THESIS PROPOSAL Aruna Ravi I D 1000 565191 Date May 15 2009 1 Student PERFORMANCE ANALYSIS OF DIRAC VIDEO CODEC PROPOSAL The objective of this thesis is to implement Dirac video codec encoder and decoder 1 based on several input test sequences and compare its performance with H 264 MPEG 4 Part 10 AVC 13 16 OVERVIEW OF THE DIRAC CODEC Dirac is a hybrid motion compensated state of the art video codec that uses modern techniques such as wavelet transforms and arithmetic coding 1 It is an open technology designed to avoid patent infringement and can be used without the payment of license fees It is well suited to the business model of public service broadcasters as it can be easily recreated for new platforms Dirac is aimed at applications ranging from HDTV high definition television to web streaming There are two parts in the Dirac development process A a compression specification for the bit stream and decoder and B software for compression and decompression The software is not intended simply to provide reference coding and decoding It is a prototype implementation that can freely be modified and deployed The decoder in particular is designed to be fast and more agile The resulting specification is simple and straightforward to implement and optimize for real time performance 1 Architecture In the Dirac codec image motion is tracked and the motion information is used to make a prediction of a later frame A transform is applied to the prediction error between the current frame and the previous frame aided by motion compensation and the transform coefficients are quantized and entropy coded 1 The encoder has the architecture as shown in Fig 1 whilst the decoder performs the inverse operations Fig 2 2 Fig 1 Encoder architecture the decoder performs the inverse operations 1 2 Notations Vin Input video frame P Previous frame Motion compensated Motion compensated prediction error MCPE e Vin P eTQ MCPE after wavelet transformation and quantization e MCPE after inverse transformation Vlocal P e Fig 2 Decoder architecture Notations P Motion compensated previous frame e Motion compensated prediction error MCPE Prediction error after inverse wavelet transformation and quantization Vout Output video frame 3 Temporal and spatial redundancies are removed by motion estimation motion compensation and discrete wavelet transform respectively Dirac uses a more flexible and efficient form of entropy coding called arithmetic coding which packs the bits efficiently into the bit stream 1 Wavelet Transform The 2D discrete wavelet transform provides Dirac with the flexibility to operate at a range of resolutions This is because wavelets operate on the entire picture at once rather than focusing on small areas at a time The wavelet transform is constructed by repeated filtering of signals into low and high frequency parts For two dimensional signals this filtering occurs both horizontally and vertically At each stage the low horizontal low vertical frequency sub band is split further resulting in logarithmic frequency decomposition into sub bands Wavelet transforms have been proven to provide a more efficient technique than block transforms with still images Within the Dirac wavelet filters the data is encoded in 3 stages as shown in Fig 3 Daubechies wavelet filters are used to transform and divide the data in subbands which then are quantized with the corresponding RDO rate distortion optimization parameters and then variable length encoded These three stages are then reversed at the decoder 5 Fig 3 Dirac wavelet transform architecture 5 The transform packs most of the information into only a few sub bands at low frequency as shown in Fig 4 which allows compression to be achieved This process can be repeated to achieve higher levels of wavelet transform Fig 5 A Dirac coded picture is free from block artifacts and is clearly superior in the case of moving images 1 Fig 4 Stages of wavelet transform 4 1 Fig 5 Wavelet transform frequency decomposition 5 Scaling and Quantization Scaling involves taking frame data after application of wavelet transform and scaling the coefficients to perform quantization Quantization employs a rate distortion optimization algorithm to strip information from the frame data that results in as little visual distortion as possible Dirac uses a dead zone quantization as shown in Fig 6 which differs from orthodox quantization by making the first set of quantization step twice as wide This allows Dirac to perform coarser quantization on smaller values 5 Fig 6 Dead zone quantizer with quality factor QF 5 Entropy Coding This is applied after wavelet transform to minimize the number of bits used It consists of three stages binarization context modeling and arithmetic coding as shown in Fig 7 The purpose of the first stage is to provide a bit stream with easily analyzable statistics that can be encoded using arithmetic coding which can adapt to those statistics reflecting any local statistical features The context modeling in Dirac is based on the principle that whether a coefficient is small or not is well predicted by its neighbors and its parents 3 Arithmetic coding performs lossless compression and is both flexible and efficient 5 Fig 7 Dirac entropy coding architecture 6 The non zero values in the higher frequency sub bands of the wavelet transform are often in the same part of the picture as they are in lower frequency sub bands Dirac creates statistical models of these correlations and arithmetic coding allows us to exploit these correlations to achieve better compression The motion information estimated at the encoder also uses statistical modeling and arithmetic coding to compress it into the fewest number of bits This compressed data is put into the bit stream to be used by the decoder as part of the compressed video Motion Estimation Exploits temporal redundancy in video streams by looking for similarities between adjacent frames Dirac implements hierarchical motion estimation Fig 8 in three distinct stages Fig 8 Hierarchical motion estimation 10 In the first stage pixel accurate motion vectors are determined for each block and each reference frame by hierarchical block matching In the second stage these pixel accurate vectors are refined by searching sub pixel values in the immediate neighborhood In the final stage mode decisions are made for each macro block determining the macro block splitting level and the 6 prediction mode used for each prediction unit This last stage
View Full Document
Unlocking...