Digital Video Compression Fundamentals and Standards docx

Digital Video Compression Digital Video Compression Fundamentals and Standards Fundamentals and Standards Speaker: Wei-Yi Wei Speaker: Wei-Yi Wei Advisor: Prof. Jian-Jung Ding Advisor: Prof. Jian-Jung Ding Digital Image and Signal Processing Digital Image and Signal Processing Lab Lab GICE, National Taiwan University GICE, National Taiwan University 2008/12/26 Digital Video Compression Fundamentals an d Standards 2 Outline Outline Introduction Introduction Video Compression Standards Video Compression Standards Simulation Reference Software Simulation Reference Software Future Work and Conclusions Future Work and Conclusions 2008/12/26 Digital Video Compression Fundamentals an d Standards 3 Outline Outline Introduction Introduction Video Compression Standards Video Compression Standards Simulation Reference Software Simulation Reference Software Future Work and Conclusions Future Work and Conclusions 2008/12/26 Digital Video Compression Fundamentals an d Standards 4 Introduction (1/2) Introduction (1/2) Why video compression technique is important ? Why video compression technique is important ? One movie video without compression One movie video without compression 720 x 480 pixels per frame 720 x 480 pixels per frame 30 frames per second 30 frames per second Total 90 minutes Total 90 minutes Full color Full color The total quantity of data = 167.96 G Bytes !! 2008/12/26 Digital Video Compression Fundamentals an d Standards 5 Introduction (2/2) Introduction (2/2) What is the difference between video compression and image What is the difference between video compression and image compression? compression? Temporal Redundancy Temporal Redundancy Coding method to remove redundancy Coding method to remove redundancy Intraframe Coding Intraframe Coding Remove spatial redundancy Remove spatial redundancy Interframe Coding Interframe Coding Remove temporal redundancy Remove temporal redundancy 2008/12/26 Digital Video Compression Fundamentals an d Standards 6 The most intuitive method to remove The most intuitive method to remove Spatiotemporal redundancy Spatiotemporal redundancy 3-Dimensional DCT 3-Dimensional DCT Remove spatiotemporal correlation Remove spatiotemporal correlation Good for low motion video Good for low motion video Bad for high motion video Bad for high motion video 1 1 1 3 0 0 0 8 (2 1) (2 1) (2 1) ( , , ) ( ) ( ) ( ) ( , , ) cos cos cos 2 2 2 N N N t x y x u y v t w F x y t C u C v C w x y t N N N N π π π − − − = = = + + +       = Ψ             ∑∑∑ for 0, , 1 , 0, , 1 and 0, , 1 1/ 2 for 0 where 8 and ( ) 1 otherwise u N v N w N k N C k = − = − = −  =  = =    2008/12/26 Digital Video Compression Fundamentals an d Standards 7 The most popular method to remove The most popular method to remove temporal redundancy temporal redundancy The Block-Matching Algorithm The Block-Matching Algorithm 2008/12/26 Digital Video Compression Fundamentals an d Standards 8 Matching Function Matching Function The dissimilarity between two blocks and The dissimilarity between two blocks and The matching criteria The matching criteria Mean square error (MSE) Mean square error (MSE) High precision is needed High precision is needed Mean absolute difference (MAD) Mean absolute difference (MAD) Low precision is enough Low precision is enough 2 ( , ) ( - )M u v u v= ( , ) | - |M u v u v= 1 1 , 1 ( , ) [ ( , ), ( , )] p q n n x y Vy Vx D s t M x y x V y V − = = = Ψ Ψ + + ∑ ∑ n ψ 1n− ψ ( , )D s t 2008/12/26 Digital Video Compression Fundamentals an d Standards 9 The Exhaustive Block-Matching Algorithm Reference Frame Current Frame Search Range Motion Vector 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 11 22 33 44 55 66 77 88 12 22 33 44 55 66 77 88 12 22 33 44 55 66 77 88 12 22 33 44 55 66 77 88 12 22 33 44 55 66 77 88 11 23 34 44 55 66 77 88 11 23 34 44 55 66 77 88 11 23 34 44 55 66 77 88 11 23 34 44 55 66 77 88 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 0 0 |A|=12 2008/12/26 Digital Video Compression Fundamentals an d Standards 10 Fast Block-Matching Algorithms Fast Block-Matching Algorithms EBMA needs Intensive computation EBMA needs Intensive computation Fast Algorithm is needed Fast Algorithm is needed Find the possible local optimal Find the possible local optimal [ ] [ ] [ ] 720 480 128 128 16 16 30 170 G instructions/sec 16 16   × × × × × × =     Reference Frame Current Frame [...]... 2008/12/26 Digital Video Compression Fundamentals an 14 Multiresolution Motion Estimation (2/3) Variable block size method 2008/12/26 Digital Video Compression Fundamentals an 15 Multiresolution Motion Estimation (3/3) 2008/12/26 Digital Video Compression Fundamentals an 16 Outline Introduction Video Compression Standards Simulation Reference Software Future Work and Conclusions 2008/12/26 Digital Video Compression. .. Work and Conclusions 2008/12/26 Digital Video Compression Fundamentals an 17 The Development of Video Compression Standards 2008/12/26 Digital Video Compression Fundamentals an 18 The MPEG-1 Standard Group of Pictures Motion Estimation Motion Compensation Differential Coding DCT Quantization Entropy Coding 2008/12/26 Digital Video Compression Fundamentals an 19 Group of Pictures (1/2) I-frame (Intracoded... Target frame Digital Video Compression Fundamentals an 25 The MPEG-2 Standard Field/Frame DCT Coding Field/Frame Prediction Mode Selection Alternative Scan Order Various Picture Sampling Formats User Defined Quantization Matrix 2008/12/26 Digital Video Compression Fundamentals an 26 Progressive Scan and Interlaced Scan Progressive Scan 2008/12/26 Interlaced Scan Digital Video Compression Fundamentals. .. 2-D logarithm Search Method Three Steps Search Method 2008/12/26 Digital Video Compression Fundamentals an 11 2-D logarithm Search Method 1 1 2 1 1 2 4 2 4 3 4 1 4 4 3 2008/12/26 Digital Video Compression Fundamentals an 12 Three Step Search Method 2 1 1 2 2 2 2 3 3 3 1 3 2 3 3 3 3 2 2 1 1 1 2008/12/26 1 1 1 Digital Video Compression Fundamentals an 13 Multiresolution Motion Estimation (1/3) The number... referred to as inter-frame 2008/12/26 Digital Video Compression Fundamentals an 20 Group of Pictures (2/2) The distance between two nearest P-frame or P-frame and I-frame denoted by M The distance between two nearest I-frames denoted by N Forward Motion Compensation GOP I 2008/12/26 B B P B B N=9 M=3 P B B Bidirectional Motion Compensation Digital Video Compression Fundamentals an I 21 The MPEG-1 Encoder... 4:2:0 2008/12/26 Digital Video Compression Fundamentals an 30 The MPEG-2 Encoder (1/2) SNR Enhanced Layer + Base Layer Bits Enhance Basic quality requirement For SDTV Enhanced Layer High quality service For HDTV Entropy Coding Q Q-1 + DCT Entropy Coding Q Q-1 Q-1 + Motion Compensation + Bits Base Base Layer IDCT Motion Estimation + + Frame Memory 2008/12/26 Digital Video Compression Fundamentals an... Estimation CAVLC 2008/12/26 Digital Video Compression Fundamentals an 32 Variable Block Size The fixed block size may not be suitable for all motion objects Improve the flexibility of comparison Reduce the error of comparison 7 types of blocks for selection 16 x 16 16 x 8 0 0 8x8 1 8x4 0 2008/12/26 0 1 8 x 16 0 1 4x8 0 1 8x8 0 1 2 3 4x4 0 1 2 3 Digital Video Compression Fundamentals an 33 Multiple... the target frame in many situations 2008/12/26 Digital Video Compression Fundamentals an 34 Integer Transform for Reducing The Spatial Redundancy (1/2) The transform matrix C 1 1  2 1 C=  1 −1  4×4 Block Size  1 −2 Separable Integer Transform 1 ÷ −1 −2 ÷ −1 1 ÷ ÷ 2 −1  1 The transform coefficients are CXCT 2008/12/26 Digital Video Compression Fundamentals an 35 Integer Transform for Reducing... Progressive Scan 2008/12/26 Interlaced Scan Digital Video Compression Fundamentals an 27 Field/Frame DCT Coding The field type DCT Fast motion video The frame type DCT Slow motion video Field DCT Coding 2008/12/26 Luminance MB Frame DCT Coding Digital Video Compression Fundamentals an 28 Alternative Scan Order Zigzag scan order Frame DCT Alternative scan order Field DCT 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1... Memory 2008/12/26 Digital Video Compression Fundamentals an 22 The MPEG-1 Encoder (2/4) Differential Coding ^ D(t ) = ψ (t ) − ψ (t ) ψ (t ) is the input image ^ is the predictive image ψ (t ) DCT N −1 N −1 2  π (2 x + 1)u   π (2 y + 1)v  F (u , v) = C (u )C (v)∑∑ f ( x, y ) cos  cos    N 2N 2N     x =0 y =0 for u = 0, , N − 1 and v = 0, , N − 1 1/ 2 for k = 0  where N = 8 and C (k ) =  . Conclusions Future Work and Conclusions 2008/12/26 Digital Video Compression Fundamentals an d Standards 3 Outline Outline Introduction Introduction Video Compression Standards Video Compression Standards Simulation. Conclusions Future Work and Conclusions 2008/12/26 Digital Video Compression Fundamentals an d Standards 18 The Development of Video The Development of Video Compression Standards Compression Standards 2008/12/26 Digital. Digital Video Compression Digital Video Compression Fundamentals and Standards Fundamentals and Standards Speaker: Wei-Yi Wei Speaker: Wei-Yi Wei Advisor: