1. FAST MODE DECISION IN H264/AVC VIDEO CODEC NIRANJAN MULAY (0393251) CHEN GAO(0401840) (EL6123: PROJECT PRESENTATION) 05/06/2010

2. Outline: Introduction to H.264/AVC coding standard Mode decisions in H.264/AVC - Intra Block - Inter Block RDO algorithm and the need for FMD FMD (for Intra and Inter) Literature survey: edge-map based FMD Study of x264 code and encoding options Implementation: -Generation of MB mode statistics file from X264 -Visualize the modes in Matlab -Intra FMD; Inter FMD Summary and future work

3. Introduction to H.264/AVC Coding Standard The key features of H.264:  Improved Intra prediction: Directional spatial prediction  Enhanced Temporal Prediction: -Motion compensation with variable block sizes from 4x4 to 16x16: reduces ‘prediction error’ -Quarter-pel accurate motion estimation -Multiple reference for motion estimation -Weighted prediction (for B and P frames)  DCT-like integer transform: No mismatch between encoder and decoder

4. Introduction to H.264/AVC Coding Standard(Cntd)  Efficient entropy coding: -Uses arithmetic entropy coding, has option for VLC coding -Context adaptive entropy coding: 2 options – CAVLC and CABAC  Variable size (primarily 4x4 along with 8x8,16x16) transform: - Smaller size helps to represent a signal in locally adaptive manner which reduces ringing artifacts. - Generally high frequency=> 4x4 and low frequency=> 16x16  In-loop deblocking filter: Reduces blocking artifacts, improves quality.  Special Error Resilient Tools

5. H.264 Intra Modes:  Intra 4x4 : useful for a MB with significant detail  Intra 16x16 : good for coding very smooth areas (Intra 8x8 chroma: similar to intra 16x16)  I_PCM : no prediction or transform

6. ‘Intra 16x16’:  Mode 0 (vertical): extrapolation from upper samples.  Mode 1 (horizontal): extrapolation from left samples.  Mode 2 (DC): mean of upper and left-hand samples.  Mode 3 (Plane): plane prediction based on a linear spatial interpolation by using the upper and left-hand samples of the MB.

7. Intra 4x4’: ‘Intra 4x4’: Figure:4x4 luma prediction mode

8. Intra 4x4 Intra 4x4(Cntd):  Mode 0: Vertical  Mode 1: Horizontal  Mode 2: DC prediction  Mode 3: Diagonal down-left  Mode 4: Diagonal down-right  Mode 5: Vertical-right  Mode 6: Horizontal-down  Mode 7: Vertical-left  Mode 8: Horizontal-up

9. H.264 Inter Modes:  Hierarchical Decision  Level-1 (Partition): Compute RD-cost for: 16x16, 16x8, 8x16, 8x8.  Level-2 (Sub-Partition): If level-1 => 8x8, Then, compute RD cost of 8x4,4x8 and 4x4 Select the most optimal block!  P_Skip Mode

10. RDO Algorithm  Formula: RD_cost(s,c,MODE|Qp) = D +. R  ------------------------------------------------------------------------------  Computational Complexity of brute-force RDO:  INTRA block: Total Modes = 4 (16x16) + 9 (4x4) + 1 (I_PCM) + 4 (chroma_8x8) = 18 Total # of RDO calculations = M8 * ( M4*16 + M16) Theoretical Bound for a MB: 4 x (9x16+4)= 592!  INTER block: Total Modes = [ 7+1 (P_SKIP) ] + Intra counterparts HUGE Computations!! Problem for real time application => So, Need of FMD!

11. Intra : Edge-Histogram approach FMD-Intra : Edge-Histogram approach  Main Idea: Use Prediction in Edge Direction Generate edge map using Sobel operator Build edge direction histogram Fast intra mode decision

12. Generate Edge Map  Sobel Operator (Compute Gradients):

13. Edge Direction Histogram for Intra_4x4

14. FMD for Intra_4x4 Contd… As per observations in Reference[5]: - The ideal 4x4 mode is either the primary mode or one of the two neighboring modes - DC mode (Mode 2) is always evaluated - Total Modes = 1(Prime) + 2 (neighbors) + DC = 4

15. 6x16 Edge Direction Histogram for Intra_16x16 Total Modes = 1(Prime) + DC = 2

16. Inter Fast Mode Decision-Inter  Main idea: If we can reasonably decide that MB is temporally stationary or spatially homogeneous, we can encode MB using larger block-size and safely skip all other modes!

17. Stationary Region Determination  Refers to the stillness between consecutive frames in the temporal dimension  Evaluate Zero-MV Diff :  If (Diff “Stationary” So, choose16x16 mode and skip other sizes !  Threshold Ts = 200 (Reference[6])

18. Homogeneous Region Determination  Refers to texture similarities inside a single video frame  Edge amplitude computation is already done in fast intra mode decision  Threshold values (Reference[6]): for 16x16 block : 20000 for 8x8 block : 5000

19. Flow Chart of FMD_Inter

21. Wait... Changing the mode:Theory to Practice!

22. H.264/AVC Profiles  H264/AVC Profiles

23. Q. What is X264 ?  ‘x.264’ :  Open source H264/AVC encoder by VideoLAN  ‘C’ code library, Platform : Linux  Optimized as compared to reference JSVM software  Bunch of encoding options!  We finalized the options for “benchmarking” performance of Non- FMD vs FMD case E.g.: Command to encode ‘foreman_qcif.yuv’ sequence…./x264 -o foreman_qcif.264 foreman_qcif.yuv 176x144 -- profile baseline --frame 30 --verbose --keyint 15 --min- keyint 15 --no-scenecut --bframes 0 --ref 1 --slices 1 --fps 15 --qp 25 --partitions all --weightp 0 --me esa --subme 7 - -no-chroma-me --no-8x8dct --trellis 0 --no-fast-pskip -- visualize

24. X264 Coding Options:  --keyint 15/--min-keyint 15: Sets GOP size to 15  --bframes 0: Disables B-frame  --slices 1: Sets 1 slices per frame  --ref 1: Only 1 frame can be used as reference  --me esa: Select exhaustive motion estimation  --no-chroma-me: Ignore chroma in motion estimation  --qp 25: Fixed quantization step-size  --partitions all: Do all possible partitions  --no-scenecut: Disables adaptive I-frame decision

25. Implementation I: ‘Generation of Mode Statistics’  Intra MB: 3 Types :: I_4x4=0 ( 11 Modes), I_16x16=2 (4 Modes), I_PCM=3,  Inter MB: 3 Types :: P_L0=4, P_8x8=5, P_SKIP=6  P_LO (Level-1): can have 3 Partitions: D_16x8=14, D_8x16=15, D_16x16=16  P_8x8 (Level-2): has D_8x8 partition and can have 4 Sub-partitions: D_L0_8x8=3, D_L0_4x4=0, D_L0_8x4=1, D_L0_4x8=2

26. Implementation II: ‘Visualization Utility’ I-Frame RED : Intra_4x4 CYAN: Intra_16x16 P-Frame GREEN: P_SKIP BLUE: P_8X8 (and below) MAGENTA: P_16x16,P_16x8, P_8x16 Motive: “Seeing is Believing !” Let’s see a Demo…

27. Key observations:  I- Frame:  16x16 size chosen for spatially homogeneous region  4x4 size chosen for a MB with many spatial details/local edges -------------------------------------------------------------------------------------  P-Frame: % of Skipped% of Inter% of Intra Akiyo78.221.80 Football6.68112.4 Foreman17.581.90.6

28. Contd… Though H.264 allows variable size MC up-to 4x4 size…  Real world video sequences: Certain percentage of ‘Skipped’ blocks  Spatially Homogeneous regions gets best compensated with 16x16 (such blocks have similar motion; very seldom split to smaller blocks)  Temporally Stationary blocks ( e.g. stationary background even with strong edges) gets best compensated with 16x16 or P_SKIP  Nonetheless, Blocks containing motion boundaries or motion in smaller objects benefit from 8x8 or 4x4 MC

29. Implementation III: FMD Intra in x264 Block SizeTotal # of modes# of modes selected Luma(Y)4x494 Luma(Y)16x1642 Chroma(U,V)8x843 or 2  ~1000 lines of C code: Edge Map computation, Prime mode computation based on histogram, Modification of mode decision logic in.x264  Number of candidate modes in Intra-FMD:

30. Results: Intra FMD (All I frames, Qp=25) RESULTS △ TIME(%) △ PSNR_Y △ PSNR_U △ PSNR_V △ PSNR_AVG Mobile -30.22 -0.0220.0070.006 -0.016 Akiyo -35.81 -0.181-0.091 -0.165 Paris -39.15 -0.0670.006-0.02 -0.055 Foreman -38.88 -0.154-0.014-0.042 -0.125 Football -38.84 -0.084-0.066-0.068 -0.198 Avg. Time Saving: 36.70% Avg. PSNR drop: 0.11 dB

31. Results: Intra FMD (PSNR vs R) Sequence: Mobile, Coding: All I, Qp= 37,33,29,25 Avg PSNR drop: 0.044 dB, Avg. Increase in R: ~6%, Avg Time Saving: 37.51%

32. Summary and future work: To Conclude:  Learnt x264 code-flow, different encoding options  Matlab ‘mode visualization script’ is ready  Intra-FMD ready, Inter-FMD (in progress)  Important: FMD framework is ready! Different FMD algorithms can be plugged in to evaluate prime mode selection… Future Work:  Inter FMD  FMD enhancement: Analysis of different modes with conditional probabilistic model

33. Reference  [1] URL: http://www.videolan.org/developers/x264.html  [2] Thomas Wiegand, Gary J Sullivan, “Overview of the H264/AVC Video Coding Standard”, IEEE Transactions on Circuits and Systems for Video Technology, Vol. 13, No. 7,July 2003  [3]URL: http://www.vcodex.com/files/H.264_overview.pdf White Paper: An Overview of H.264 Advanced Video Codinghttp://www.vcodex.com/files/H.264_overview.pdf  [4] Iain E G Richardson, “H.264 and MPEG4 Video Compression”, WILEY Publications, 2003  [5] Feng Pan et al, “Fast Mode Decision Algorithm for Intra-prediction in H264/AVC Video Coding”, IEEE Transactions on Circuits and Systems for Video Technology, Vol. 15, No. 7,July 2005  [6] D. Wu et al, “Fast Intermode Decision in H264/AVC Video Coding”, IEEE Transactions on Circuits and Systems for Video Technology, Vol. 15, No. 6,July 2005  [7] Rui Su, Guizhong Liu, Tongyu Zhang,”Fast Mode Decision Algorithm for Intra Prediction In H264/AVC”, ICASSP-2006