequence. The encoder 600 is configured to select an entropy coding scheme 662 used to provide an encoded index 660 indicative of the selected processing scheme 540 depending on the characteristics described by the one or more syntax elements 520. Further, the encoder 600 is configured to provide, as the coded representation 614 of the video sequence, a bitstream that includes the one or more syntax elements 520 and that includes a coded index 660. The coded index value may represent a value of the coded index 660.
[0186] 13 shows an encoder 1300 for hybrid video coding according to a further embodiment. The encoder 1300 is configured to provide an encoded representation 1314 of a video sequence based on input video content 512. The encoder 1300 is configured to determine one or more syntax elements 520 associated with a portion 585 of the video sequence.
[0187] According to one embodiment, the encoder 1300 is configured to select a processing scheme 540 to be applied to a portion 585 of a video sequence based on characteristics described by one or more syntax elements 520, the processing scheme being fully specified by characteristics of the portion (585) and fractional portions of motion vectors, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 585 of the video sequence.
[0188] According to an alternative embodiment, the encoder 1300 is configured to select a processing scheme 540 to be applied to a portion 585 of a video sequence based on the motion vector precision described by one or more syntax elements 520 .
[0189] According to another alternative embodiment, the encoder 1300 is configured to select a processing scheme 540 to be applied to a portion 585 of a video sequence based on characteristics described by one or more syntax elements 520, at PU or CU granularity.
[0190] Further, the encoder 1300 is configured to provide, as the coded representation 1314 of the video sequence, a bitstream that includes one or more syntax elements 520.
[0191] According to one embodiment, the encoder 1300 is configured to select another processing scheme other than the standard processing scheme for the HPEL position of the sample to be obtained if the determined one or more syntax elements 520 indicate MV precision of HEPL and if the prediction mode is one of translation-inter, translation merge, and combined inter / intra.
[0192] The following description applies equally to the encoder 500 according to FIG. 5, the encoder 600 according to FIG. 6, and the encoder 1300 according to FIG. 13, unless stated otherwise.
[0193] For example, the encoder 500, 600, 1300 receives a video sequence as input video content and provides an encoded representation 514, 614, 1314 of the video sequence, which may include multiple video frames. The syntax determination module 521 can obtain the portion 585 of the video sequence, for example by dividing the video sequence or its video frames, or by obtaining the portion 585 from a module for dividing the video sequence. The portion 585 can be the result of partitioning the video sequence, for example, into coding tree units or coding units or coding tree blocks or coding blocks, for example, as described with respect to FIG. 3. For example, the portion 585 can result from partitioning the video sequence used for inter- or intra-picture prediction, for example by the prediction module 44, 58 of FIG. 1 or FIG. 2. The syntax determination module 521 can further determine one or more syntax elements 520 based on the portion 585 of the video sequence and optionally further portions.
[0194] The encoder 500, 600, 1300 optionally comprises a processing module 544 for applying the processing scheme 540 to the portion 585 to obtain a prediction signal as indicated by the dashed arrow. For example, the prediction signal may be used to obtain an encoded representation 514, 614, 1314 of a video sequence. For example, the encoding of the input video content 512 into the encoded representation 514, 614, 1314 may be performed according to the scheme described with respect to the encoder 10 of FIG. 1, where the picture 12 may represent the video content 512 and the bitstream 14 may include the encoded representation 514, 614, 1314. For example, the processing module 544 may comprise or correspond to the prediction module 44, such that the processing scheme 540 may represent an input of the prediction module 44, which may use the processing scheme 540 for MCP to provide the prediction signal 26, for example to use the AMVR introduced in the introductory part of the description.
[0195] According to one embodiment, the processing scheme 540 is a filter or set of filters. For example, the one or more filters may be FIR filters, such as interpolation filters.
[0196] For example, the processing scheme 540 may be an interpolation filter, such as the 8-tap and 7-tap filters described in section 2.1, or a similar interpolation filter, which may be used, for example, to obtain fractional sample positions for MCP. In an embodiment of the present invention, two additional 1D 6-tap low-pass FIR filters may be supported and applied instead of the HEVC 8-tap filter (with coefficients h[i] listed in Table 2). In the following, the interpolation filters may serve as an example of a processing scheme.
[0197] For example, the processing scheme 540 may have spectral characteristics, such as low-pass or high-pass characteristics, so that the encoders 500, 600, 1300 may be configured to select between processing schemes having different characteristics. According to one embodiment, the encoder 500, 600 is configured to select different processing schemes for different blocks within a video frame or picture.
[0198] For example, a video frame of a video sequence may be divided into a number of blocks or portions, for example as illustrated by blocks 80, 82, 83 in FIG. 3, by, for example, the encoder 500, 600, 1300. For example, the encoder 500, 600, 1300 may obtain a prediction signal 26 for each of the portions 80, 82, 83. The encoder 500, 600, 1300 may select a processing scheme 540 for different portions of the video frame individually. In this context, a portion or block of a video frame may refer to a regular sized block or a first level block resulting from a first recursion of the division of the video frame, such as a coding tree block or coding tree unit, for example, coding block 80 or 82. The encoder 500, 600, 1300 may also select different processing schemes for different sub-portions of such portions or first level blocks or coding tree units, for example, coding block 83. The encoders 500, 600, 1300 may even select different processing schemes for different portions or blocks of a coded block or coded unit.
[0199] Thus, the encoders 500, 600, 1300 can adapt the interpolation filters at a fine granularity of pictures or video frames of a video sequence, allowing adaptation to local characteristics within one slice, tile or picture, i.e., at a granularity smaller than the size of the slice or tile or picture.
[0200] Thus, the encoder 500, 600, 1300 may determine information regarding the size or shape of the portion 585, such as a block size or block shape, and provide this information as one or more syntax elements 520. The encoder 500, 600, 1300 may also determine or obtain further parameters that may be used in the context of an intra-picture or inter-picture prediction model. For example, the encoder 500, 600, 1300 may determine an MV precision to be applied to the portion 585, for example, by the processing module 544 to obtain a prediction signal for encoding the portion 585.
[0201] Thus, the one or more syntax elements 520 may include at least one of a motion vector precision, a fractional sample position, a block size, a block shape, a number of prediction hypotheses, a prediction mode, availability of a coded residual signal, spectral characteristics of the coded residual signal, a reference picture signal, loop filter data, a motion vector length, or an adaptive motion vector resolution mode.
[0202] For example, the encoder 500, 600, 1300 may infer the processing scheme 540 from already determined characteristics or processing steps, which may be indicated by one or more syntax elements 520. For example, for HEVC and VVC merge modes, the interpolation filters to use may be inherited from the selected merge candidate. In another embodiment, it may be possible to override the interpolation filters inherited from the merge candidate by explicit signaling.
[0203] According to one embodiment, the encoder 500, 600, 1300 is configured to determine a set of processing schemes that are permissible in terms of the characteristics described by one or more syntax elements and to select the processing scheme 540 from the determined set. For example, the processing scheme selector 541 may infer from the one or more syntax elements 520 which processing schemes of the processing scheme entities are compatible with or can be advantageously applied to the portion 585 in order to obtain an efficient encoding of the video sequence, or in particular the portion 585.
[0204] In other words, one or more processing schemes are available or permissible or selectable for the portion 585 depending on one or more characteristics of the portion 585 indicated by one or more syntax elements 520. Thus, only a single processing scheme may be permissible for the portion 585. In this case, the encoder 500, 600 may select only the permissible processing scheme as the processing scheme 540. Since the processing scheme may in this case be uniquely identified from the one or more syntax elements 520, the encoder 500, 600 may skip providing the coded index 550, 560 in the coded representation 514, 614 to avoid redundant information and signaling overhead in the bitstream. If multiple processing schemes are permissible for the portion 585, the encoder may determine the permissible processing schemes as a set of processing schemes permissible for the portion 585 and select the processing scheme 540 from this set of processing schemes. In the latter case, the encoder 500, 600 may encode an encoded index 560, 660 in the encoded representation 514, 614 to indicate the selected processing scheme 540. The embodiment described in Section 6 may be an example of such an embodiment.
[0205] Thus, in some cases, processing schemes or interpolation filters can be adapted to different portions of a video frame without the need to explicitly signal the processing scheme or interpolation filter applied, and therefore, additional bits may not be required in the bitstream containing the coded representations 514, 516.
[0206] According to an embodiment that may correspond to the encoder 1300, the processing scheme is in all cases uniquely determined based on the characteristics described by one or more syntax elements, so that no explicit indication of the processing scheme 540 in the encoded representation 1314 with a coded index dedicated to the processing scheme 540 is required. Thus, according to an embodiment, the encoder 1300 may not need to decide whether to code a coded index or not, or may not need to select a coding scheme.
[0207] According to one embodiment, the encoder 500, 600, 1300 is configured to determine the set of available processing schemes depending on the motion vector precision. For example, the available (or allowable) interpolation filter set may be determined by the motion vector precision selected for the block (signaled using a dedicated syntax element).
[0208] For example, one or more syntax elements 520 may indicate the MV precision selected for portion 585, e.g., by using a scheme for signaling the MV precision described in Section 6.1.2 with respect to Table 3.
[0209] According to one embodiment, the encoders 500, 600, 1300 are configured to select between quarter-sample motion vector resolution, half-sample motion vector resolution, full-sample motion vector resolution, and four-sample motion vector resolution.
[0210] According to one embodiment, the encoder 500, 600, 1300 is configured to selectively determine a set of processing schemes for a motion vector precision between a maximum motion vector precision and a minimum motion vector precision, or between a maximum motion vector precision and a full sample motion vector precision, and select a processing scheme 540 from the determined set.
[0211] For example, in one embodiment, multiple sets of interpolation filters may be supported for half a luma sample precision, and a default set of interpolation filters may be used for all other possible motion vector precisions.
[0212] According to one embodiment of the encoder 500, 600, if the AMVR mode indicates HPEL precision (amvr mode=2), the interpolation filters may be explicitly signaled by one syntax element (if_idx) per CU. When if_idx is equal to 0, it indicates that the filter coefficients h[i] of the HEVC interpolation filter are used to generate the HPEL position.
[0213] When if_idx is equal to 1, it indicates that the filter coefficients h[i] of the flat-top filter (see table below) are used to generate the HPEL position. When if_idx is equal to 2, it indicates that the filter coefficients h[i] of the Gaussian filter (see table below) are used to generate the HPEL position.
[0214] When a CU is coded in merge mode and merge candidates are generated by copying motion vectors from neighboring blocks without modifying the motion vectors, the index if_idx is also copied. If if_idx is greater than 0, the HPEL position is filtered using one of two non-HEVC filters (flat-top or Gaussian).
[0215] [Table 2]
[0216] To obtain the coded index 560, 660, the encoder 500, 600 can apply a mapping rule that maps the index 550 or an index indicating the selected processing scheme 540 to the coded index 560, 660. For example, the mapping rule may correspond to a binarization scheme. The encoder 500, 600 can adapt the coded index 560, 660 according to several allowable processing schemes so that the coded index 560, 660 requires a low number of bits.
[0217] For example, the encoder 500, 600 may select a processing scheme, and therefore a mapping rule, depending on the motion vector precision and / or the fractional sample position or fractional portion of the motion vector.
[0218] The encoder 500, 600 may further adapt the number of bins used to provide the encoded index 560, 660 depending on the characteristics described by one or more syntax elements 520, for example by adapting a mapping rule.
[0219] Therefore, the encoder 500, 600 selects a mapping rule that maps index values to encoded index values such that the number of bins representing the encoded index values is adapted to the number of processing schemes allowable for the portion 585 of the video sequence.
[0220] For example, the number of bins used to provide the coded index 560 may depend on the number of processing schemes allowed in terms of the property. For example, if only one processing scheme is allowed, the number of bins may be 0. Table 2 shows examples of three allowable processing schemes and up to three bins for the coded index 560, 660, along with corresponding mapping rules for mapping index values to coded index values.
[0221] For example, Table 2 shows one embodiment of a mapping rule for mapping if_idx to an encoded index value. The mapping rule may be different for different MV precisions.
[0222] In the embodiment described in Table 2, the interpolation filter may be modified only for the HPEL position. If the motion vector points to an integer luma sample position, the interpolation filter is not used in all cases. If the motion vector points to a reference picture position representing a full sample x (or y) position and a half sample y (or x) position, the selected half-pel filter is applied only in the vertical (or horizontal) direction. Furthermore, only the filtering of the luma component is affected; for the chroma components, conventional chroma interpolation filters are used in all cases.
[0223] The if_idx index may correspond to the index 550, for example as shown in Table 2, and the binarization value of if_idx may correspond to the encoded index 560, 660. Thus, the index encoder 561 may derive the encoded index 560 from the index 550 based on a binarization scheme, for example the binarization scheme shown in the last column of Table 2. In the context of the encoder 600, the binarization scheme shown in the last column of Table 2 may correspond to the entropy coding scheme 662 or may be part of the entropy coding scheme 662 selected by the coding scheme selector 661, based on which the coder 690 may provide the encoded index 660, which may correspond to, for example, the binarization value shown in Table 2, or an encoded representation thereof. For example, a characteristic for which the encoder 600 may select the entropy coding scheme 662 may correspond to the MV precision. Thus, Table 2 shows an example where one or more of the syntax elements 520 indicate that the MV precision of the portion 585 is HPEL.
[0224] According to one embodiment, the encoder 500, 600 is configured to determine whether a single processing scheme 540 is acceptable for the portion 585 of the video sequence, and selectively omit inclusion of the coded index in response to a finding that only a single processing scheme 540 is acceptable for the portion 585 of the video sequence. For example, the encoder 500 may omit inclusion of the coded index 560 in the coded representation 514. The encoder 600 may, for example, select an entropy coding scheme 662 that indicates that the coded index 660 is not included in the coded representation 614, or that the number of bins or bits used for the coded index 660 in the bitstream is zero.
[0225] According to one embodiment, the adaptive filter selection is extended to the chroma components. For example, the encoder 500, 600 can select the processing scheme 540 depending on the characteristics of the chroma components. The chroma components may be part of a chroma-coded block that may be associated with a luma-coded block. For example, the portion 585 may be part of a luma-coded block and / or part of a chroma-coded block. 4. Decoder according to Figures 7 and 8
[0226] 7 illustrates a decoder 700 for hybrid video coding according to one embodiment. The decoder 700 is configured to provide output video content 712 based on an encoded representation 714 of a video sequence. The encoded representation 714 may correspond to an encoded representation 514, 614, 1314 provided by an encoder 500, 600, 1300. The decoder 700 is configured to obtain a bitstream as the encoded representation 714 of the video sequence, and to identify from the bitstream one or more syntax elements 520 associated with a portion 785 of the video sequence that may correspond to a portion 585 of the input video content 512.
[0227] According to one embodiment, the decoder 700 comprises a syntax identifier 721 for identifying one or more syntax elements 520, possibly coded in the representation 714 coded by the encoder 500, 600. Furthermore, the decoder 700 is configured to identify a processing scheme 540 for the portion 785 of the video sequence based on the one or more syntax elements 520 and the index 550, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer 401 and / or fractional positions 411, 415, 417 within the portion 785 of the video sequence, and the decoder 700 is configured to assign different processing schemes to a given coded index value depending on the one or more syntax elements 520.
[0228] For example, the processing scheme selector 741 may determine an index 550 from the encoded representation 714 based on one or more syntax elements 520, and may then use the one or more syntax elements 520 and / or the index 550 to determine the processing scheme 540. Depending on the one or more syntax elements 520, the processing scheme selector 741 may also infer that the encoded representation 714 and the one or more syntax elements 520 do not include the index 550, or may infer the index 550 from the one or more syntax elements 520.
[0229] According to an alternative embodiment, the decoder 700 is configured to select a processing scheme 540 for a portion 785 of a video sequence in response to one or more syntax elements 520, the processing scheme being fully specified by characteristics of the portion 785 and fractional parts of motion vectors, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence. According to another alternative embodiment, the decoder 700 is configured to select the processing scheme 540 for the portion 785 of the video sequence depending on the motion vector precision.
[0230] According to another alternative embodiment, the decoder 700 is configured to determine the processing scheme 540 of the portion 785 of the video sequence at PU or CU granularity and depending on one or more syntax elements 520 . Further, the decoder 700 is configured to apply, for example by a processing module 744, the identified processing scheme 540 to the portion 785 of the video sequence.
[0231] 8 illustrates a decoder 800 for hybrid video coding according to another embodiment. The decoder 800 is configured to provide an output video content 812 based on an encoded representation 714 of a video sequence. The encoded representation 714 may correspond to an encoded representation 514, 614 provided by the encoder 500, 600. The decoder 800 is configured to obtain a bitstream as the encoded representation 714 of the video sequence and to identify from the bitstream one or more syntax elements 520 associated with a portion 785 of the video sequence that may correspond to a portion 585 of the input video content 512. For example, the decoder 700 comprises a syntax identifier 821 for identifying one or more syntax elements 520 that may have been encoded in the encoded representation 714 by the encoder 500, 600. Furthermore, the decoder 800 is configured to determine, in response to the one or more syntax elements 520, an entropy coding scheme 662 to be used in the bitstream to provide an encoded index indicative of a processing scheme 540 of the portion 785 of the video sequence, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence. For example, the decoder 800 comprises a coding scheme selector 861, which may correspond to the coding scheme selector 661 of the encoder 600, and determines the coding scheme 662 based on the one or more syntax elements 520. The decoder 800 is further configured to identify, in the bitstream, the encoded index and decode the index, for example by the processing scheme selector 841, if the determined entropy coding scheme 662 defines a number of bins greater than zero. The decoder 800 is further configured to identify a processing scheme 540 based on the one or more syntax elements 520 and based on the decoded index if the number of bins defined by the determined coding scheme is greater than zero. For example, the processing scheme selector 841 may identify the processing scheme 540.Further, the decoder 800 is configured to apply the identified processing scheme 540 to the portion 785 of the video sequence by a processing module 844, which may correspond to, for example, processing module 744.
[0232] The decoders 700, 800 may obtain the output video content 712, 812 from the encoded representations 514, 614, 1314 provided by the encoders 500, 600 based on the input video content 512 such that the output video content 712, 812 may correspond to the input video content 512 except for coding losses such as quantization losses. Thus, the description of the encoders 500, 600 and their features may equally apply to the decoders 700, 800, although some processes may be performed in reverse. For example, features of the encoders 500, 600 referring to encoding may correspond to features of the decoders 700, 800 referring to decoding.
[0233] For example, the decoder 700, 800 may read or decode one or more syntax elements 520 from the encoded representation 714, 814, 1314 as provided by the encoder 500, 600, 1300. Thus, the encoder 500, 600, 1300 and the decoder 700, 800, 1300 may apply equivalent assumptions or knowledge to determine the processing scheme 540 from the one or more syntax elements 520.
[0234] For example, the coded index 560, 660 may indicate information that the decoder 700, 800 cannot infer from one or more syntax elements 520, such that some embodiments of the decoder 700, 800 may determine the processing scheme 540 based on the coded index 560, 660. For example, some embodiments of the decoder 700, 800 may decode the coded index 560, 660 from the coded representation 714, 814 to obtain an index, e.g., the index 550, if, for example, the decoder 700, 800 infers from one or more syntax elements 520 that some processing schemes are acceptable. However, it is pointed out that not all embodiments of the decoder 700 are configured to infer whether some processing schemes are acceptable. Some embodiments of the decoder 700 may be configured to infer the processing scheme 540 from one or more syntax elements of the coded representation 1314, for example, in all cases or independently of the indication of one or more syntax elements 520. Thus, embodiments of the decoder 700 can infer the processing scheme 540 from one or more syntax elements 520 without decoding an index explicitly dedicated to the processing scheme 540 .
[0235] Thus, features, functions, advantages and examples of one or more syntax elements 520, processing scheme 540, portion 585, index 550, optionally encoded index 560, 660 and further details of encoders 500, 600, 1300 may equally apply to decoders 700, 800, although the description of the elements is not explicitly repeated.
[0236] For example, the decoders 700, 800 may be implemented in the decoder 20 of Figure 2. For example, the processing modules 744, 844 may comprise or correspond to the prediction module 58, such that the processing scheme 540 may represent an input for the prediction module 58, which may use the processing scheme 540 for the MCP to provide a predicted signal 59. The relationship between the encoders 500, 600 and the decoders 700, 800 may correspond to the relationship between the encoder 10 and the decoder 20, as described in Section 1.
[0237] According to one embodiment, the decoder 700, 800 is configured to determine whether a single processing scheme 540 is acceptable for a portion 785 of the video sequence, and, in response to a finding that only a single processing scheme 540 is acceptable for the portion 785 of the video sequence, selectively omit identifying the coded index.
[0238] For example, the decoder 700 may determine whether a single processing scheme 540 is acceptable for the portion 785 based on one or more syntax elements 520. Depending on the result, the decoder 700 may identify or decode an encoded index that may correspond to the encoded index 560, 660, or the decoder 700 may omit or skip identifying an encoded index and select a single acceptable processing scheme for the portion 785.
[0239] According to a further example, the decoder 800 can infer from the entropy coding scheme 662 whether to decode an encoded index from the encoded representation 814. If an encoded index is to be decoded, the decoder 800 can decode an encoded index that may correspond to the encoded index 560, 660 to determine an index that may correspond to the index 550 and determine the processing scheme 540 based on the index. If there is no encoded index to be decoded, the decoder 800 can omit or skip decoding the encoded index and determine the processing scheme 540 based on the one or more syntax elements 520.
[0240] According to one embodiment, the decoder 700, 800 is configured to select a filter depending on the fractional part of a motion vector. For example, a motion vector may be represented by two components that refer to two different directions. For example, a motion vector may be represented by a horizontal component and a vertical component. Horizontal and vertical can represent two orthogonal directions in the plane of a picture or video frame. Thus, the fractional part of a motion vector can refer to a component of a MV that has a fractional resolution.
[0241] According to one embodiment, the decoder 700, 800 is configured to select different filters for filtering in a first direction and filtering in a second direction depending on the fractional part of the motion vector. For example, the decoder 700, 800 may select different filters according to a first component of the MV representing a component of the MV in a first direction and a second component of the MV representing a component of the MV in a second direction.
[0242] The encoded representations 514, 614, 714, 1314 may be part of a video bitstream provided by the encoder 500, 600, 1300, which video bitstream includes encoded representations of multiple portions of a video sequence of the input video content.
[0243] According to one embodiment, a video bitstream for hybrid video coding that describes a video sequence includes one or more syntax elements 520 that relate to portions 585 of the video sequence. The video bitstream includes a processing scheme 540 index bitstream element for at least some of the portions of the video sequence, the presence of the processing scheme 540 index bitstream element varying depending on one or more other syntax elements 520. For example, the video bitstream may correspond to or include a coded representation 514, 614, 714.
[0244] According to one embodiment, a video bitstream for hybrid video coding, the video bitstream describing a video sequence, includes one or more syntax elements 520 associated with portions 585 of the video sequence. The video bitstream includes an index bitstream element of a processing scheme 540 for at least some of the portions of the video sequence, where a number of bins of the index bitstream element of the processing scheme 540 varies depending on the one or more syntax elements 520 associated with the portions 585 of the video sequence. For example, the video bitstream may correspond to or include the coded representations 514, 614, 714.
[0245] According to one embodiment, a video bitstream for hybrid video coding describing a video sequence includes one or more syntax elements 520 associated with portions 585 of the video sequence. The video bitstream includes an index bitstream element of a processing scheme 540 for at least some of the portions of the video sequence, and the meaning of a codeword of the index bitstream element of the processing scheme 540 varies depending on the one or more syntax elements 520 associated with the portions 585 of the video sequence. For example, the video bitstream may correspond to or include the coded representations 514, 614, 714. 5. Method for hybrid video coding according to Figs. 9, 10, 11 and 12
[0246] Fig. 9 shows a flow chart of a method 5000 for hybrid video coding according to an embodiment. The method 5000 for providing an encoded representation 514 of a video sequence based on an input video content 512, the method 5000 comprising a step 5001 of determining one or more syntax elements 520 associated with a portion 585 of the video sequence. Step 5002 comprises selecting a processing scheme 540 to be applied to the portion 585 of the video sequence based on characteristics described on the one or more syntax elements 520, the processing scheme 540 for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 585 of the video sequence. Step 5003 comprises encoding an index indicating the selected processing scheme 540, such that a given encoded index value represents different processing schemes depending on the characteristics described by the one or more syntax elements 520. Step 5004 comprises providing a bitstream comprising the one or more syntax elements 520 and the encoded index as an encoded representation of the video sequence.
[0247] Fig. 10 shows a flow chart of a method 6000 for hybrid video coding according to an embodiment. The method 600 for providing an encoded representation of a video sequence based on an input video content, the method 600 comprising a step 6001 of determining one or more syntax elements 520 associated with a portion 585 of the video sequence. A step 6002 comprises selecting a processing scheme 540 to be applied to the portion 585 of the video sequence based on characteristics described on the one or more syntax elements 520, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 585 of the video sequence. A step 6003 comprises selecting an entropy coding scheme to be used for providing an encoded index indicative of the selected processing scheme 540 depending on the characteristics described by the one or more syntax elements 520. A step 6004 comprises providing a bitstream including the one or more syntax elements 520 and including the encoded index as an encoded representation of the video sequence.
[0248] Fig. 11 shows a flow chart of a method 7000 for hybrid video coding according to an embodiment. The method 7000 for providing an output video content based on an encoded representation of a video sequence, the method 7000 comprising a step 7001 of obtaining a bitstream as the encoded representation of the video sequence. Step 7002 comprises identifying from the bitstream one or more syntax elements 520 associated with the portion 785 of the video sequence. Step 7003 comprises identifying a processing scheme 540 for the portion 785 of the video sequence based on the one or more syntax elements 520 and an index, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence, the method assigning different processing schemes to a given encoded index value depending on the one or more syntax elements 520. Step 7004 comprises applying the identified processing scheme 540 to the portion 785 of the video sequence.
[0249] Fig. 12 shows a flow chart of a method 8000 for hybrid video coding according to an embodiment. The method 8000 for providing an output video content based on an encoded representation of a video sequence, the method 8000 comprising a step 8001 of obtaining a bitstream as an encoded representation of the video sequence. Step 8002 comprises identifying from the bitstream one or more syntax elements 520 associated with the portion 785 of the video sequence. Step 8003 comprises determining, in response to the one or more syntax elements 520, an entropy coding scheme used in the bitstream for providing an encoded index indicative of a processing scheme 540 of the portion 785 of the video sequence, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence. Step 8004 comprises identifying, in the bitstream, an encoded index and decoding the index if the determined entropy coding scheme defines a number of bins greater than zero. Step 8005 includes identifying a processing scheme 540 based on the one or more syntax elements 520 and based on the decoded index if the number of bins defined by the determined entropy coding scheme is greater than zero. Step 8006 includes applying the identified processing scheme 540 to the portion 785 of the video sequence.
[0250] 14 shows a flowchart of a method 7100 for hybrid video coding according to an embodiment. The method 7100 is for providing an output video content 712 based on an encoded representation 1314 of a video sequence, the method 7100 comprising: obtaining 7101 a bitstream as the encoded representation 714 of the video sequence; identifying 7102 from the bitstream one or more syntax elements 520 associated with a portion 785 of the video sequence; selecting 7103 a processing scheme 540 for the portion 785 of the video sequence in response to the one or more syntax elements 520, the processing scheme being fully specified by characteristics of the portion 785 and fractional parts of motion vectors, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence; and applying 7104 the identified processing scheme 540 to the portion 785 of the video sequence.
[0251] 15 shows a flowchart of a method 7200 for hybrid video coding according to an embodiment. The method 7200 is for providing an output video content 712 based on an encoded representation 1314 of a video sequence, the method 7200 comprising: obtaining 7201 a bitstream as the encoded representation 714 of the video sequence; identifying 7202 from the bitstream one or more syntax elements 520 associated with a portion 785 of the video sequence; selecting 7203 a processing scheme 540 for the portion 785 of the video sequence depending on a motion vector precision, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence; and applying 7204 the identified processing scheme 540 to the portion 785 of the video sequence.
[0252] 16 shows a flowchart of a method 7300 for hybrid video coding according to an embodiment for providing an output video content 712 based on an encoded representation 1314 of a video sequence, the method 7300 comprising the steps of: obtaining 7301 a bitstream as the encoded representation 714 of the video sequence; identifying 7302 from the bitstream one or more syntax elements 520 associated with a portion 785 of the video sequence; and determining 7303 a processing scheme 540 for the portion 785 of the video sequence in response to the one or more syntax elements 520 with a PU or CU granularity, the processing scheme 540 for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence; and applying 7304 the identified processing scheme 540 to the portion 785 of the video sequence.
[0253] 17 shows a flowchart of a method 1310 for hybrid video coding according to one embodiment. The method 1310 is for providing an encoded representation 1314 of a video sequence based on an input video content 512, the method 1310 including: determining 1311 one or more syntax elements 520 associated with a portion 585 of the video sequence; selecting 1312 a processing scheme 540 to be applied to the portion 585 of the video sequence based on characteristics described by the one or more syntax elements 520, the processing scheme being fully specified by characteristics of the portion 585 and fractional parts of motion vectors, the processing scheme 540 being for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 585 of the video sequence; and providing 1313 a bitstream including the one or more syntax elements 520 as the encoded representation 1314 of the video sequence.
[0254] 18 shows a flowchart of a method 1320 for hybrid video coding according to one embodiment. The method 1320 is for providing an encoded representation 1314 of a video sequence based on an input video content 512, the method 1320 including: determining 1321 one or more syntax elements 520 associated with a portion 585 of the video sequence; selecting 1322 a processing scheme 540 to be applied to the portion 585 of the video sequence based on a motion vector precision described by the one or more syntax elements 520, the processing scheme 540 for obtaining samples 401, 411, 415, 417 for motion compensated prediction at integer and / or fractional positions within the portion 785 of the video sequence; and providing 1323 a bitstream including the one or more syntax elements 520 as the encoded representation 1314 of the video sequence.
[0255] 19 shows a flowchart of a method 1330 for hybrid video coding according to an embodiment. The method 1330 is for providing an encoded representation 1314 of a video sequence based on an input video content 512, the method 1330 including: determining 1331 one or more syntax elements 520 associated with a portion 585 of the video sequence; selecting 1332 a processing scheme 540 to be applied to the portion 585 of the video sequence based on characteristics described by the one or more syntax elements 520 with PU or CU granularity, the processing scheme 540 for obtaining samples 401, 411, 415, 417 for motion compensation prediction at integer and / or fractional positions within the portion 785 of the video sequence; and providing 1333 a bitstream including the one or more syntax elements 520 as the encoded representation 1314 of the video sequence. 6. Further embodiments
[0256] In many, or possibly all, current video coding standards, including the current VVC draft [5], the interpolation filters are fixed for each fractional sample position. ("Fractional sample position" herein refers to the relative position between the four surrounding FPEL or integer positions). An embodiment of the present invention allows for the use of different interpolation filters (e.g., processing schemes) for the same fractional sample position in different parts (e.g., portions) of a video sequence. The aforementioned techniques for switching interpolation filters (see the "Background of the Invention" section) have the disadvantages of allowing switching between different interpolation filters only at a coarse granularity (i.e., slice level) and that in some cases (for the AIF and its variants) the selection may be explicitly signaled by transmitting individual FIR filter coefficients.
[0257] By allowing the interpolation filters to be switched at a finer granularity (e.g., PU / CU, or CTU level), it is possible to better adapt to local picture characteristics (e.g., in addition to the high-precision HEVC interpolation filters, it may be beneficial to use filters with different characteristics, e.g., filters with stronger low-pass characteristics to attenuate high-frequency noise components). However, explicit signaling of the interpolation filters used for each PU may be too expensive in terms of the required bitrate. Therefore, embodiments of the present invention avoid this signaling overhead by restricting the interpolation filters available to a part or region of a video sequence according to the characteristics of a given part or region (e.g., PU / CU, CTU, slice, frame, etc.) of the video sequence that are also known to the decoder (e.g., because they are indicated by one or more syntax elements 520 that may be part of the coded representation 514, 614, 1314). If only one interpolation filter remains available, no further signaling is necessary. If multiple remain available, only which of the available interpolation filters will be used may be transmitted. (Note that here and below, "interpolation filter" refers to the broader sense of any signal processing applied to obtain a reference sample at a given, possibly fractional, sample position, and may also include filtering of FPEL positions.)
[0258] According to an embodiment of the present invention, changing the used interpolation filter may only be possible if the current block has certain characteristics (given by the values of other decoded syntax elements). In this case, two cases can be distinguished:
[0259] Case 1: For given characteristics of a block (determined by other syntax elements), only one specific set of interpolation filters is supported. Here, the term "set of interpolation filters" refers to the fact that the actually used filters also depend on the subsample position (as in conventional video coding with subsample precision). The characteristics then determine the set of interpolation filters used, and the fractional part of the motion vector determines the actual filters (or horizontal and vertical filters, if filtering is specified by separable filters). Since the interpolation filters are fully specified by the characteristics of the block and the fractional part of the motion vector, no additional syntax elements are transmitted in this case.
[0260] Case 2: Multiple sets of interpolation filters are supported for a given characteristic of a block. In this case, the set of interpolation filters can be selected by coding a corresponding syntax element (e.g., an index into a list of supported sets). The interpolation filter used (from the selected set) can again be determined by the fractional part of the motion vector.
[0261] Note that both cases may be supported in the same picture. This means that for blocks with the first property, a fixed set of interpolation filters may be used (this set may be the same as the set of conventional interpolation filters or may deviate from the conventional set). Also, for blocks with the second property, multiple sets of interpolation filters may be supported, the selected filter being determined by the transmitted filter index (or any other means). Furthermore, blocks may have a third property, in which case another set of interpolation filters (one or more) may be used. These characteristics of portions or regions of a video sequence (portions of a video sequence) that can be used to limit the available interpolation filters include:
[0262] -MV precision (e.g. QPEL, HPEL, FPEL, etc.), i.e. different sets of interpolation filters may be supported for different motion vector precisions -Fractional sample position -Block size, or more generally, block shape The number of prediction hypotheses (i.e., 1 for uniprediction, 2 for biprediction, etc.), e.g. an additional smoothing filter for uniprediction only - Prediction mode (e.g., translation inter, affine inter, translation merge, affine merge, combined inter / intra) Availability of the coded residual signal (e.g., the coded block flag as in HEVC) -Spectral characteristics of the coded residual signal - a reference picture signal (reference blocks defined by motion data), e.g. Block edges from co-located PUs in the reference block High-frequency or low-frequency characteristics of the reference signal Loop filter data, e.g. Sample Adaptive Offset Filter (SAO) Edge offset or band offset classification from Deblocking filter determination and boundary strength -MV length, e.g. allowing additional smoothing filters only for long MVs or certain directions Combinations of two or more of these properties (eg, MV precision and fractional sample positions) are also possible (see above).
[0263] In one embodiment of the present invention, the available interpolation filters may be directly linked to the MV precision. In particular, additional or different interpolation filters (with respect to HEVC or VVC interpolation filters) may be available only for blocks that are coded using a particular MV precision (e.g., HPEL) and are applied, for example, only to HPEL positions. Signaling of the MV precision may use a similar scheme to AMVR (see Sec. 6.1.2) for VVC in this case. If only one interpolation filter (i.e., replacing the corresponding HEVC or VVC interpolation filter) is available, no further signaling is required. If more than one interpolation filter is available, an indication of the used interpolation filter may be signaled only for blocks using a given MV precision, while for all other blocks the signaling remains unchanged.
[0264] In one embodiment of the present invention, for HEVC or VVC merge modes, the interpolation filters used may be inherited from the selected merge candidate. In another embodiment, it may be possible to override the interpolation filters inherited from the merge candidate by explicit signaling. 6.1 Further embodiments of the invention
[0265] In one embodiment of the present invention, the set of available interpolation filters may be determined by the motion vector precision selected for the block (signaled using a dedicated syntax element): for half a luma sample precision, multiple sets of interpolation filters may be supported, while for all other possible motion vector precisions, a default set of interpolation filters is used. 6.1.1 Interpolation Filter Sets
[0266] In an embodiment of the present invention, two additional 1D 6-tap low-pass FIR filters may be supported, which may be applied instead of the HEVC 8-tap filter (with coefficients h[i] as described above). 6.1.2 Signalling
[0267] An extension of the AMVR CU syntax with truncated unary binarization from the current VVC draft with HPEL resolution, adding one CABAC context for the additional bin: - An index is only present if at least one MVD is greater than 0 - Index equal to 0 indicates QPEL precision and HEVC interpolation filters
[0268] - Index equal to 1 indicates that all MVDs have HPEL MV precision and must be shifted to QPEL precision before being added to the MVP rounded to HPEL precision. If the index is equal to 2 or 3, it indicates FPEL or 4PEL precision, respectively, and no interpolation filter is applied.
[0269] [Table 3]
[0270] The context of the first bin that switches between QPEL and other precisions can be chosen among three contexts depending on the value of the AMVR mode of the neighboring blocks to the left and above the current block: if both neighboring blocks have QPEL AMVR mode, the context is 0, otherwise (if only one neighboring block does not use QPEL AMVR mode), the context is 1, otherwise (both neighboring blocks do not use QPEL AMVR mode), the context is 2.
[0271] If the AMVR mode indicates HPEL precision (amvr mode = 2), the interpolation filters may be explicitly signaled by one syntax element (if_idx) per CU. When if_idx is equal to 0, it indicates that the filter coefficients h[i] of the HEVC interpolation filter are used to generate the HPEL position.
[0272] When if_idx is equal to 1, it indicates that the filter coefficients h[i] of the flat-top filter (see Table 2 in Section 3) are used to generate the HPEL position.
[0273] When if_idx is equal to 2, it indicates that the filter coefficients h[i] of the Gaussian filter (see Table 2 in Section 3) are used to generate the HPEL position.
[0274] If the CU is coded in merge mode and the merge candidates are generated by copying motion vectors from neighboring blocks without modifying the motion vectors, the index if_idx is also copied. If if_idx is greater than 0, the HPEL position is filtered using one of two non-HEVC filters (flat-top or Gaussian).
[0275] In this particular embodiment, the interpolation filter may be modified only for the HPEL position. If the motion vector points to an integer luma sample position, the interpolation filter is not used in all cases. If the motion vector points to a reference picture position representing a full sample x (or y) position and a half sample y (or x) position, the selected half-pel filter is applied only in the vertical (or horizontal) direction. Furthermore, only the filtering of the luma component is affected; for the chroma components, the conventional chroma interpolation filter is used in all cases. Possible variations of the described embodiments may include the following aspects.
[0276] The set of interpolation filters for HPEL precision also includes modified filters for full sample positions, i.e., if a half-sample accurate motion vector points to a full sample position in the x (or y) direction, stronger low-pass filtering in the horizontal (or vertical) direction may also be applied.
[0277] If full luma sample or 4 luma sample motion vector precision is selected, reference sample filtering may also be applied, where either a single filter or multiple filters can be supported. The adaptive filter selection is extended to the chroma components. The filter selected further depends on whether uni- or bi-prediction is used. 7. Implementation Alternatives:
[0278] Although some aspects are described in the context of an apparatus, it will be apparent that these aspects also represent a description of a corresponding method, with blocks or apparatus corresponding to method steps or features of method steps. Similarly, aspects described in the context of a method step also represent a description of the corresponding block or item or feature of the corresponding apparatus. Some or all of the method steps may be performed by (or using) a hardware apparatus, such as, for example, a microprocessor, a programmable computer or electronic circuitry. In some embodiments, one or more of the most important method steps may be performed by such an apparatus.
[0279] Depending on specific implementation requirements, the embodiments of the present invention can be implemented in hardware or software. The embodiments can be implemented using a digital storage medium, such as a floppy disk, DVD, Blu-ray, CD, ROM, PROM, EPROM, EEPROM or flash memory, having electronically readable control signals stored therein and cooperating (or capable of cooperating) with a programmable computer system to perform the respective methods. Thus, the digital storage medium may be computer readable.
[0280] Some embodiments according to the invention comprise a data carrier having electronically readable control signals which cooperate with a programmable computer system to perform one of the methods described herein.
[0281] Generally, embodiments of the present invention may be implemented as a computer program product having a program code operative to perform one of the methods when the computer program product runs on a computer. The program code may for example be stored on a machine readable carrier. Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine readable carrier.
[0282] In other words, an embodiment of the inventive method is a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.
[0283] A further embodiment of the inventive method is therefore a data carrier (or digital storage medium or computer readable medium) comprising, and recorded thereon, the computer program for performing one of the methods described herein. The data carrier, digital storage medium or recording medium is typically tangible and / or non-transitory.
[0284] A further embodiment of the inventive method is therefore a data stream or a sequence of signals representing the computer program for performing one of the methods described herein. The data stream or the sequence of signals can for example be arranged to be transferred via a data communication connection, for example the Internet.
[0285] A further embodiment comprises a processing means, for example a computer, or a programmable logic device, configured to or adapted to perform one of the methods described herein. A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.
[0286] Further embodiments according to the invention include an apparatus or system configured to transfer (e.g. electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may be, for example, a computer, a mobile device, a memory device, etc. The apparatus or system may, for example, comprise a file server for transferring the computer program to the receiver.
[0287] In some embodiments, a programmable logic device (e.g., a field programmable gate array) may be used to perform some or all of the functions of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor to perform one of the methods described herein. In general, the methods are preferably performed by any hardware device.
[0288] The apparatus described herein may be implemented using a hardware apparatus, or using a computer, or using a combination of a hardware apparatus and a computer. The apparatus described herein, or any components of the apparatus described herein, may be implemented at least in part in hardware and / or software.
[0289] The methods described herein may be performed using a hardware apparatus, or using a computer, or using a combination of a hardware apparatus and a computer.
[0290] Any of the methods described herein, or components of the apparatus described herein, may be implemented at least in part by hardware and / or software.
[0291] The above-described embodiments are merely illustrative of the principles of the present invention. It is understood that modifications and variations of the configurations and details described herein will be apparent to others skilled in the art. It is therefore intended to be limited only by the scope of the appended claims and not by the specific details presented as illustrative and explanatory of the embodiments herein.
[0292] References [1] Yuri Vatis, Bernd Edler, Dieu Thanh Nguyen, Joern Ostermann, “Two-dimensional non-separable Adaptive Wiener Interpolation Filter for H.264 / AVC”, ITU-T SG 16 / Q.6 Doc. VCEG-Z17, Busan, South Korea, April 2005.
[0293] [2] S. Wittmann, T. Wedi, “Separable Adaptive Interpolation Filter,” document COM16-C219, ITU-T, June 2007.
[0294] [3] M. Karczewicz, Y. Ye, and P. Chen, “Switched interpolation filter with offset,” document COM16-C463, ITU-T, Apr. 2008.
[0295] [4] Y. Ye and M. Karczewicz, “Enhanced Adaptive Interpolation Filter,” document COM16-C464, ITU-T, Apr. 2008.
[0296] [5] B. Bross, J. Chen, S. Liu (editors), “Versatile Video Coding (Draft 4),” document JVET-M1001, Jan. 2019.
Claims
1. encoding a syntax element comprising an indication of an Adaptive Motion Vector Resolution (AMVR) mode and an indication of whether an inter prediction mode for a coded block of a picture is translation or affine; selecting, in response to the indicated AMVR mode and the indication of whether the inter prediction mode is translational or affine, a finite impulse response (FIR) interpolation filter from a plurality of FIR interpolation filters to obtain a sample corresponding to a half-sample position of a reference picture for the coded block, the plurality of FIR interpolation filters comprising a first interpolation filter having six non-zero coefficients and a second interpolation filter having eight non-zero coefficients, the second interpolation filter having a greater edge preserving characteristic than the first interpolation filter; applying the selected FIR interpolation filter to samples of the reference picture to obtain the samples corresponding to the half-sample positions; predicting samples of the coded block using the samples corresponding to the half-sample positions; A video encoder configured to:
2. The video encoder of claim 1 , wherein the first interpolation filter has coefficients of 3, 9, 20, 20, 9, and 3.
3. 2. The video encoder of claim 1, wherein the second interpolation filter has coefficients of -1, 4, -11, 40, 40, -11, 4, and -1.
4. The video encoder of claim 1 , wherein the plurality of FIR interpolation filters comprises a plurality of filters having six non-zero coefficients.
5. decoding a syntax element from a video bitstream, the syntax element comprising an indication of an Adaptive Motion Vector Resolution (AMVR) mode and an indication of whether an inter prediction mode for a coded block of a picture is translation or affine; selecting, in response to the indicated AMVR mode and the indication of whether the inter prediction mode is translational or affine, a finite impulse response (FIR) interpolation filter from a plurality of FIR interpolation filters to obtain a sample corresponding to a half-sample position of a reference picture for the coded block, the plurality of FIR interpolation filters comprising a first interpolation filter having six non-zero coefficients and a second interpolation filter having eight non-zero coefficients, the second interpolation filter having a higher edge preserving property than the first interpolation filter; applying the selected FIR interpolation filter to samples of the reference picture to obtain the samples corresponding to the half-sample positions; predicting samples of the coded block using the samples corresponding to the half-sample positions; a video decoder configured to:
6. 6. The video decoder of claim 5, wherein the first interpolation filter has the following coefficients: 3, 9, 20, 20, 9, 3.
7. 6. The video decoder of claim 5, wherein the second interpolation filter has coefficients of -1, 4, -11, 40, 40, -11, 4, and -1.
8. The video decoder of claim 5 , wherein the plurality of FIR interpolation filters comprises a plurality of filters having six non-zero coefficients.
9. The video decoder of claim 5 , wherein none of the syntax elements are explicitly dedicated to indicating the selected FIR interpolation filter.
10. The video decoder of claim 5 , configured to select FIR interpolation filters separately for different coded units of the picture.
11. 1. A method of video encoding, comprising: encoding a syntax element comprising an indication of an Adaptive Motion Vector Resolution (AMVR) mode and an indication of whether an inter prediction mode for a coded block of a picture is translation or affine; selecting, in response to the indicated AMVR mode and the indication of whether the inter prediction mode is translational or affine, a finite impulse response (FIR) interpolation filter from a plurality of FIR interpolation filters to obtain a sample corresponding to a half-sample position of a reference picture for the coded block, the plurality of FIR interpolation filters comprising a first interpolation filter having six non-zero coefficients and a second interpolation filter having eight non-zero coefficients, the second interpolation filter having a greater edge preserving characteristic than the first interpolation filter; applying the selected FIR interpolation filter to samples of the reference picture to obtain the samples corresponding to the half-sample positions; predicting samples of the coded block using the samples corresponding to the half-sample positions; A method for providing the above.
12. The method of claim 11 , wherein the first interpolation filter has the following coefficients: 3, 9, 20, 20, 9, 3.
13. The method of claim 11, wherein the second interpolation filter has coefficients of -1, 4, -11, 40, 40, -11, 4, and -1.
14. The method of claim 11 , wherein the plurality of FIR interpolation filters comprises a plurality of filters having six non-zero coefficients.
15. 1. A method of video decoding, comprising: decoding a syntax element from a video bitstream, the syntax element comprising an indication of an Adaptive Motion Vector Resolution (AMVR) mode and an indication of whether an inter prediction mode for a coded block of a picture is translation or affine; selecting, in response to the indicated AMVR mode and the indication of whether the inter prediction mode is translational or affine, a finite impulse response (FIR) interpolation filter from a plurality of FIR interpolation filters to obtain a sample corresponding to a half-sample position of a reference picture for the coded block, the plurality of FIR interpolation filters comprising a first interpolation filter having six non-zero coefficients and a second interpolation filter having eight non-zero coefficients, the second interpolation filter having a higher edge preserving property than the first interpolation filter; applying the selected FIR interpolation filter to samples of the reference picture to obtain the samples corresponding to the half-sample positions; predicting samples of the coded block using the samples corresponding to the half-sample positions; A method for providing the above.
16. The method of claim 15 , wherein the first interpolation filter has the following coefficients: 3, 9, 20, 20, 9, 3.
17. 16. The method of claim 15, wherein the second interpolation filter has coefficients of -1, 4, -11, 40, 40, -11, 4, and -1.
18. The method of claim 15 , wherein the plurality of FIR interpolation filters comprises a plurality of filters having six non-zero coefficients.
19. The method of claim 15 , wherein none of the syntax elements are explicitly dedicated to indicating the selected FIR interpolation filter.
20. The method of claim 15 , further comprising selecting FIR interpolation filters separately for different coded units of the picture.
Citation Information
Patent Citations
Encoder, decoder, method, and video bitstream for hybrid video coding, and computer program product
JP7547590B2
