Contents Page
|
|  Yüklə 2.59 Mb.
|
4 The scalable and the non-scalable syntaxThe full syntax can be divided into two major categories: One is the non-scalable syntax, which is structured as a super set of the syntax defined in ISO/IEC 11172-2. The main feature of the non-scalable syntax is the extra compression tools for interlaced video signals. The second is the scalable syntax, the key property of which is to enable the reconstruction of useful video from pieces of a total bitstream. This is achieved by structuring the total bitstream in two or more layers, starting from a standalone base layer and adding a number of enhancement layers. The base layer can use the non-scalable syntax, or in some situations conform to the ISO/IEC 11172-2 syntax. 4.1 Overview of the non-scalable syntaxThe coded representation defined in the non-scalable syntax achieves a high compression ratio while preserving good image quality. The algorithm is not lossless as the exact sample values are not preserved during coding. Obtaining good image quality at the bitrates of interest demands very high compression, which is not achievable with intra picture coding alone. The need for random access, however, is best satisfied with pure intra picture coding. The choice of the techniques is based on the need to balance a high image quality and compression ratio with the requirement to make random access to the coded bitstream. A number of techniques are used to achieve high compression. The algorithm first uses block-based motion compensation to reduce the temporal redundancy. Motion compensation is used both for causal prediction of the current picture from a previous picture, and for non-causal, interpolative prediction from past and future pictures. Motion vectors are defined for each 16-sample by 16-line region of the picture. The prediction error, is further compressed using the discrete cosine transform (DCT) to remove spatial correlation before it is quantised in an irreversible process that discards the less important information. Finally, the motion vectors are combined with the quantised DCT information, and encoded using variable length codes. 4.1.1 Temporal processingBecause of the conflicting requirements of random access and highly efficient compression, three main picture types are defined. Intra coded pictures (I-Pictures) are coded without reference to other pictures. They provide access points to the coded sequence where decoding can begin, but are coded with only moderate compression. Predictive coded pictures (P-Pictures) are coded more efficiently using motion compensated prediction from a past intra or predictive coded picture and are generally used as a reference for further prediction. Bidirectionally-predictive coded pictures (B-Pictures) provide the highest degree of compression but require both past and future reference pictures for motion compensation. Bidirectionally-predictive coded pictures are never used as references for prediction (except in the case that the resulting picture is used as a reference in a spatially scalable enhancement layer). The organisation of the three picture types in a sequence is very flexible. The choice is left to the encoder and will depend on the requirements of the application. Figure I-1 illustrates an example of the relationship among the three different picture types. 
Figure 1 Example of temporal picture structure 4.1.2 Coding interlaced videoEach frame of interlaced video consists of two fields which are separated by one field-period. The specification allows either the frame to be encoded as picture or the two fields to be encoded as two pictures. Frame encoding or field encoding can be adaptively selected on a frame-by-frame basis. Frame encoding is typically preferred when the video scene contains significant detail with limited motion. Field encoding, in which the second field can be predicted from the first, works better when there is fast movement. 4.1.3 Motion representation - macroblocksAs in ISO/IEC 11172-2, the choice of 16 by 16 macroblocks for the motion-compensation unit is a result of the trade-off between the coding gain provided by using motion information and the overhead needed to represent it. Each macroblock can be temporally predicted in one of a number of different ways. For example, in frame encoding, the prediction from the previous reference frame can itself be either frame-based or field-based. Depending on the type of the macroblock, motion vector information and other side information is encoded with the compressed prediction error in each macroblock. The motion vectors are encoded differentially with respect to the last encoded motion vectors using variable length codes. The maximum length of the motion vectors that may be represented can be programmed, on a picture-by-picture basis, so that the most demanding applications can be met without compromising the performance of the system in more normal situations. It is the responsibility of the encoder to calculate appropriate motion vectors. The specification does not specify how this should be done. 4.1.4 Spatial redundancy reductionBoth source pictures and prediction errors have high spatial redundancy. This specification uses a block-based DCT method with visually weighted quantisation and run-length coding. After motion compensated prediction or interpolation, the resulting prediction error is split into 8 by 8 blocks. These are transformed into the DCT domain where they are weighted before being quantised. After quantisation many of the DCT coefficients are zero in value and so two-dimensional run-length and variable length coding is used to encode the remaining DCT coefficients efficiently. 4.1.5 Chrominance formatsIn addition to the 4:2:0 format supported in ISO/IEC 11172-2 this specification supports 4:2:2 and 4:4:4 chrominance formats. 4.2 Scalable extensionsThe scalability tools in this specification are designed to support applications beyond that supported by single layer video. Among the noteworthy applications areas addressed are video telecommunications, video on asynchronous transfer mode networks (ATM), interworking of video standards, video service hierarchies with multiple spatial, temporal and quality resolutions, HDTV with embedded TV, systems allowing migration to higher temporal resolution HDTV etc. Although a simple solution to scalable video is the simulcast technique which is based on transmission/storage of multiple independently coded reproductions of video, a more efficient alternative is scalable video coding, in which the bandwidth allocated to a given reproduction of video can be partially re-utilised in coding of the next reproduction of video. In scalable video coding, it is assumed that given a coded bitstream, decoders of various complexities can decode and display appropriate reproductions of coded video. A scalable video encoder is likely to have increased complexity when compared to a single layer encoder. However, this standard provides several different forms of scalabilities that address non-overlapping applications with corresponding complexities. The basic scalability tools offered are: data partitioning, SNR scalability, spatial scalability and temporal scalability. Moreover, combinations of these basic scalability tools are also supported and are referred to as hybrid scalability. In the case of basic scalability, two layers of video referred to as the lower layer and the enhancement layer are allowed, whereas in hybrid scalability up to three layers are supported. The following Tables provide a few example applications of various scalabilities. Table 1 Applications of SNR scalability
Table 2 Applications of spatial scalability
Table 3. Applications of temporal scalability
4.2.1 Spatial scalable extensionSpatial scalability is a tool intended for use in video applications involving telecommunications, interworking of video standards, video database browsing, interworking of HDTV and TV etc., i.e., video systems with the primary common feature that a minimum of two layers of spatial resolution are necessary. Spatial scalability involves generating two spatial resolution video layers from a single video source such that the lower layer is coded by itself to provide the basic spatial resolution and the enhancement layer employs the spatially interpolated lower layer and carries the full spatial resolution of the input video source. The lower and the enhancement layers may either both use the coding tools in this specification, or the ISO/IEC 11172-2 standard for the lower layer and this specification for the enhancement layer. The latter case achieves a further advantage by facilitating interworking between video coding standards. Moreover, spatial scalability offers flexibility in choice of video formats to be employed in each layer. An additional advantage of spatial scalability is its ability to provide resilience to transmission errors as the more important data of the lower layer can be sent over channel with better error performance, while the less critical enhancement layer data can be sent over a channel with poor error performance. 4.2.2 SNR scalable extensionSNR scalability is a tool intended for use in video applications involving telecommunications, video services with multiple qualities, standard TV and HDTV, i.e., video systems with the primary common feature that a minimum of two layers of video quality are necessary. SNR scalability involves generating two video layers of same spatial resolution but different video qualities from a single video source such that the lower layer is coded by itself to provide the basic video quality and the enhancement layer is coded to enhance the lower layer. The enhancement layer when added back to the lower layer regenerates a higher quality reproduction of the input video. The lower and the enhancement layers may either use this specification or ISO/IEC 11172-2 standard for the lower layer and this specification for the enhancement layer. An additional advantage of SNR scalability is its ability to provide high degree of resilience to transmission errors as the more important data of the lower layer can be sent over channel with better error performance, while the less critical enhancement layer data can be sent over a channel with poor error performance. 4.2.3 Temporal scalable extensionTemporal scalability is a tool intended for use in a range of diverse video applications from telecommunications to HDTV for which migration to higher temporal resolution systems from that of lower temporal resolution systems may be necessary. In many cases, the lower temporal resolution video systems may be either the existing systems or the less expensive early generation systems, with the motivation of introducing more sophisticated systems gradually. Temporal scalability involves partitioning of video frames into layers, whereas the lower layer is coded by itself to provide the basic temporal rate and the enhancement layer is coded with temporal prediction with respect to the lower layer, these layers when decoded and temporal multiplexed to yield full temporal resolution of the video source. The lower temporal resolution systems may only decode the lower layer to provide basic temporal resolution, whereas more sophisticated systems of the future may decode both layers and provide high temporal resolution video while maintaining interworking with earlier generation systems. An additional advantage of temporal scalability is its ability to provide resilience to transmission errors as the more important data of the lower layer can be sent over channel with better error performance, while the less critical enhancement layer can be sent over a channel with poor error performance. 4.2.4 Data partitioning extensionData partitioning is a tool intended for use when two channels are available for transmission and/or storage of a video bitstream, as may be the case in ATM networks, terrestrial broadcast, magnetic media, etc. The bitstream is partitioned between these channels such that more critical parts of the bitstream (such as headers, motion vectors, low frequency DCT coefficients) are transmitted in the channel with the better error performance, and less critical data (such as higher frequency DCT coefficients) is transmitted in the channel with poor error performance. Thus, degradation to channel errors are minimised since the critical parts of a bitstream are better protected. Data from neither channel may be decoded on a decoder that is not intended for decoding data partitioned bitstreams. INTERNATIONAL STANDARD 13818-2 RECOMMENDATION ITU T H.262 INFORMATION TECHNOLOGY - GENERIC CODING OF MOVING PICTURES AND ASSOCIATED AUDIO INFORMATION: VIDEO |