Article Preview
TopIntroduction
Broadcasting, Video Content needs a better video coding strategy with a good compression format; H.264 / AVC is one such compression format which is suitable for all forms of transmission, storage and retrieval even with a loss of data. It is a block based form of compression format developed as a standard by international telecommunication union for telephony (ITU-T) video coding experts group (VCEG) with the ISO/IEC JTCI moving picture experts group (MPEG) (Schwarz, 2007). As an extension to advanced video coding (AVC), scalable video coding (SVC) offers a layered approach for scalability, which comprises of one base layer and one or more enhancement layer. Also to provide video services with reduced fidelity, low spatial resolution SVC does single encode and multiple decode. A video is subdivided into more number of subset bit streams. A subset bit stream is a partial bit stream made by discarding unnecessary packets of video. This partial bit stream is lesser in data size which is more useful in bandwidth constraints. SVC has a unique feature of extracting the video with the help of the partial bit stream (Schwarz, 2007), even from error prone network. The base layer is subset bit stream which decodes the low resolution video and enhancement layer is, one can decode the bit streams of base layer along with already encoded enhancement layers.
The bit stream is scalable, if part of it removed and is still decodable with the resultant (Schwarz, 2007). It has three forms of scalability – temporal, spatial and quality Scalability. In temporal, frame rate is considered, in spatial, resolution is considered and in quality, PSNR (Peak Signal-Noise-Ratio) is considered. In all forms of scalability, either a frame (temporal), size (spatial) or PSNR level in decibel (quality) will be discarded in the subset bit stream. The mode partition of MBs in the current frame is most similar to the reference frame. In the Spatial scalability video is coded at different formats such as quarter common intermediate format (QCIF), common intermediate format (CIF), standard definition television (SDTV), high definition television (HDTV), etc. With the samples of lower resolution video subset bit stream higher resolution video bit stream can be predicted in spatial scalability. For each macro block at the base layer the corresponding up sampled macro blocks at enhancement layers tend to have the same mode partition. In SNR/Quality/fidelity scalability the multiple levels of quality bit stream for a single spatial resolution bit stream video. High quality video can be obtained by decoding the samples of predicted lower quality subset bit stream.
SVC defines nine prediction modes for INTRA 4 x 4, four prediction modes for INTRA 16 x 16, SKIP mode and seven MB modes for INTER prediction such as INTER 16 x 16, INTER 16 x 8, INTER 8 x 16, INTER 8 x 8, INTER 8 x 4, INTER 4 x 8 and INTER 4 x 4 (Wiegand, 2003). Deciding the best mode for a current frame from previous frames introduces complexity in any video compression format; advanced video coding offers a single layered approach generates single bit stream, whereas scalable video coding generates multiple bit stream.
The computational complexity is increased in the enhancement layer due to the inclusion of INTRA BL (Intra Base Layer) mode, while this mode is included for improving coding efficiency. In addition to that, there is a Base Layer Skip (BLS) mode in the enhancement layer. It is triggered by the base mode flag. These will up sample all motion vectors and residual values without searching the base layer.
In single layer coding under each spatial layer, motion compensated prediction and intra prediction modes are used. Also to improve the coding efficiency, interlayer prediction mechanisms are implemented (Schwarz, 2007) for various spatial resolution. In SVC, three inter layer prediction mechanisms are added such as macro block mode prediction and its associated motion parameters, residual signal prediction and intra texture prediction.