Exam 6: Mpeg Video Coding I -- Mpeg-1 and 2, Basic Audio Compression Techniques, Mpeg Audio Compression, Computer and Multimedia Networks

Speech coders using block-wise coding usually introduce delays. If the length of the speech analysis frame is 25 msec, give a reasonable estimate of the delay on the receiver side.

Work out the following problem of MPEG motion vector search in detail. The Target (Current) Frame is a B-frame. The size of macroblocks is $4 \times 4$ and the search range is the entire picture. The macroblock in question (darkened) in the frame has its upper-left corner at $\left(x_{t}, y_{t}\right)$ . It contains 12 dark pixels, each with intensity value 10. the other 4 pixels are part of the background, which has a uniform intensity value of 100 . There are obstacles in the scene having light pixels with value 50 . The complete scene is described in the figure. What is the best motion vector for coding this frame?

In MPEG's polyphase filter bank, if $24 \mathrm{kHz}$ is divided into 32 equal-width frequency subbands, (a) What is the size of each subband? (b) How many critical bands, at worst, does a subband overlap?

Multiplexing enables users to share the capacity over a physical link. How does FDM (Frequency Division Multiplexing) work?

Switching is important for interconnection of telephone networks. What are the two main switching technologies? How do they work?

Both UDP and TCP are used in the Transport layer in TCP/IP networks. Discuss the differences between UDP and TCP.

What is the compression ratio of MPEG audio if stereo audio sampled with 16 bits per sample at $48 \mathrm{kHz}$ is reduced to a bitstream of $256 \mathrm{kbps}$ ?