In another example, the parameters of the SAO in HEVC Draft 8 are as follows: 1) a filter type that is 0 if no SAO is to be applied, 1 for BO type filtering, and 2 for EO type filtering; 2) four offset values that are the absolute values of the actual offset values; 3) the signs of each non-zero offset value (for BO type filtering); 4) a band position (for BO type filtering) indicating the left-most band of four consecutive bands where the offsets are to be applied; and 5) an EO class (for EO type filtering) that indicates the edge direction (See Table 3). The signs for the offset values for EO type filtering are inferred from the category, i.e., plus for categories 1 and 2, and minus for categories 3 and 4.

TABLE 3

EO Class
Edge Direction

0
1D 0-degree edge offset

1
1D 90-degree edge offset

2
1D 135-degree edge offset

3
1D 45-degree edge offset

Unless otherwise explicitly stated herein, embodiments of the invention are not limited to any particular SAO parameter set/SAO implementation.

As used herein in describing various embodiments, the term “SAO information” refers to the data that is encoded in the video bit stream to communicate the SAO parameters to a decoder. The “SAO information” may be different in different embodiments. For example, SAO information may be the directly encoded SAO parameters, one or more prediction indicators, or a combination there of.

FIG. 5 shows a block diagram of a digital system that includes a source digital system 500 that transmits encoded video sequences to a destination digital system 502 via a communication channel 516. The source digital system 500 includes a video capture component 504, a video encoder component 506, and a transmitter component 508. The video capture component 504 is configured to provide a video sequence to be encoded by the video encoder component 506. The video capture component 504 may be, for example, a video camera, a video archive, or a video feed from a video content provider. In some embodiments, the video capture component 504 may generate computer graphics as the video sequence, or a combination of live video, archived video, and/or computer-generated video.

The video encoder component 506 receives a video sequence from the video capture component 504 and encodes it for transmission by the transmitter component 508. The video encoder component 506 receives the video sequence from the video capture component 504 as a sequence of pictures, divides the pictures into largest coding units (LCUs), and encodes the video data in the LCUs. The video encoder component 506 may be configured to perform SAO filtering and SAO parameter signaling during the encoding process as described herein. An embodiment of the video encoder component 506 is described in more detail herein in reference to FIG. 6.

The transmitter component 508 transmits the encoded video data to the destination digital system 502 via the communication channel 516. The communication channel 516 may be any communication medium, or combination of communication media suitable for transmission of the encoded video sequence, such as, for example, wired or wireless communication media, a local area network, or a wide area network.

The destination digital system 502 includes a receiver component 510, a video decoder component 512 and a display component 514. The receiver component 510 receives the encoded video data from the source digital system 500 via the communication channel 516 and provides the encoded video data to the video decoder component 512 for decoding. The video decoder component 512 reverses the encoding process performed by the video encoder component 506 to reconstruct the LCUs of the video sequence. The video decoder component 512 may be configured to perform SAO filtering during the decoding process as described herein. An embodiment of the video decoder component 512 is described in more detail below in reference to FIG. 8.

The reconstructed video sequence is displayed on the display component 514. The display component 514 may be any suitable display device such as, for example, a plasma display, a liquid crystal display (LCD), a light emitting diode (LED) display, etc.

In some embodiments, the source digital system 500 may also include a receiver component and a video decoder component and/or the destination digital system 502 may include a transmitter component and a video encoder component for transmission of video sequences both directions for video steaming, video broadcasting, and video telephony. Further, the video encoder component 506 and the video decoder component 512 may perform encoding and decoding in accordance with one or more video compression standards. The video encoder component 506 and the video decoder component 512 may be implemented in any suitable combination of software, firmware, and hardware, such as, for example, one or more digital signal processors (DSPs), microprocessors, discrete logic, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), etc.

FIG. 6 shows a block diagram of the LCU processing portion of an example video encoder. A coding control component (not shown) sequences the various operations of the LCU processing, i.e., the coding control component runs the main control loop for video encoding. The coding control component receives a digital video sequence and performs any processing on the input video sequence that is to be done at the picture level, such as determining the coding type (I, P, or B) of a picture based on the high level coding structure, e.g., IPPP, IBBP, hierarchical-B, and dividing a picture into LCUs for further processing.

In addition, for pipelined architectures in which multiple LCUs may be processed concurrently in different components of the LCU processing, the coding control component controls the processing of the LCUs by various components of the LCU processing in a pipeline fashion. For example, in many embedded systems supporting video processing, there may be one master processor and one or more slave processing modules, e.g., hardware accelerators. The master processor operates as the coding control component and runs the main control loop for video encoding, and the slave processing modules are employed to off load certain compute-intensive tasks of video encoding such as motion estimation, motion compensation, intra prediction mode estimation, transformation and quantization, entropy coding, and loop filtering. The slave processing modules are controlled in a pipeline fashion by the master processor such that the slave processing modules operate on different LCUs of a picture at any given time. That is, the slave processing modules are executed in parallel, each processing its respective LCU while data movement from one processor to another is serial.

The LCU processing receives LCUs of the input video sequence from the coding control component and encodes the LCUs under the control of the coding control component to generate the compressed video stream. The LCUs in each picture are processed in row order. The CUs in the CU structure of an LCU may be processed by the LCU processing in a depth-first Z-scan order. The LCUs 600 from the coding control unit are provided as one input of a motion estimation component 620, as one input of an intra-prediction component 624, and to a positive input of a combiner 602 (e.g., adder or subtractor or the like). Further, although not specifically shown, the prediction mode of each picture as selected by the coding control component is provided to a mode selector component and the entropy encoder 634.

The storage component 618 provides reference data to the motion estimation component 620 and to the motion compensation component 622. The reference data may include one or more previously encoded and decoded pictures, i.e., reference pictures.

The motion estimation component 620 provides motion data information to the motion compensation component 622 and the entropy encoder 634. More specifically, the motion estimation component 620 performs tests on CUs in an LCU based on multiple inter-prediction modes (e.g., skip mode, merge mode, and normal or direct inter-prediction), PU sizes, and TU sizes using reference picture data from storage 618 to choose the best CU partitioning, PU/TU partitioning, inter-prediction modes, motion vectors, etc. based on a rate distortion coding cost. To perform the tests, the motion estimation component 620 may divide an LCU into CUs according to the maximum hierarchical depth of the quadtree, and divide each CU into PUs according to the unit sizes of the inter-prediction modes and into TUs according to the transform unit sizes, and calculate the coding costs for each PU size, prediction mode, and transform unit size for each CU.

The motion estimation component 620 provides the motion vector (MV) or vectors and the prediction mode for each PU in the selected CU partitioning to the motion compensation component 622 and the selected CU/PU/TU partitioning with corresponding motion vector(s), reference picture index (indices), and prediction direction(s) (if any) to the entropy encoder 634.

The motion compensation component 622 provides motion compensated inter-prediction information to the mode decision component 626 that includes motion compensated inter-predicted PUs, the selected inter-prediction modes for the inter-predicted PUs, and corresponding TU sizes for the selected CU partitioning. The coding costs of the inter-predicted CUs are also provided to the mode decision component 626.

The intra-prediction component 624 provides intra-prediction information to the mode decision component 626 and the entropy encoder 634. More specifically, the intra-prediction component 624 performs intra-prediction in which tests on CUs in an LCU based on multiple intra-prediction modes, PU sizes, and TU sizes are performed using reconstructed data from previously encoded neighboring CUs stored in the buffer 628 to choose the best CU partitioning, PU/TU partitioning, and intra-prediction modes based on a rate distortion coding cost. To perform the tests, the intra-prediction component 624 may divide an LCU into CUs according to the maximum hierarchical depth of the quadtree, and divide each CU into PUs according to the unit sizes of the intra-prediction modes and into TUs according to the transform unit sizes, and calculate the coding costs for each PU size, prediction mode, and transform unit size for each PU. The intra-prediction information provided to the mode decision component 626 includes the intra-predicted PUs, the selected intra-prediction modes for the PUs, and the corresponding TU sizes for the selected CU partitioning. The coding costs of the intra-predicted CUs are also provided to the mode decision component 626. The intra-prediction information provided to the entropy encoder 634 includes the selected CU/PU/TU partitioning with corresponding intra-prediction modes.

The mode decision component 626 selects between intra-prediction of a CU and inter-prediction of a CU based on the intra-prediction coding cost of the CU from the intra-prediction component 624, the inter-prediction coding cost of the CU from the inter-prediction component 620, and the picture prediction mode provided by the mode selector component. Based on the decision as to whether a CU is to be intra- or inter-coded, the intra-predicted PUs or inter-predicted PUs are selected, accordingly.

The output of the mode decision component 626, i.e., the predicted PUs, is provided to a negative input of the combiner 602 and to a delay component 630. The associated transform unit size is also provided to the transform component 604. The output of the delay component 630 is provided to another combiner (i.e., an adder) 638. The combiner 602 subtracts each predicted PU from the original PU to provide residual PUs to the transform component 604. Each resulting residual PU is a set of pixel difference values that quantify differences between pixel values of the original PU and the predicted PU. The residual blocks of all the PUs of a CU form a residual CU block for the transform component 604.

The transform component 604 performs block transforms on the residual CU to convert the residual pixel values to transform coefficients and provides the transform coefficients to a quantize component 606. More specifically, the transform component 604 receives the transform unit sizes for the residual CU and applies transforms of the specified sizes to the CU to generate transform coefficients.

The quantize component 606 quantizes the transform coefficients based on quantization parameters (QPs) and quantization matrices provided by the coding control component and the transform sizes. The quantized transform coefficients are taken out of their scan ordering by a scan component 608 and arranged sequentially for entropy coding. In essence, the coefficients are scanned backward in highest to lowest frequency order until a coefficient with a non-zero value is located. Once the first coefficient with a non-zero value is located, that coefficient and all remaining coefficient values following the coefficient in the highest to lowest frequency scan order are serialized and passed to the entropy encoder 634.

The entropy encoder 634 entropy encodes the relevant data, i.e., syntax elements, output by the various encoding components and the coding control component to generate the compressed video bit stream that is provided to a video buffer 636 for transmission or storage. The syntax elements are encoded according to the syntactical order specified in HEVC. This syntactical order specifies the order in which syntax elements should occur in a compressed video bit stream. Among the syntax elements that are encoded are flags indicating the CU/PU/TU partitioning of an LCU, the prediction modes for the CUs, and the ordered quantized transform coefficients for the CUs. The entropy encoder 634 also codes relevant data from the in-loop filtering component 616 such as the LCU specific SAO information for each LCU. The LCU SAO information is signaled on an LCU-by-LCU basis rather than all together in slice headers as in the prior art. In some embodiments, the SAO information for an LCU may immediately precede the LCU data in the encoded bit stream as shown in FIG. 4. In some embodiments, the SAO information may immediately follow the LCU data in the encoded bit stream.

The LCU processing includes an embedded decoder. As any compliant decoder is expected to reconstruct an image from a compressed bit stream, the embedded decoder provides the same utility to the video encoder. Knowledge of the reconstructed input allows the video encoder to transmit the appropriate residual energy to compose subsequent pictures. To determine the reconstructed input, i.e., reference data, the ordered quantized transform coefficients for a CU provided via the scan component 608 are returned to their original post-transform arrangement by an inverse scan component 610, the output of which is provided to a dequantize component 612, which outputs a reconstructed version of the transform result from the transform component 604.

The dequantized transform coefficients are provided to the inverse transform component 614, which outputs estimated residual information representing a reconstructed version of a residual CU. The inverse transform component 614 receives the transform unit size used to generate the transform coefficients and applies inverse transform(s) of the specified size to the transform coefficients to reconstruct the residual values.

The reconstructed residual CU is provided to the combiner 638. The combiner 638 adds the delayed selected CU to the reconstructed residual CU to generate a reconstructed CU, which becomes part of reconstructed picture data. The reconstructed picture data is stored in a buffer 628 for use by the intra-prediction component 624 and is provided to an in-loop filter component 616.

The in-loop filter component 616 applies various filters to the reconstructed picture data to improve the quality of the reference picture data used for encoding/decoding of subsequent pictures. FIG. 7 shows the in-loop filter component 616 in more detail. The filters in the in-loop filter component 616 include a deblocking filter 704 and a SAO filter 706. The in-loop filter component 616 may apply these filters, for example, on an LCU-by-LCU basis. The two filters may be applied sequentially as shown in FIG. 7. That is, the deblocking filter 704 may be first applied to the reconstructed data. Then, the SAO 706 may be applied to the deblocked reconstructed picture data. Referring again to FIG. 6, the final filtered reference picture data is provided to storage component 618.

The deblocking filter 704 operates to smooth discontinuities at block boundaries, i.e., TU and CU block boundaries, in a reconstructed picture. In general, for each LCU of the reconstructed picture, the SAO filter 706 determines the best offset values, i.e., band offset values or edge offset values, to be added to pixels of that LCU to compensate for intensity shift that may have occurred during the block based coding of the picture, applies the offset values to the reconstructed LCU, and determines the SAO information to be encoded in the bit stream for the LCU. The operation of embodiments of the SAO filter 706 of the in-loop filter component 616 is described in more detail herein in reference to the method of FIG. 10.

FIG. 8 shows a block diagram of an example video decoder. The video decoder operates to reverse the encoding operations, i.e., entropy coding, quantization, transformation, and prediction, performed by the video encoder of FIG. 6 to regenerate the pictures of the original video sequence. In view of the above description of a video encoder, one of ordinary skill in the art will understand the functionality of components of the video decoder without detailed explanation.

The entropy decoding component 800 receives an entropy encoded (compressed) video bit stream and reverses the entropy coding to recover the encoded PUs and header information such as the prediction modes, the encoded CU and PU structures of the LCUs, and the LCU specific SAO information for each LCU. If the decoded prediction mode is an inter-prediction mode, the entropy decoder 800 then reconstructs the motion vector(s) as needed and provides the motion vector(s) to the motion compensation component 810.

The inverse quantization component 802 de-quantizes the quantized transform coefficients of the residual CU. The inverse transform component 804 transforms the frequency domain data from the inverse quantization component 802 back to the residual CU. That is, the inverse transform component 804 applies an inverse unit transform, i.e., the inverse of the unit transform used for encoding, to the de-quantized residual coefficients to produce the residual CUs.

A residual CU supplies one input of the addition component 806. The other input of the addition component 806 comes from the mode switch 808. When an inter-prediction mode is signaled in the encoded video stream, the mode switch 808 selects predicted PUs from the motion compensation component 810 and when an intra-prediction mode is signaled, the mode switch selects predicted PUs from the intra-prediction component 814.

The motion compensation component 810 receives reference data from storage 812 and applies the motion compensation computed by the encoder and transmitted in the encoded video bit stream to the reference data to generate a predicted PU. That is, the motion compensation component 810 uses the motion vector(s) from the entropy decoder 800 and the reference data to generate a predicted PU.

The intra-prediction component 814 receives reconstructed samples from previously reconstructed PUs of a current picture from the buffer 807 and performs the intra-prediction computed by the encoder as signaled by an intra-prediction mode transmitted in the encoded video bit stream using the reconstructed samples as needed to generate a predicted PU.

The addition component 806 generates a reconstructed CU by adding the predicted PUs selected by the mode switch 808 and the residual CU. The output of the addition component 806, i.e., the reconstructed CUs, supplies the input of the in-loop filter component 816 and is also stored in the buffer 807 for use by the intra-prediction component 814.

The in-loop filter component 816 applies the same filters to the reconstructed picture data as the encoder, i.e., a deblocking filter and SAO, in the same order to improve the quality of the reconstructed picture data. The output of the in-loop filter component 816 is the decoded pictures of the video bit stream. Further, the output of the in-loop filter component 816 is stored in storage 812 to be used as reference data by the motion compensation component 810.

FIG. 9 shows the in-loop filter component 816 in more detail. The filters in the in-loop filter component 816 include a deblocking filter 904 and a SAO filter 906. The deblocking filter 904 operates in the same manner as the deblocking filter of the encoder. In general, for each reconstructed LCU, the SAO filter 906 applies the offset values determined by the encoder for the LCU to the pixels of the LCU. More specifically, the SAO filter 906 receives decoded LCU specific SAO information from the entropy decoding component 800 for each reconstructed LCU, determines the SAO parameters for the LCU from the SAO information, and applies the determined offset values to the pixels of the LCU according to values of other parameters in the SAO parameter set. The operation of embodiments of the SAO filter 906 of the in-loop filter component 816 for each LCU is described in more detail herein in reference to the method of FIG. 12.

FIG. 10 is a flow diagram of a method for SAO filtering and SAO parameter signaling that may be performed in a video encoder, e.g., the encoder of FIG. 6. In general, in this method, SAO parameters are determined for each LCU in a picture, SAO filtering is performed on each LCU according to the SAO parameters determined for the LCU, and SAO information for each LCU is encoded in the bit stream interleaved with the LCU data. In an encoder, method steps 1000 and 1002 may be performed by an SAO filter, e.g., SAO 706 of FIG. 7, and method step 1004 may be performed by an entropy encoder, e.g., entropy encoder 634 of FIG. 6.

Referring now to FIG. 10, SAO parameters are determined 1000 for LCUs in a reconstructed picture (after a deblocking filter is applied). That is, SAO parameters are determined for each LCU in the reconstructed picture. Any suitable technique may be used for determining the LCU SAO parameters. For example, a region-based approach for determining SAO parameters that uses quadtree partitioning such as the one previously described herein may be used. In such an approach, the same SAO parameters are used for all LCUs in a region. In another example, a fixed picture partitioning that is LCU-aligned may be used in which SAO parameters are determined for each partition. The same SAO parameters would apply to all LCUs in a partition. In another example, SAO parameters may be determined individually for each LCU.

Once the SAO parameters for an LCU are determined, the SAO information to be encoded in the bit stream for that LCU is also determined. In some embodiments, the actual values of the SAO parameters determined for an LCU are the SAO information for the LCU. As is explained in more detail below, in some embodiments, a prediction protocol may be used to predict some or all of the SAO parameters for an LCU in order to reduce the size of the SAO information encoded in the bit stream. In such embodiments, the SAO information for an LCU is determined according to the prediction protocol.

SAO filtering is then performed 1002 on the reconstructed picture according to the SAO parameters determined for the LCUs. More specifically, SAO filtering is performed on each LCU according to the particular SAO parameters determined for that LCU. In general, the SAO filtering applies the specified offsets in the SAO parameters to pixels in the LCU according to the filter type indicated in the SAO parameters.

The LCU specific SAO information for each LCU is also entropy coded 1004 into the compressed bit stream on an LCU by LCU basis, i.e., the LCU specific SAO information is interleaved with the LCU data rather than being encoded in the slice header as in the prior art. In some embodiments, the SAO information for an LCU may immediately precede the LCU data in the encoded bit stream as shown in the example of FIG. 4. In some embodiments, the SAO information may immediately follow the LCU data in the encoded bit stream.

Direct encoding of the SAO parameters for each LCU may increase the encoded bit stream size over the prior art in which SAO parameters for regions were coded instead. To reduce the number of bits needed for encoding SAO parameters, rather than encoding the values of some or all of the LCU SAO parameters, in some embodiments of the method of FIG. 10, the values may be predicted according to an agreed upon prediction protocol between the encoder and the decoder and prediction indicators encoded in lieu of the actual values to reduce the number of bits to be encoded. In essence, depending on the prediction protocol, one or prediction indicators indicate to the decoder how an actual parameter value (or more than one actual parameter value) is to be determined.

In some embodiments, offset values are predicted based on previously encoded offset values and the differences (deltas) signaled as prediction indicators for the offset values. In such embodiments, the SAO information for an LCU includes the prediction indicators in lieu of the offset values. In some such embodiments, a prediction indicator for an offset value is the delta between the offset value and the previously encoded offset value. More specifically, the offset values for an LCU may be signaled in a known order. The actual value of the first offset value is encoded. The absolute value of this first offset value is then used as the predictor for the second offset value. The difference between the absolute value of the first offset value and the actual second offset value is determined, and this delta is encoded as the prediction indicator for the second offset value. The absolute value of this delta is then used as the predictor for the third offset value. The difference between the absolute value of the delta and the actual value of the third offset value is determined, and this delta is encoded as the prediction indicator for the third offset value. This process is repeated for each of the remaining offset values. For example, depending on the particular SAO implementation, there may be 16 SAO offsets to be signaled (see Table 2). The agreed upon order for signaling these offsets may be according to band order from lowest band number to highest band number. If the offsets are 2, 2, 4, 5, 6, 6, 7, 7, 3, 12, 2, 2, 3, 6, 6, 7, these offset values would be signaled as deltas (prediction indicators) 2, 0, 2, 1, 1, 0, 1, 0, −4, 9, 10, 0, 1, 3, 0, 1.

In some embodiments, a prediction protocol for LCU SAO parameters may be used in which an LCU is “merged” with a neighboring LCU in the same slice, i.e., a spatially neighboring LCU, such that the LCU shares the SAO parameters of the neighboring LCU. In general, the SAO parameters of the neighboring LCU serve as the predictors for the SAO parameters of the LCU if the SAO parameters of the LCU and the neighboring LCU are the same. In such a prediction protocol, a merge indicator, e.g., a merge flag, is encoded for each LCU to indicate whether or not merging is used. If merging is used, the SAO parameters of the LCU are not encoded. If merging is not used, the actual SAO parameters of the LCU are encoded in addition to the merge indicator.

Different schemes may be used to identify a neighboring LCU for possible merging. For example, in some embodiments, a single neighboring LCU, e.g., the immediate left neighboring LCU or the immediate top neighboring LCU, may be considered for merging with the current LCU. If the SAO parameters of the current LCU and the single neighboring LCU are the same, the merge indicator is set to indicate merging of the current LCU and is encoded in the bit stream. The actual SAO parameters for the current LCU need not be signaled as the decoder can use the SAO parameters of the neighboring LCU. If the SAO parameters of the current LCU and the single neighboring LCU are not the same or the neighboring LCU is not available, e.g., the current LCU is at a left or top slice boundary, the merge indicator is set to indicate no merging and the merge indicator and the other SAO parameters of the current LCU are encoded in the bit stream.

In some embodiments, multiple spatially neighboring LCUs, i.e., two or more neighboring LCUs, may be considered for merging with the current LCU. In such embodiments, the SAO parameters of the current LCU are compared to the SAO parameters of each of the spatially neighboring LCU candidates (if available) in a specified order, and the first neighboring LCU with matching parameters, if any, is selected for merging with the current LCU. If a match is found, an identifier for the selected neighboring LCU, e.g., an index, is encoded in the bit stream along with the merge indicator set to indicate merging. If no match is found, a merge indicator set to indicate no merging is encoded in the bit stream along with the other SAO parameters of the current LCU.

For example, in some embodiments, the four spatially neighboring LCUs indicated in FIG. 5 may be considered for merging with the current LCU. In this figure, the X denotes the current LCU and the numbered blocks are the spatially neighboring LCUs to be considered. Note that depending on the location of the current LCU in the slice, one or more of these spatially neighboring LCUs may not be available. For example, if the current LCU is at the left boundary of the slice, neighboring LCU 0 and neighboring LCU 2 will not be available to be considered for merging with the current LCU. The number inside each block indicates the order in which the neighboring LCUs are to be considered for merging with the current LCU. This number also serves as an index to identify the neighboring LCU to the decoder if that LCU is selected for merging. In some embodiments, a subset of these spatially neighboring LCUs may be used. For example, only LCU 0 (the immediate left neighboring LCU) and LCU 1 (the immediate top neighboring LCU) may be considered for merging.

In some embodiments, a prediction protocol for LCU SAO parameters may be used in which an LCU is “merged” with a temporally co-located LCU such that the LCU shares the SAO parameters of the co-located LCU. In general, a co-located LCU or temporally co-located LCU is a square area in a reference picture having the same coordinates, size, and shape of an LCU in a picture currently being encoded or decoded. For purposes of merging, the temporally co-located LCU for a current LCU is a square area in a reference picture having the same coordinates, size, and shape of the current LCU. If the SAO parameters of the current LCU and the co-located LCU are the same, the merge indicator is set to indicate merging of the current LCU and is encoded in the bit stream. The actual SAO parameters for the current LCU need not be signaled as the decoder can use the SAO parameters of the co-located LCU. If the SAO parameters of the current LCU and the co-located LCU are not the same or the co-located LCU is not available, the merge indicator is set to indicate no merging and the merge indicator and the other SAO parameters of the current LCU are encoded in the bit stream.

In some embodiments, a prediction protocol for LCU SAO parameters may be used in which multiple spatially neighboring LCUS and the co-located LCU are considered for merging with the current LCU. In some such embodiments, the SAO parameters of the current LCU are compared to the SAO parameters of each of the spatially neighboring LCU candidates and the co-located LCU (if available) in a specified order, and the first LCU candidate with matching parameters, if any, is selected for merging with the current LCU. If a match is found, an identifier for the selected candidate LCU, e.g., an index, is encoded in the bit stream along with the merge indicator set to indicate merging. If no match is found, a merge indicator set to indicate no merging is encoded in the bit stream along with the other SAO parameters of the current LCU.

For example, in some embodiments, the spatially neighboring LCUs may be the LCUs indicated in FIG. 5 with the ordering as indicated in FIG. 5 and the co-located LCU may be assigned an index of 4, making it the last LCU to be considered. An SAO merge candidate list is constructed by considering each of the five candidate LCUs in the order specified. If a candidate LCU is available, it is added as the next entry in the SAO merge candidate list. Thus, if all five candidate LCUs are available, the ordered candidate LCUs in the SAO merge list will be LCU 0, LCU 1 LCU 2, LCU 3, and the temporally co-located LCU, LCU 4. If, for example, LCU 0 and LCU 2 are not available, the ordered candidate LCUs in the SAO merge list will be LCU 1, LCU 3, and LCU 4.

The SAO parameters are then compared to the SAO parameters of each candidate LCU in the SAO merge candidate list in order. The first candidate LCU in the list having the same SAO parameters, if any, is selected for merging with the current LCU. If a match is found, an identifier for the selected LCU candidate, e.g., the index of the location of the selected LCU candidate in the SAO merge candidate list, is encoded in the bit stream along with the merge indicator set to indicate merging. If no match is found, a merge indicator set to indicate no merging is encoded in the bit stream along with the SAO parameters of the current LCU. In some embodiments, if only one of the candidate LCUs was available, i.e., only one candidate LCU was in the constructed SAO merge candidate list, and the SAO parameters matched, the LCU candidate identifier is not encoded in the bit stream, i.e., only a merge indicator set to indicate merging is encoded.

In some embodiments, a prediction protocol for LCU SAO parameters may be used in which the parameters of multiple reference LCUs are combined to predict the SAO parameters of the current LCU. The reference LCUs may be spatially neighboring LCUs or may be spatially neighboring LCUs and the temporally co-located LCU. On the encoder side, an SAO filter type for the current LCU is selected as the one occurring most frequently in the parameters of the reference LCUs. If two or more filter types occur with the same frequency, the one with the smaller filter type index is chosen. The value of each offset for the current LCU is then computed as the average value of that offset in each of the reference LCUs having the filter type selected. The encoder then uses some criteria, e.g., coding cost, to decide whether these predicted parameters or the actual SAO parameters of the LCU should be used. If the predicted parameters are chosen, the merge indicator is set to indicate merging and is encoded in the bit stream. If the actual SAO parameters are chosen, the merge indicator is set to indicate no merging and is encoded in the bit stream along with the actual SAO parameters. In addition, the predicted parameters would be used to apply SAO filtering to the LCU in the encoder rather than the actual SAO parameters Note that on the decoder side, if merging is indicated, the decoder would repeat the process used in the encoder to determine the predicted parameters.

FIG. 12 is a flow diagram of a method for SAO filtering that may be performed in a video decoder, e.g., the decoder of FIG. 8. In general, in this method, the SAO information for an LCU encoded in a compressed video bit stream as per an embodiment of the method of FIG. 10 is entropy decoded, the SAO parameters for the LCU are determined from the decoded SAO information, and SAO filtering is performed on the reconstructed LCU according to the determined SAO parameters. In a video decoder, method steps 1202 and 1204 may be performed by an SAO filter, e.g., SAO 906 of