(a) A video decoder receives a series of coded frames in one Group of Pictures (
ID: 2294038 • Letter: #
Question
(a) A video decoder receives a series of coded frames in one Group of Pictures (GOP) with the frame sequence "I B2 B P4 Bi Be P Bs B" The video is compressed using 4:2:0 chroma subsampling with a frame resolution of 800x600 pixels and 8 bits per pixel. The average bit rate for the compressed video is 4x10° bps. The video compressed at the required quality has an average I-frame compression ratio of 10:1, an average P-frame compression ratio of 30: and an average B-frame compression ratio of 72:1. Assume that the compression ratio includes all required headers. (i) (i) Determine the display order of the GOP. Compute the frame rate (frame per second) for the video.Explanation / Answer
B. MPEG 2 uses different quantization values in each of the quantization matrices because;
MPEG -2 uses variable quantization in which variable step size quantization is used for each coefficient. As the human eye is more sensitive to low frequency quantization noise so these low frequency coefficients are quantized by keeping the smaller step size and high frequency coefficients which are less important are quantized by keeping the larger step size during the quantization process. Usually the first value of the quantization matrix represents the average value OR mean value and based on the video frame information rest of the values of the quantization matrix represents the amount of variation from its average value. That’s why quantization values in each of the quantization matrices are different.
MPEG 2 uses different quantization matrices for inter block and intra block coding because:
In Intra block coding scheme its own data is used to reduce the spatial redundancy. Entire data in in the form of I frame (ie Intra Frames). This coding scheme do not require any reference frame and can be coded and then decoded separately.
Where as in Interblock encoding predictive and bidirectional predictive frames are used they needs the previous OR the next frame as the reference frames. Every small section / block of picture estimation tries to find the similar block in the reference frame. When the match is found that sub block is assumed as a interblock otherwise the block is considered as intra block.
As these two coding scheme uses different methods of operation hence the quantization matrices for these two are different
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.