PolytechnicUniversity, Dept. Electrical and Computer Engineering

EE4414 Multimedia Communication System II

Fall 2003, Yao Wang

______

First Exam (10/21/2003 11-12:50) (Make Up)

Closed-book, 1 sheet of notes (single or double sided) allowed, no peeking into neighbors!

  1. (20 pt) Consider the following two raster scan formats: progressive scan using 20 frames/second, 200 lines/frame, and interlaced scan using 40 fields/second, 100 lines/field. For each scan format, determine
  2. The overall line rate (lines/second)
  3. The maximum temporal frequency (how many cycles of black and white frames alternating at most) the system can handle
  4. The maximum vertical frequency (how many cycles of black and white lines alternating within each frame or field) the system can handle
  5. The maximum frequency (cycles per second) in the 1D waveform of the raster signal, assuming the image aspect ratio is 1:1.
  6. Based on your results, explain the pros and cons of these two scan formats.

For parts (b)-(d), assuming a kell factor=1 for simplicity.

  1. (30 pt) In NTSC color TV transmission, the I and Q components are multiplexed with the Y component to form a single composite signal, using the QAM technique.
  2. (10 pt) Draw the block diagram for multiplexing Y, I, and Q together. Your block diagram should include the actual operation involved in QAM modulation (i.e., don’t use a block labeled as QAM modulator).
  3. (5 pt) Assume that the bandwidth of the Y, I, and Q signals are 4.2, 0.5, and 0.5 MHz respectively, and the audio has bandwidth 200 KHz, and is modulated to 4.5 MHz. The line rate is fl=30*525=15750 lines/second=15.75 KHz. We need to choose the color subcarrier frequency fc appropriately. Which of the following three choices is the best? Explain why briefly.
  4. fc =100.5 fl =1.58 MHz
  5. fc =225.5 fl =3.55 MHz
  6. fc =225 fl =3.54 MHz
  1. (15 pt) Draw the block diagram for a color TV receiver to separate Y, I, and Q components from the composite signal, assuming the composite signal is already demodulated to the baseband. To separate Y from I and Q, you can choose to use either a low-pass filter or a comb filter. Specify the range of the cut-off frequency for any low-pass filter that you may need. Your block diagram should include the actual operation involved in QAM demodulation (i.e., don’t use a block labeled as QAM demodulator).
  1. (15 pt) The figure below shows two interlaced video frames.
  2. Generate the field data associated with each frame.
  3. Deinterlace field 1 of frame 2 using field averaging. Write down the deinterlaced field.
  4. Now try line averaging. Write down the deinterlaced field.
  5. Now try line and field averaging.


  1. (15 pt) Draw a block diagram of a block-based hybrid video encoder. For simplicity, assume a block is always coded in the P-mode. Your block-diagram should include all the major operations involved, starting from when an input macroblock gets into the encoder, ending when the bits for this macroblock go out.
  1. (5 pt) What is the main benefit of using motion-compensated temporal prediction in video coding, compared to code a video frame directly? What are some of the problems due to motion-compensated temporal prediction?
  1. (15 pt ) For a video of 30 frames/second, 720x480 pixels/frame, what is the number of operations needed per second to accomplish integer-pel EBMA if we use block size of 16x16, search range of –30 to 30? (count one subtraction and taking absolute value, and sum of two numbers as one operation). (Note it is Ok to write down your total count in terms of the above parameters, without calculating the actual final number). In general, how does the operation count vary with the search range, search accuracy (integer vs. half-pel), the frame size, and the block size? What parameters affect the accuracy of the predicted image?