Performance Analysis and Comparison of H.264 and VP6Siddhartha Mukkamala (1000571314)([email protected])Under guidance of Dr.K.R.RaoVP6 codec was first introduced in May 2003 and was officiallyreleased in October 2003. It is a proprietary video codecdeveloped by On2 Technologies as a successor to earlier effortssuch as VP3 and VP5.Can provide higher visual quality when using at lower bitratesMacromedia has licensed this for its Flash family products.VP6 is the preferred video compression format for use withFlash Player 8 and higher.What is VP6 ?Only two frame typesIntra, or I-frames, may be reconstructed from their compressedrepresentation with no reference to other frames in the sequence. I-framesprovide entry points into the bitstream that do not require precedingframes to be decoded providing a method for fast random access .Inter, prediction or P-frames, are encoded differentially with respect to apreviously encoded reference frame in the sequence. This reference framemay either be the reconstruction of the immediately previous frame in thesequence or a stored previous frame known as the Golden Frame.The alternative prediction, or Golden Frame, is a frame buffer that bydefault holds the last decoded I-frame but it may be updated at any time. Aflag in the frame header indicates to the decoder whether or not to updatethe Golden Frame buffer.Frames in VP6Two Different Profiles - Simple and AdvancedEach Frame header contains a flag ,Vp6Profile,which indicates the profile that was used to code it.In both profiles the BoolCoder is used for encoding .The BoolCoder is a simplified binary arithmetic coder allowingtokens to be encoded with fractions of a bit. It is much moreefficient in terms of compression performance than the Huffmancoder, but this comes with a significantly increased computationalcomplexity.Coding Profiles in VP6H.264 was developed by ITU-T Video Coding Experts Grouptogether with the ISO/IEC Moving Picture Experts Group in 2003.Key video compression (codec) scheme in MPEG-4 part 10 format for digital media exchange.H.264 provides better image quality compared to previousstandards when reaching its limits. It does not break into blocksbut degrades much more smoothly, making the image softer.H.264H.264 attains the MPEG-2 quality compression at a lower bit rate but withincreased computational effort, which is not a big deal in modern supercomputers age. It will provide the same quality as MPEG-2 at half the daterate as shown in Figure belowH.264 [2]Video quality MPEG-2 Vs H.264 [20]AVC EncoderH.264 Encoder [1]AVC DecoderH.264 Decoder [1]Encoder--Prediction: (Motion estimation and compensation)Intra Prediction [2] Inter Prediction[2] How does an H.264 codec work ?Intra Prediction : Intra-prediction uses the macroblocks from thesame image for prediction. Two types of prediction schemes are usedfor the luminance component. These two schemes can be referred asINTRA_4x4 and INTRA_16x16 [16]. In INTRA_4x4, a macroblock of size16x16 samples is divided into 16 4x4 subblocks. Intra prediction schemeis applied individually to these 4x4 subblocks. There are nine differentprediction modes supported as shown in Fig belowHow does an H.264 codec work ? [2]4x4 Luma prediction (intra-prediction) modes in H.264 [10]In mode 0, the samples of the macroblock are predicted from the neighboringsamples on the top. In mode 1, the samples of the macroblock are predictedfrom the neighboring samples from the left. In mode 2, the mean of all theneighboring samples is used for prediction. Mode 3 is in diagonally down-leftdirection. Mode 4 is in diagonal down-right direction. Mode 5 is in vertical-rightdirection. Mode 6 is in horizontal-down direction. Mode 7 is in vertical-leftdirection. Mode 8 is in horizontal up direction. The predicted samples arecalculated from a weighted average of the prediction samples A to M.How does an H.264 codec work ? [3]As shown in below figure four modes are used for prediction of 16x16 intraprediction of luminance components. The three modes, mode 0 (vertical),mode 1 (horizontal) and mode 2 (DC) are similar to the prediction modes of4x4 block. In the fourth mode, the linear plane function is fitted in theneighboring samples.How does an H.264 codec work ? [4]Inter Prediction : Inter-prediction is used to reduce the temporal correlationwith the use of motion estimation and compensation algorithms. An image isdivided into macroblocks; each 16x16 macroblock is further partitioned into16x16, 16x8, 8x16, 8x8 sized blocks. A 8x8 sub-macroblock can be furtherpartitioned in 8x4, 4x8, 4x4 sized blocks. Figure in the next slide illustrates thepartitioning of a macroblock and a sub-macroblock. The input videocharacteristics govern the block size. A smaller block size ensures less residualdata; however smaller block sizes also mean more motion vectors and hencemore number of bits required to encode theses motion vectors .How does an H.264 codec work ? [5]How does an H.264 codec work ? [6]Macroblock partitions for motion estimation/motion compensation–16x16, 16x8, 8x16 and 8x8 [9]Macroblock sub-partitions for motion estimation/ motion compensation – 8x8, 8x4, 4x8 and 4x4 [9]Deblocking Filter : H.264 suffer from blocking artifacts due to block-based transform in intra andinter-prediction coding, and quantization of transform coefficients. Deblocking filter reduces theartifacts at the block boundaries and prevents the propagation of accumulated noise. Thepresence of the filter however adds to the complexity of the system as shown in the figure below.Filtering is applied to horizontal or vertical edges of 4x4 blocks in a macroblock.How does an H.264 codec work ? [7]Boundaries in a macroblock to be filtered (luma boundaries shown with solid lines and chroma boundaries shown with dotted lines) [1]Subtracts the prediction from the current block---residualsResidual samples are transformed using 4x4 or 8x8 integer transform (DCT)The output of the transform – a block of transform coefficients, is quantized.Encoding—Variable length coding, Arithmetic codingHow does an H.264 codec work ? [8]Profile---Set of coding tools or algorithms that can be used in generating a bit stream (specific encoding tech).Level---Places constraints on certain key parameters of the bit stream.H.264 Profiles and levelsH.264 Profiles [1]Baseline Profile (BP): Primaryly for lower-cost applications with limitedcomputer resources. In BP macro blocks need not to be in raster
View Full Document
Unlocking...