cs526 logo
rainbow animatio

CS525 S2006 Midterm Exam

Enter your CS Unix machine login and the student ID as password and submit the midterm before 11:59pm 3/13/2006.   You can work at home or at PC lab where you can access it through a web browser.  For multiple-choice questions, you must choose either yes or no for each answer.

After filled in the answers in the text areas and selected the answers, please print a copy of the web page with your answers before you hit the submit button. Note that "save file" menuitem does not save the data you enter. It only saves the HTML source file. Submit your answers by pressing the submit button at the end of this web page.  You will get a confirm web page with all your answers. Save or print  the answer confirmation web page for your record. If you have problem getting the confirm web page or submitting midterm answers using the web page, just email me the answers.

Your name:
Your login on CS UNIX machines:
Your password ( SID):

    1. Why RTP is running on top of UDP and not on TCP?

    2. What is delay jitter and what is the cause of it?

    3. How can one remove delay jitter appeared in  RTP-based streaming applications?

    4. Given 1Mbps session bandwidth, 5 senders,  100 receivers, and an average of 100 byte RTCP packet size,  how often should the receiver sends the RR report messages (reference material at  Page 14 of the RTP handout)?

  2. Introduction to Video
    1. What is HDTV?
    2. 1. a video signal standard used in Japan Yes No
      2. 26MHz is allocated per channel Yes No
      3. video and audio signals are sent at different frequencies Yes No
      4. Each field contains 1025 scanning lines Yes No
    3. Relationship among bandwidth, scaning lines, frame rates, horizontal blanking (retract) period, horizontal resolution.
    4. 1. If other parameters are fixed,  higher frame rate implies more video bandwidth needed  Yes No
      2. If other parameters are fixed,  more scanning lines implies higher frame rate Yes No
      3. If other parameters are fixed,  shorter horizontal blanking period implies higher frame rates Yes No
      4. If other parameters are fixed,  longer horizontal blanking period implies higher resolution. Yes No
    5. HDTV vs. 35mm film
    6. 1. HDTV has higher resolutions Yes No
      2. HDTV has higher aperture response on wider (thick) HDTV line patterns. Yes No
      3. TV has higher aperture response on narrower (thin) HDTV line patterns. Yes No
    7. The day  light consists of lights of many colors.  Yes No
    8. Satuation of color in a light source indciates how bright is its color? Yes No
    9. YIQ video signal
    10. 1. Human eye are less sensitive to the luminence than color satuation Yes No
      2. Human eye are more sensitive to Q signal than I signal. Yes No
      3. When digitized TV signals, it make sense to give more bits to Q signal than I signal.  Yes No
      4. These three signal components are mixed during TV broadcast transmission. Yes No
  3. JPEG
    1. Zig-Zag Sequence and Run length Coding.
      Given the DCT coefficients with 64, followed by 48zeros, -100, and the rest are zeros
      1.What is the intermediate encoding sequence for the above DCT coefficnets?

      2.How many  bits will be used  to encode this intermediate encoding sequence?

      3.Consider that 8 bits are used to encode the value of each original 64 values. What is compression ratio achieved by the combination of DCT, runlength and VLI coding?

  4. MPEG
    1. MPEG encoded frames generate fix data bit rate.  Yes No
    2. Which of the following approaches can achieve  higher compression ratio?
      1. 1. Increase the number of quantization steps in the encoding of I frames Yes No
        2. Increase the search area in reference frame for matching the macroblock Yes No
        3. Use longer video sequence with fewer reference frames Yes No
    3. Assume we have mpeg file with IBBPBBPBBI encoding sequence. the macroblock at (200, 100) of reference frame 1 happens to match best with the macroblock at (256, 128) of frame 4. Further assume that the macroblocks are encoded with <macroblock type><motion vector>[<motion vector>]<prediction error matrix's encoded bit stream>. Here we have a forwarding prediction case. Use the fomula, unknowin the MPEG paper. Here subscript 2 is for the reference frame and 1 is for the frame with macroblock to be matched. What are the motion vector values to be encoded?

    4. Follow the above assumption and assume that the prediction error matrix's encoded bit stream is 70bits, each element in the motino vector is encoded in 10 bits. What is the compression ratio for the macroblock on (256,128) of frame 1?.

  5. Audio Processing
    1. How big is a raw 1 minutes, stereo, audio file sampled with 44kHz and 16 bits per sample using PCM coding?
    2. MP3 Audio Masking
    3. 1. Take advantages of the defection in human perception. Yes No
      2. Divide audio signal into subbands. Yes No
      3. If the surrounding subbands have signal below certain threshould, they are masked out and not represented. Yes No
      4. The masking thresholds before and after the reference signal are typically different. Yes No
  6. MIDI.
    The following is a MIDI file similar to Ballade2.mid.
    4d 54 68 64    4-character chunk ID: "MThd"
    00 00 00 06    length of the header: 6 bytes
    00 01          file format: format 1
    00 10          # of tracks: 16 tracks in this file
    00 c0      MSb=0, resolution per quarter note: 192 ticks per quarter note
    4d 54 72 6b    4-character chunk ID: "MTrk"
    00 00 00 7c    length of the track: 124 Bytes

    delta time track event   Intepretation
    ========= ===========    =====================================
    00 ff 7f 03 00 00 41     meta event, (?) 3 bytes
    00 ff 58 04 04 02 18 08  Time signature meta event, 4/4 meter?
    00 ff 51 03 0a 49 6d     Tempo meta event, 674157 microsec/quarternote
    00 ff 59 02 00 00        Key signature meta event, C major
    00 ff 01 2a 43 6f ...    Copyright meta event?

    4d 54 72 6b              4-character chunk ID: "MTrk"
    00 00 09 61              length of the chunk data: 2401 bytes

    delta track event   Interpretation
    ===== ===========   =====================================
    00    ff 21 01 01   meta event?
    00    ff 03 14 50 ... instrument name, 20 bytes, Piano Treble
    00    c1 02         Program change on channel 1, patch #2
    00    b1 07 61      Control change on channel 1, volume level 91
                             (loud, 80 normal)
    00    b1 0a 7f      continuous controller(MIDI knob), max. postion(127)
    00    91 37 7f      Note on channel 1,note=55(G,octave3), vel=127(loud)
    60    39 7f         note=57, vel=127(loud) after 96/192 quarternote
                          (A, octave3), running status without command
    60    3c 7f         note=60(C, octave4), vel=127 after eighth note
    5e    41 7f         note=65(F, octave4), vel=127 after 94/192 qnote
    8140  43 7f         note=67(G, octave4), vel=127 after 192/192 qnote
    10    81 37 40      Note off, note=55, vel=64(med.) after 16/192 qnote
    8130  91 41 7f      Note on, note=65, vel=127 after 176/192 quarternote
    06    81 3c 40      Note off, note=60, vel=64 after 6/192 quarternote
    0c    37 40         note=55, vel=64, after 12/192 quarternote
    4e    91 37 7f      Note on, note=55, vel=127, after 78/192 qnote

    1. How long is the playtime of the first note (G Octave3) in terms of seconds?

    2. If the above MIDI code after 4e 91 37 7f is followed by 00 91 30 60. Explain these four bytes using the format similar to the above interpretation.

If you feel some of the questions are ambiguous, state the problem # and your assumptions on the answers.