(19) United States
`(12) Patent Application Publication (10) Pub. No.: US 2007/0009047 A1
`(43) Pub. Date:
`Jan. 11, 2007
`Shim et al.
`
`US 20070009047A1
`
`(54) METHOD AND APPARATUS FOR HYBRID
`ENTROPY ENCOOING AND DECODING
`
`(75) Inventors: Woo-sung Shim, Yongin-si (KR):
`Austin Lobo, Suwon-si (KR)
`Correspondence Address:
`SUGHRUE MION, PLLC
`2100 PENNSYLVANIA AVENUE, N.W.
`SUTE 8OO
`WASHINGTON, DC 20037 (US)
`(73) Assignee: SAMSUNG ELECTRONICS CO.,
`LTD.
`
`(21) Appl. No.:
`
`11/482,811
`
`(22) Filed:
`
`Jul. 10, 2006
`
`(30)
`
`Foreign Application Priority Data
`
`Jul. 8, 2005 (KR)............................ 10-2005-0061779
`
`Publication Classification
`
`(51) Int. Cl.
`(2006.01)
`H04N 7/2
`(52) U.S. Cl. ........................................................ 375/240.26
`(57)
`ABSTRACT
`An apparatus and method for entropy encoding and decod
`ing of video data. The method for entropy encoding video
`data includes dividing the video data into at least one partial
`Video data portion and determining an entropy encoding
`mode to be used for each of the partial video data portion
`between at least two different entropy encoding modes and
`performing entropy encoding on each of the partial video
`data portions using the determined entropy encoding mode.
`
`CURRENTPCTURE
`F
`
`
`
`
`
`108
`
`110
`
`112
`
`150
`
`FORMATION
`UNIT
`
`QUANTIZATION
`UNIT
`
`REARRANGEMENT
`UNIT
`
`ENTROPY
`ENCODER
`
`MOTION
`ESMAION
`
`100 -
`
`PREDCTION
`UNIT
`
`
`
`
`
`FRAME
`MEMORY
`
`
`
`RECONSTRUCTED
`PICTURE F.
`
`INVERSE
`TRANS
`FORMATION
`UNIT
`
`INVERSE
`OUANTIZATION
`UNIT
`
`
`
`
`
`
`
`
`
`
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 1
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 1 of 8
`
`US 2007/0009047 A1
`
`FIG. 1 (PRIOR ART)
`
`
`
`10
`
`SLICE
`
`20
`
`MBTYPE
`
`MBPRED CODED ERROR DATA
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 2
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 2 of 8
`
`US 2007/0009047 A1
`
`ÅdOH|NE
`
`HEC]OONE
`
`& '0IJI
`
`ESHEAN||
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 3
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 3 of 8
`
`US 2007/0009047 A1
`
`FIG. 3
`
`160 SYNTAX
`ELEMENT
`
`
`
`
`
`VIDEO DATA
`CONTROLLER
`
`
`
`SYNTAX
`ELEMENT
`
`BITSTREAM
`
`FIG. 4
`
`
`
`2OO
`
`21 O
`
`220
`
`HEADER
`
`FIRST DATA REGION SECOND DATA REGION
`
`DATA REGION
`START POSITION (A)
`
`OFFSET
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 4
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 4 of 8
`
`US 2007/0009047 A1
`
`
`
`
`
`
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 5
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 5 of 8
`
`US 2007/0009047 A1
`
`FIG 6
`-
`
`18O
`
`181
`
`CONTEXT
`STORING
`UNIT
`
`
`
`SYNTAX
`ELEMENT
`
`BTSTREAM
`
`FIG 7
`
`
`
`
`
`
`
`
`
`
`
`DETERMINE ENTROPY ENCODING
`MODE OF VIDEO DATA
`
`300
`
`PERFORMENTROPY ENCODING
`USING DETERMINED ENTROPY
`ENCODING MODE
`
`3O2
`
`STORE POSITION INFORMATION OF
`ENTROPY ENCODED DATA IN
`HEADE OF DATA
`
`304
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 6
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 6 of 8
`
`US 2007/0009047 A1
`
`
`
`7070099 | 7
`
`9 "DIJI
`
`Ul
`H
`4 C.
`
`
`
`
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 7
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 7 of 8
`
`US 2007/0009047 A1
`
`FIG. 9
`
`510
`
`BITSTREAM
`
`CONTROLLER
`
`500
`
`SYNTAXELEMENT
`
`SYNTAX ELEMENT
`
`FIG 10
`
`
`
`
`
`
`
`RECEIVE BITSTREAM AND DETERMINE
`ENTROPY DECODING MODE
`
`6OO
`
`PERFORMENTROPY DECODING USNG
`DETERMINED ENTROPY DECODING MODE
`
`610
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 8
`
`

`

`Patent Application Publication Jan. 11, 2007 Sheet 8 of 8
`
`US 2007/0009047 A1
`
`FIG 11
`
`
`
`START
`
`
`
`YE S
`
`IS SYNTAX
`ELEMENT TO BE
`DECODED IS FIRST SYNTAX
`ELEMENT2
`
`700
`
`
`
`PERFORMINITIALIZATION
`
`NO
`
`
`
`
`
`720
`IS SYNTAX ELEMENT
`ENTROPY ENCODED
`USING CABACP
`
`PERFORMENTROPY DECODING
`USING CABAC
`
`PERFORMENTROPY
`DECODING USING CAVLC
`
`740
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 9
`
`

`

`US 2007/0009047 A1
`
`Jan. 11, 2007
`
`METHOD AND APPARATUS FOR HYBRD
`ENTROPY ENCOOING AND DECODING
`
`CROSS-REFERENCE TO RELATED PATENT
`APPLICATION
`0001. This application claims the benefit of Korean
`Patent Application No. 10-2005-0061779, filed on Jul. 8,
`2005, in the Korean Intellectual Property Office, the disclo
`sure of which is incorporated herein in its entirety by
`reference.
`
`BACKGROUND OF THE INVENTION
`
`0002)
`1. Field of the Invention
`0003 Methods and apparatuses consistent with the
`present invention relate to video compression encoding and
`decoding, and more particularly, to entropy encoding and
`decoding of video data, in which different entropy coding
`methods are used for video compression.
`0004 2. Description of the Related Art
`0005 Conventionally, since computation power or
`memory for processing multimedia data and a transmission
`bandwidth are limited, complexity is limited and research
`aims at developing an optimal video processing technique in
`a specific bitrate area. However, as the development of
`hardware and Software for processing multimedia has been
`accelerated and channel bandwidth for servicing multimedia
`has been increased and diversified, there has emerged a
`demand for an encoding technique having higher encoding
`efficiency and adaptive to various network environments. To
`meet the demand, the Video Coding Expert Group (VCEG)
`affiliated with the International Telecommunications Union
`Telecommunication Standardization Sector (ITU-T) and the
`Moving Picture Expert Group (MPEG) of the International
`Organization for Standardization (ISO)/International Elec
`trotechnical Commission (IEC) jointly have established the
`video encoding standard called H.264 or MPEG-4 Part 10
`advanced video coding (AVC). H.264 improves encoding
`efficiency by 50% or more in comparison to existing stan
`dards such as the MPEG-4 Part 2 visual codec and considers
`error robustness and network friendliness to cope with the
`rapidly changing wireless environment and Internet envi
`rOnment.
`0006 H.264 uses context adaptive variable length coding
`(hereinafter, will be referred to as CAVLC) and context
`adaptive binary arithmetic coding (hereinafter, will be
`referred to as CABAC) as entropy encoding methods for
`compressing video data.
`0007 FIG. 1 illustrates a slice syntax according to con
`ventional H.264.
`0008 Referring to FIG. 1, the slice syntax includes a slice
`header 10 and slice data 20.The slice header 10 defines a
`slice type and an encoded picture including slices and may
`include a command associated with reference picture man
`agement. The slice data 20 includes information about coded
`macroblocks (MB) and/or skipped (non-encoded) macrob
`locks (skip run). In addition, each macroblock includes
`header components and coded error data.
`0009. According to H.264, in upper layers of a slice layer,
`Syntax elements are encoded into a fixed or variable length
`binary code. In the slice layer or lower layers thereof, syntax
`
`elements are encoded using variable length coding (VLC) or
`CABAC according to an entropy encoding mode.
`0010) If entropy coding mode flag is 0, residual data is
`encoded using CAVLC and other variable-length coded
`units are encoded using an Exp-Golomb code. If entropy
`coding mode flag is 1, H.264 syntax elements are encoded
`and decoded using CABAC. In other words, according to the
`value of entropy coding mode flag, the slice data 20 is
`encoded using one of entropy encoding methods CAVLC
`and CABAC.
`0011 Context modeling with respect to the other syntax
`elements except for the residual data, e.g., mb type indicat
`ing, a macroblock type, coded block pattern indicating
`coded data pattern, and mb qp delta indicating a quantizer
`parameter (QP) transmitted as a delta value from a previous
`QP value is complicated. In other words, the other syntax
`elements except for the residual data occupy a smaller
`portion of the entire video data than the residual data, but
`they increase the complexity of context modeling when
`being encoded using CABAC. On the other hand, context
`modeling of the residual data is simpler than that of the other
`Syntax elements.
`0012 However, according to conventional H.264, slice
`data is encoded using only one of CAVLC or CABAC and
`variation in the complexity of entropy encoding according to
`characteristics of syntax elements is not considered.
`
`SUMMARY OF THE INVENTION
`0013 The present invention provides an apparatus and
`method for entropy encoding and decoding, in which the
`complexity of entropy encoding for video compression is
`reduced and the compression rate of entropy encoding is
`improved.
`0014. According to one aspect of the present invention,
`there is provided a method for entropy encoding video data.
`The method includes dividing the video data into at least one
`partial video data portion and determining an entropy encod
`ing mode to be used for each of the partial video data
`portions from between at least two different entropy encod
`ing modes and performing entropy encoding on each of the
`partial video data portions using the determined entropy
`encoding mode.
`00.15 According to another aspect of the present inven
`tion, there is provided an apparatus for entropy encoding
`Video data. The apparatus includes a controller which
`divides the video data into at least one partial video data
`portion and determines an entropy encoding mode to be used
`for each of the partial video data portions from between
`different entropy encoding modes and an entropy encoding
`unit which performs entropy encoding on each of the partial
`Video data portions according to the determined entropy
`encoding mode.
`0016. According to still another aspect of the present
`invention, there is provided a method for entropy decoding
`an input bitstream that is entropy encoded using at least one
`entropy encoding modes. The method includes dividing the
`input bitstream into bitstream data portions according to the
`entropy encoding modes and determining an entropy decod
`ing mode for each of the bitstream data portions and
`performing entropy decoding on each of the bitstream data
`portions using the determined entropy decoding mode.
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 10
`
`

`

`US 2007/0009047 A1
`
`Jan. 11, 2007
`
`0017 According to yet another aspect of the present
`invention, there is provided an apparatus for entropy decod
`ing an input bitstream that is entropy encoded using at least
`one entropy encoding mode. The apparatus includes a con
`troller which divides the input bitstream into bitstream data
`portions according to the entropy encoding modes and
`determines an entropy decoding mode for each of the
`bitstream data portions and an entropy decoding unit which
`performs entropy decoding on each of the bitstream data
`portions using the determined entropy decoding mode.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`0018. The above and other aspects of the present inven
`tion will become more apparent by describing in detail
`exemplary embodiments thereof with reference to the
`attached drawings in which:
`0.019
`FIG. 1 illustrates a slice syntax according to con
`ventional H.264;
`0020 FIG. 2 is a block diagram of a video encoder using
`an entropy encoder according to an exemplary embodiment
`of the present invention;
`0021
`FIG. 3 is a block diagram of an entropy encoder
`according to an exemplary embodiment of the present
`invention;
`0022 FIG. 4 illustrates data encoded by an entropy
`encoder according to an exemplary embodiment of the
`present invention;
`0023 FIG. 5 is a block diagram of a context adaptive
`binary arithmetic coding (CABAC) unit of FIG. 3;
`0024 FIG. 6 is a block diagram of a context adaptive
`variable length coding (CAVLC) unit of FIG. 3;
`0.025
`FIG. 7 is a flowchart illustrating an entropy encod
`ing method according to an exemplary embodiment of the
`present invention;
`0026 FIG. 8 is a block diagram of a video decoder using
`an entropy decoder according to an exemplary embodiment
`of the present invention;
`0027 FIG. 9 is a block diagram of an entropy decoder
`according to an exemplary embodiment of the present
`invention;
`0028 FIG. 10 is a flowchart illustrating an entropy
`decoding method according to an exemplary embodiment of
`the present invention; and
`0029 FIG. 11 is a flowchart illustrating a method of
`decoding a bitstream using an entropy decoding method
`according to an exemplary embodiment of the present
`invention.
`
`DETAILED DESCRIPTION OF THE
`INVENTION
`0030 First, terms used in the following description will
`be defined.
`0031 CABAC: context adaptive binary arithmetic cod
`ing
`0032 CAVLC: context adaptive variable length coding
`
`& G.
`
`0033 UVLC: universal variable length code. In particu
`lar, an Exp Golumb code as a kind of UVLC is used in
`H.264 video codec.
`0034) I PCM: mb type in which sampled values without
`being predicted, transformed, or quantized are directly trans
`mitted. In CABAC, when mb type is I PCM, all I PCM
`data is read from arithmetically encoded slices and a binary
`arithmetic coder is restarted.
`0035) MB: macroblock
`0036) CABAC offset: term that is newly defined in the
`present invention. CABAC offset is a pointer indicating a
`slice data portion that is CABAC entropy coded from the
`beginning of slice data. CABAC offset is stored in a slice
`header.
`0037 ae(v): syntax element that is context-adaptive arith
`metic entropy-coded
`0038 me(v): syntax element that is mapped Exp
`Golomb-coded
`0039 se(v): syntax element that is signed integer Exp
`Golomb-coded
`0040 te(v): syntax element that is truncated Exp
`Golomb-coded
`0041 u(n): unsigned integer using n bits. When n is “v
`on a syntax table, the number of bits changes according to
`other syntax elements. Parsing of u(n) is specified by the
`return value of a function read bit(n) analyzed by the binary
`expression of an unsigned integer in which a most signifi
`cant bit (MSB) is first written.
`0042 read bit(n): function which reads n bits of a bit
`stream and moves a bitstream pointer by the read in bits.
`When n is 0, read bit(n) returns 0 and does not move the
`bitstream pointer.
`0043 ue(v): syntax element that is unsigned integer Exp
`Golomb-coded
`0044 f(n): a fixed pattern bit string written using n bits.
`Parsing of f(n) is specified by the return value of the function
`read bit(n).
`0045 codIRange and codIOffset: variables indicating the
`range of an arithmetic coding engine and an offset state,
`respectively.
`0046 FIG. 2 is a block diagram of a video encoder using
`an entropy encoder according to an exemplary embodiment
`of the present invention.
`0047 Referring to FIG. 2, a video encoder 100 may
`include a motion estimation unit 102, a motion compensa
`tion unit 104, an intraprediction unit 106, a transformation
`unit 108, a quantization unit 110, a rearrangement unit 112,
`an entropy encoder 150, an inverse quantization unit 116, an
`inverse transformation unit 118, a filter 120, and a frame
`memory 122.
`0048. The motion estimation unit 102, the motion com
`pensation unit 104, and the intraprediction unit 106 predict
`a current picture to be encoded using an intra mode or an
`inter mode, and the transformation unit 108 transforms a
`difference Dn between a predicted picture and the current
`picture into a frequency domain. The quantization unit 110
`quantizes transformed data.
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 11
`
`

`

`US 2007/0009047 A1
`
`Jan. 11, 2007
`
`0049. The entropy encoder 150 according to an exem
`plary embodiment of the present invention removes a sta
`tistically redundant component of video data and generates
`a compressed bitstream or file that can be transmitted or
`stored. The compressed bitstream includes a coded motion
`vector parameter, coded residual data, and header informa
`tion. In particular, the entropy encoder 150 according to the
`present invention generates compressed data bitstreams cor
`responding to syntax elements using at least one different
`entropy encoding method according to the types of the
`syntax elements. The entropy encoder 150 according to the
`present invention performs entropy encoding on video data
`having high correlation using CABAC and on video data
`having low correlation using CAVLC. For example, when
`Syntax elements according to H.264 are entropy coded, the
`entropy encoder 150 performs entropy encoding on all
`Syntax elements included in residual data and a syntax
`element end of slice flag using CABAC and on the remain
`ing syntax elements using CAVLC. This is because the
`residual data visually have higher correlation than other
`Syntax elements.
`0050 FIG. 3 is a block diagram of the entropy encoder
`150 according to an exemplary embodiment of the present
`invention.
`0051 Referring to FIG. 3, the entropy encoder 150
`according to the present invention may include a controller
`160 and an entropy encoding unit 165. The entropy encoding
`unit 165 includes a CABAC unit 170 and a CAVLC unit 180.
`0.052 The controller 160 receives video data output from
`a video encoder and determines which entropy encoding
`mode is to be used to entropy-encode the received video data
`according to the type of the received video data.
`0053 According to H.264, a residual block and motion
`estimation information are encoded by the CABAC unit 170
`or the CAVLC unit 180 and output in the form of a final
`bitstream. Here, the residual block indicates a block having
`a difference between pixel values of the original block and
`pixel values of a reconstructed block as a pixel value. The
`reconstructed block is obtained by calculating a motion
`vector through motion estimation with respect to the original
`block and reconstructing the original block using the calcu
`lated motion vector.
`0054 As a result of video encoding according to H.264,
`after discrete cosine transformation (DCT) is performed on
`a residual macroblock in units of 4x4 residual blocks, the
`controller 160 receives DCT coefficients of the 4x4 residual
`blocks. The controller 160 generates a plurality of syntax
`elements for each of the 4x4 residual blocks, determines
`whether to perform CABAC or CAVLC according to the
`type of each of the generated syntax elements, and outputs
`the generated syntax elements to the CABAC unit 170 or the
`CAVLC unit 180.
`0055. The CABAC unit 170 performs CABAC on all
`Syntax elements included in residual data and a syntax
`element end of slice flag under the control of the controller
`160.
`0056. The CAVLC unit 180 performs CAVLC on the
`remaining syntax elements except for the syntax elements
`included in the residual data and the syntax element end
`of slice flag.
`
`0057) Syntax elements entropy-encoded by the CAVLC
`unit 180 are as follows.
`0058 Mb field decoding is entropy encoded using u(1),
`mb type using ue(v), transform size 8x8 flag using u(1),
`coded block pattern using me(v), mb qp delta using se(v),
`prev intra4x4 pred mode flag using u(1), rem intra4x
`4 pred mode using u(3), intra chroma pred mode using
`u(V), ref idx 10 using te(v), ref idx 11 using te(v), mVd
`10 using se(v), mvd 11 using se(v), and Sub mb type using
`ue(v). In other words, the syntax elements that are conven
`tionally encoded using ae(v), i.e., that are context adaptive
`binary arithmetic encoded, are entropy encoded using
`CAVLC such as u(n), ue(v), me(v), and se(v). This is
`because the remaining syntax elements except for residual
`data occupies a small portion of the entire video data, but
`they increase the complexity of context modelling when
`being entropy encoded using CABAC and thus, they are
`entropy encoded using CAVLC having lower complexity.
`0059. In addition, mb skip flag is encoded using u(1)
`instead of ae(v). When mb type is I PCM, a syntax element
`pcm alignment Zero bit is entropy encoded using f(1), a
`Syntax element pcm sample luma is entropy encoded using
`u(V), and a syntax element pcm sample chroma is entropy
`encoded using u(V).
`0060. In brief, the entropy encoder 150 according to an
`exemplary embodiment of the present invention performs
`CABAC on syntax elements associated with residual data
`that require simple context modelling and performs CAVLC
`on syntax elements that require complicated context mod
`elling, i.e., syntax elements including additional information
`of the residual data, thereby improving compression effi
`ciency while reducing complexity in encoding.
`0061
`FIG. 4 illustrates data encoded by the entropy
`encoder 150 according to an exemplary embodiment of the
`present invention.
`0062 Referring to FIG. 4, data that are entropy encoded
`by the CABAC unit 170 and CAVLC unit 180 are stored in
`a first data region 210 and a second data region 220,
`respectively. Here, it is assumed that data that is entropy
`encoded by the CAVLC unit 180 is stored in the first data
`region 210 and data that is entropy encoded by the CABAC
`unit 170 is stored in the second data region 220. The
`controller 160 may store a difference of location between a
`start position A of the first data region 210 and a start
`position of the second data region 220 as offset information
`in a header 200. When entropy-encoded data is decoded
`according to the present invention, entropy decoding may be
`performed using the offset information stored in the header
`200 or an entropy decoding method that is predetermined
`according to the type of a syntax element to be decoded.
`0063 FIG. 5 is a block diagram of the CABAC unit 170
`of FIG. 3.
`0064. The CABAC unit 170 according to an exemplary
`embodiment of the present invention may include a binarizer
`171, a context modeler 172, a regular coding engine 173,
`and a bypass coding engine 174.
`0065. If a nonbinary valued syntax element is input, the
`binarizer 171 outputs a bin String by mapping the syntax
`element to a binary valued sequence.
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 12
`
`

`

`US 2007/0009047 A1
`
`Jan. 11, 2007
`
`Predetermined bins selected from the output bin
`0.066
`string or binary valued syntax elements (hereinafter, the bin
`string or the binary valued syntax element will be referred to
`as a bin) do not pass through the context modeler 172, but
`are encoded into a bitstream by the bypass coding engine
`174 to increase the processing speed of encoding. The
`remaining bins are input to the context modeler 172.
`0067. The context modeler 172 determines a probability
`model required to encode a currently input bin based on the
`currently input bin and a previously encoded syntax ele
`ment.
`0068 The regular coding engine 173 receives the bins
`that are input after passing through the context modeler 172
`and information about the probability model determined by
`the context modeler 172 and finally generates and outputs a
`bitstream.
`0069. The CABAC unit 170 may be used in a decoder,
`and in this case, a signal processing order may be inverse to
`the encoding process order.
`0070 FIG. 6 is a block diagram of the CAVLC unit 180
`of FIG. 3.
`0071 Motion information, prediction mode information,
`and transformation coefficient information included in the
`received syntax element are input to the CAVLC unit 182.
`The CAVLC unit 182 applies a variable length encoding
`table to the input syntax element and outputs a bitstream.
`0072 A context storing unit 181 stores information about
`a bitstream encoded by the CAVLC unit 182 and changes a
`variable length encoding table based on the information to
`allow entropy encoding to be adaptively performed.
`0073. The CAVLC unit 180 may be used in a decoder,
`and in this case, an input bitstream is processed in an order
`inverse to the encoding processing order.
`0074 FIG. 7 is a flowchart illustrating an entropy encod
`ing method according to an exemplary embodiment of the
`present invention.
`0075) Referring to FIG. 7, it is determined which entropy
`encoding mode is to be used for video data in operation 300.
`As mentioned above, when a syntax element according to
`H.264 is entropy-encoded, all syntax elements included in
`residual data and the syntax element end of slice flag are
`entropy encoded using CABAC and the remaining syntax
`elements indicating information about the residual data are
`entropy encoded using CAVLC.
`0.076
`Next, entropy encoding is performed according to
`the determined entropy encoding mode in operation 302. To
`discriminate positions of data that are entropy encoded using
`different entropy encoding modes, difference of location
`between the start position of entropy encoded data using
`CAVLC and a start position of the entropy encoded data
`using CABAC can be stored in a header of the data in
`operation 304.
`0077. The entropy encoder according to an exemplary
`embodiment of the present invention is not limited to that
`shown in FIG. 3, but may include a plurality of different
`entropy encoding modules for different entropy encoding
`modes and perform entropy encoding using the entropy
`encoding methods.
`
`0078 FIG. 8 is a block diagram of a video decoder using
`an entropy decoder according to an exemplary embodiment
`of the present invention.
`0079 Referring to FIG. 8, a video decoder 400 according
`to an exemplary embodiment of the present invention may
`include an entropy decoder 500, a rearrangement unit 404,
`an inverse quantization unit 406, an inverse transformation
`unit 408, a motion compensation unit 410, an intraprediction
`unit 412, a filter 414, and an adder 416.
`0080. The entropy decoder 500 receives and entropy
`decodes a compressed bitstream to generate syntax elements
`and quantized coefficients X. The inverse quantization unit
`406 and the inverse transformation unit 408 perform inverse
`quantization and inverse transformation on the decoded
`Syntax elements and the quantized coefficients X to extract
`transformation encoding coefficients, motion vector infor
`mation, and residual information. The motion compensation
`unit 410 and the intraprediction unit 412 generate a predic
`tion block using decoded information according to an
`encoded picture type. The adder 416 adds an error value D',
`to the prediction block to generate uF" which passes through
`the filter 414, thereby generating a reconstructed picture F.
`0081
`FIG. 9 is a block diagram of the entropy decoder
`500 according to an exemplary embodiment of the present
`invention.
`0082) Referring to FIG. 9, the entropy decoder 500 may
`include a controller 510 and an entropy decoding unit 520.
`The entropy decoding unit 520 may include at least one
`module for performing entropy decoding according to
`applied entropy encoding modes. In FIG. 9, the entropy
`decoding unit 520 includes a context adaptive binary arith
`metic decoding (hereinafter, will be referred to as CABAD)
`unit 521 and a context adaptive variable length decoding
`(hereinafter, will be referred to as CAVLD) unit 523.
`0083) The controller 510 determines which entropy
`decoding mode is to be used to entropy-decode current video
`data to be decoded. At this time, an entropy decoding mode
`may be predetermined according to the type of the current
`Video data or an entropy encoding mode applied to the
`current video data may be determined by extracting offset
`information stored in a header of entropy-decoded data. The
`controller 510 outputs a bitstream to the entropy decoding
`unit 520 according to the determined entropy decoding
`mode. For example, a bitstream entropy encoded using
`CABAC is output to the CABAD unit 521 and a bitstream
`entropy encoded using CAVLC is output to the CAVLD unit
`523.
`0084. The CABAD unit 521 and the CAVLD unit 523
`perform entropy decoding on input bitstreams using corre
`sponding entropy decoding modes and output syntax ele
`mentS.
`0085 FIG. 10 is a flowchart illustrating an entropy
`decoding method according to an exemplary embodiment of
`the present invention.
`0086) Referring to FIG. 10, a bitstream is received and it
`is determined which entropy decoding mode is to be used to
`entropy decode the received bitstream in operation 600.
`0087 As described above, the entropy decoding mode
`can be determined by offset information stored in a header
`
`Unified Patents, LLC v. Elects. & Telecomm. Res. Inst., et al.
`
`Ex. 1029, p. 13
`
`

`

`US 2007/0009047 A1
`
`Jan. 11, 2007
`
`of the bitstream or the entropy decoding mode can be
`predetermined according to the type of a syntax element.
`0088. In operation 610, entropy decoding is performed
`according to the determined entropy decoding mode,
`thereby outputting syntax elements corresponding to the
`bitstream.
`0089. To decode a bitstream encoded using at least one of
`the entropy encoding modes, using a plurality of pointers
`indicating data regions encoded using different entropy
`encoding modes, data pointed by a pointer is read and the
`pointer is updated to entropy-decode data of the bitstream,
`which is entropy-encoded using different entropy encoding
`modes.
`0090 FIG. 11 is a flowchart illustrating a method of
`decoding a bitstream encoded using CABAC and CAVLC
`according to an exemplary embodiment of the present
`invention
`0091 Referring to FIGS. 9 and 11, the controller 510
`determines whether a syntax element of an input bitstream
`to be decoded is the first syntax element in operation 700.
`0092) If the syntax element to be decoded is the first
`syntax element, initialization is performed in operation 710.
`0093. The initialization may include initialization for
`context parameters and initialization of an arithmetic decod
`ing engine. The initialization for context parameters is
`initiated when parsing of slice data starts. Initialization of
`ctXidx is the same as that in decoding of conventional H.264.
`However, ctxidx from 0 to 84 and from 399 to 401 that are
`conventionally entropy encoded using CABAC are not used.
`This is because syntax elements indicated by ctxidx from 0
`to 84 and from 399 to 401 are entropy encoded using
`CAVLC. Except for the syntax elements entropy encoded
`using CAVLC, context parameters are initialized for syntax
`elements indicated by ctxidx from 85 to 398 and from 402
`to 459. In addition, the initialization of the arithmetic
`decoding engine is initiated when parsing of slice data starts
`and syntax elements pcm alignment Zero bit, pcm
`sample luma, and pcm sample chroma of an I PCM-type
`macroblock are decoded.
`0094) The controller 510 determines which entropy
`encoding mode between CABAC and CAVLC is used to
`entropy encode a syntax element in operation 720 and
`performs entropy decoding using CABAD in operation 730
`or entropy decoding using CAVLD in operation 740.
`0.095 As described above, according to exemplary
`embodiments of the present invention, compression rate can
`be improved when compared to conventional CAVLC and
`complexity in computation can be reduced when compared
`to conventional CABAC.
`0096. By using different entropy encoding and decoding
`modes according to the type of video data, compression rate
`of the video data is improved and complexity in computation
`required for entropy encoding and decoding is reduced.
`0097. The present invention can also be embodied as a
`computer-readable code on a computer-readable recording
`medium. The computer-readable recording medium is any
`data storage device that can store data which can be there
`after read by a computer system. Examples of the computer
`readable recording medium include, but are not limited to,
`
`read-only memory (ROM), random-access memory (RAM),
`CD-ROMs, magnetic tapes, floppy disks, optical data Stor
`age devices, and carrier waves. The computer-readable
`recording medium can also be distributed over network
`coupled computer systems so that the computer-readable
`code is stored and executed in a distributed fashion.
`0098. While the present invention has been particularly
`shown and described with reference to an exemplary
`embodiment thereof, it will be understood by those of
`ordinary skill in the art that various changes in form and
`details may be made therein without departing from the
`spirit and scope of the present invention as defined by the
`following claims.
`
`What is claimed is:
`1. A method for entropy encoding video data, the method
`comprising:
`dividing the video data into partial video data portions and
`determining an entropy encoding mode to be used for
`each of the partial video data portions between at