Multimedia Communications Spring 2006-07 Video Quality & - - PowerPoint PPT Presentation

▶
multimedia communications
SMART_READER_LITE
LIVE PREVIEW

Multimedia Communications Spring 2006-07 Video Quality & - - PowerPoint PPT Presentation

Dec 21, 2023 160 likes •969 views

CS 584 / CMPE 584 Multimedia Communications Spring 2006-07 Video Quality & Traffic Characteristics Shahab Baqai LUMS Video Compression & Encoding Control Schemes Quality degradation resulting from Digitization

slide-1
SLIDE 1

CS 584 / CMPE 584

Multimedia Communications

Spring 2006-07

Video Quality & Traffic Characteristics

Shahab Baqai LUMS

slide-2
SLIDE 2

2

Video Compression & Encoding Control Schemes Quality degradation resulting from

– Digitization – Quantization

For fixed given values for the encoder parameters both quality and bit rate vary with scene content To achieve certain quality, bit rate, and/or delay

  • bjectives over time, encoder parameters must be

dynamically adjusted Video Encoding Control Schemes

– Constant bit rate – Open-loop variable bit rate – Constant quality variable bit rate

slide-3
SLIDE 3

3

ITS Video Quality Measure – ˆ

s

Quantitative video quality measure developed by ITS Institute for Telecommunication Sciences Measures spatial & temporal information loss 94% correlation with subjective evaluations Validated using 10 sec sequences, wide range of impairments Quality degradation measured on a scale of 1 to 5

slide-4
SLIDE 4

4

Video Quality Measure – ˆ

s

1 2 3 1 2 3

ˆ 4.77 0.272 0.356 where :spatial distortion measure : temporal distortion measure, detects lost motion : temporal distortion measure, detects added (false) motion s m m m m m m = − −

slide-5
SLIDE 5

5

Video Quality Measure – SNR

( ) ( ) ( ) ( ) ( )

2 1 10 2 1 th th th th

10log where : Luminance value for the i pixel of the n orignal frame : Luminance value for the i pixel of the n encoded/decoded frame : Number of pixels in

p p

N i i N i i i i i p

  • n

SNR

  • n

d n

  • n
  • n

N

= =

= ⎡ ⎤ − ⎣ ⎦

∑ ∑

a frame

slide-6
SLIDE 6

6

Tools & Experimental Setup

slide-7
SLIDE 7

7

Analysis of Video Traffic & Quality Characteristics

slide-8
SLIDE 8

8

Video Sequences Used

slide-9
SLIDE 9

9

Description of Video Sequences Used Video Conferencing

– Head shoulders, no scene change, little motion

Terminator 2

– Morphing, slow movement, sharp edges, several scene changes

Raiders

– Some very fast moving parts (panning), many scene changes

Star Trek

– Mixture of fast & slow movement, some irregular movement due to camera shaking, 29 scene changes

Commercials

– Mixture of real & synthetic images, second half contains very fast movement and sharp edges

slide-10
SLIDE 10

10

Constant bit rate Video Encoder Control

slide-11
SLIDE 11

11

Quantization

slide-12
SLIDE 12

12

Quantization Matrix for Intra-Coded Macro Block

slide-13
SLIDE 13

13

CBR Quality Characterization

Star Trek, V = 384 kbps, B = 19.2 kbits

slide-14
SLIDE 14

14

CBR Quality Characterization

Star Trek, V = 384 kbps, B = 384 kbits

slide-15
SLIDE 15

15

CBR Quality Characterization

Star Trek, V = 384 kbps, B = 1920 kbits

slide-16
SLIDE 16

16

CBR Quality Characterization

Star Trek, V = 1536 kbps, B = 19.2 kbits

slide-17
SLIDE 17

17

CBR Quality Characterization

Star Trek, V = 1536 kbps, B = 384 kbits

slide-18
SLIDE 18

18

CBR Quality Characterization

Star Trek, V = 1536 kbps, B = 1920 kbits

slide-19
SLIDE 19

19

CBR Quality Characterization

Video Conference, V = 384 kbps, B = 19.2 kbits

slide-20
SLIDE 20

20

CBR Quality Characterization

Video Conference, V = 384 kbps, B = 384 kbits

slide-21
SLIDE 21

21

CBR Quality Characterization

Video Conference, V = 384 kbps, B = 1920 kbits

slide-22
SLIDE 22

22

CBR Quality Characterization

Video Conference, V = 1536 kbps, B = 19.2 kbits

slide-23
SLIDE 23

23

CBR Quality Characterization

Video Conference, V = 1536 kbps, B = 384 kbits

slide-24
SLIDE 24

24

CBR Quality Characterization

Video Conference, V = 1536 kbps, B = 1920 kbits

slide-25
SLIDE 25

25

CBR Quality Characterization

Commercials, V = 384 kbps, B = 19.2 kbits

slide-26
SLIDE 26

26

CBR Quality Characterization

Commercials, V = 384 kbps, B = 384 kbits

slide-27
SLIDE 27

27

CBR Quality Characterization

Commercials, V = 384 kbps, B = 1920 kbits

slide-28
SLIDE 28

28

CBR Quality Characterization

Commercials, V = 1536 kbps, B = 19.2 kbits

slide-29
SLIDE 29

29

CBR Quality Characterization

Commercials, V = 1536 kbps, B = 384 kbits

slide-30
SLIDE 30

30

CBR Quality Characterization

Commercials, V = 1536 kbps, B = 1920 kbits

slide-31
SLIDE 31

31

Quality vs. B for Star Trek

slide-32
SLIDE 32

32

Quality vs. V for Star Trek

slide-33
SLIDE 33

33

Quality vs. V for All Sequences

slide-34
SLIDE 34

34

Equal Contours, Star Trek

min

ˆ s

slide-35
SLIDE 35

35

Equal Contours, All Sequences

min

ˆ s

slide-36
SLIDE 36

36

CBR Delay Characterization for Transmission over a circuit

Delay for a Video transmission over a circuit

slide-37
SLIDE 37

37

Delay Statistics vs. B

C = V V = 384 kbps V = 1536 kbps

slide-38
SLIDE 38

38

Delay Statistics vs. V

C = V

slide-39
SLIDE 39

39

Maximum Delay Contours

C = V

slide-40
SLIDE 40

40

Meeting Delay & Quality Objectives (1)

slide-41
SLIDE 41

41

Meeting Delay & Quality Objectives (2)

slide-42
SLIDE 42

42

Transmission at C > V (1)

slide-43
SLIDE 43

43

Transmission at C > V (2)

slide-44
SLIDE 44

44

CBR: Meeting Quality & Delay Constraints by Choosing V < C Given

– C = 1024 kbps – Dmax = 30 ms, smin = 4.2

If V is chosen to be 1024 kbps, then delay & quality

  • bjectives cannot be met

eg for

– B = 30 kbits, Dmax = 30 ms smin = 3.5

Choosing V = 768 kbps & B = 384 kbits we achieve Dmax = 25 ms smin = 4.2

In Particular when Dmax < 40 – 50 ms, it is useful to choose V to be somewhat smaller than C

D

slide-45
SLIDE 45

45

CBR Traffic Characterization

Number of Bits per Frame for Star Trek

slide-46
SLIDE 46

46

CBR Traffic Characterization

Number of Bits per Frame for Video Conferencing

slide-47
SLIDE 47

47

CBR Traffic Characterization

Frame Size Histograms for Star Trek

slide-48
SLIDE 48

48

CBR Traffic Characterization

Frame Size Histograms for Video Conferencing

slide-49
SLIDE 49

49

CBR Traffic Characterization

Frame Size Histograms for Commercials

slide-50
SLIDE 50

50

CBR Traffic Characterization

Bits per Frame Statistics V = 1536 Kbps, B = 384 kbits (Average number of bits per frame = 51200)

slide-51
SLIDE 51

51

Multiplexing CBR Streams over a Circuit

slide-52
SLIDE 52

52

Video Conference: Multiplexing Performance

slide-53
SLIDE 53

53

Commercials: Multiplexing Performance

slide-54
SLIDE 54

54

CBR: MPEG

Frame Size

slide-55
SLIDE 55

55

Open-loop Variable bit rate Video Encoder Control

slide-56
SLIDE 56

56

Quality vs. Time for Star Trek

slide-57
SLIDE 57

57

Minimum Quality vs.

q

slide-58
SLIDE 58

58

Number of Bits per Frame

slide-59
SLIDE 59

59

Average Frame Size vs.

q

slide-60
SLIDE 60

60

Maximum Frame Size vs.

q

slide-61
SLIDE 61

61

Frame Size Histograms

slide-62
SLIDE 62

62

Frame Size Autocorrelation Functions

slide-63
SLIDE 63

63

OL-VBR MPEG Encoding: Frame Size vs. Time

slide-64
SLIDE 64

64

Frame Size Histogram: MPEG vs. H.261

slide-65
SLIDE 65

65

Frame Size Histograms for I, P, & B Frames

slide-66
SLIDE 66

66

Constant Quality Variable bit rate Video Encoder Control

slide-67
SLIDE 67

67

Video Conferencing: Constant Quality Achieved

slide-68
SLIDE 68

68

CQ-VBR: Quality Statistics

slide-69
SLIDE 69

69

CQ-VBR: Number of Bits per Frame

slide-70
SLIDE 70

70

Frame Size Histograms

slide-71
SLIDE 71

71

Frame Size Statistics

slide-72
SLIDE 72

72

Frame Size vs. Target Quality

slide-73
SLIDE 73

73

CQ-VBR: Delay when Video is Transmitted over a Circuit

slide-74
SLIDE 74

74

Maximum Delay vs. C

slide-75
SLIDE 75

75

Comparison: CQ-VBR Achieves Lower Average Bit Rate

slide-76
SLIDE 76

76

Comparison: CQ-VBR Achieves Lower Average Bit Rate

slide-77
SLIDE 77

77

Joint Peak rate & Quality Controlled VBR

slide-78
SLIDE 78

78

JPQC-VBR: Peaks Eliminated

slide-79
SLIDE 79

79

JPQC-VBR: Comparison of Quality Statistics with CQ-VBR

slide-80
SLIDE 80

80

JPQC-VBR: Comparison of Frame Size Statistics with CQ-VBR