Impact of Reference Distance for Motion Compensation Prediction on Video Quality

1
Impact of Reference Distance for Motion
Compensation Prediction onVideo Quality
Yubing Wang, Mark Claypool
and Robert Kinicki WPI Computer Science
Department Worcester, MA 01609 Wang_yubing_at_emc.co
m
ACM/SPIE Multimedia Computing and Networking
(MMCN) San Jose, California, January 31, 2007
2
Introduction

• Packet loss degrades video quality.
• Reference Picture Selection (RPS) selects one of
  several previous frames as a reference frame.
• Distance from selected frame is reference
  distance
• Higher reference distance, lower quality and vice
  versa
• Two modes NACK and ACK
• A systematic study of the effects of reference
  distance on video quality.
• A set of videos are selected
• Motion and scene complexity
• H.264 encoded (supports RPS)
• Two objective measures
• PSNR
• VQM

3
Outline

• Introduction
• Background
• Hypothesis
• Methodology
• Results and Analysis
• Conclusions

4
Reference Picture Selection (ACK)
1
3
4
5
6
7
2
ACK(1)
ACK(2)
ACK(3)

• The decoder acknowledges all correctly received
  frames.
• Only the acknowledged frames are used as a
  reference.
• Error propagation is avoided entirely.
• Reference distance increases with round-trip
  delay.
• Coding efficiency decreases as reference distance
  increases.
• Video quality degrades due to limited bit-rate.

5
Reference Picture Selection (NACK)
1
3
4
5
6
7
8
2
NACK(3)

• The previous frame is used as a reference during
  the error-free transmission.
• Reference distance is always 1 regardless of RTT
• The decoder sends a NACK for the erroneous frame
  along with a reference frame
• Error propagation
• Impact of loss increases with RTT

6
RPS ACK or RPS NACK?

• Which to use?
• Depends upon loss and loss pattern
• Depends upon round-trip delay
• Depends upon the effects of reference distance
• Can quantify effects of loss and delay
• Need to quantify the effects of reference
  distance on video Quality
• Will help determine choice between NACK and ACK

7
Video Quality Measurement

• Video Quality Metrics
• SSIM (Z. Wang et al), MPQM (C. Branden et al),
• PSNR
• VQM (ITS)
• Video Quality Metric (VQM)
• Developed by the Institute for Telecommunication
  Science (ITS).
• Provides an objective measurement for perceived
  video quality based on user studies.
• Measures the perceptual effects of video
  impairments and combines them into a single
  metric.
• VQM models include Television,
  Videoconferencing, General, Developer, PSNR.

8
Outline

• Introduction
• Background
• Hypothesis
• Methodology
• Results and Analysis
• Conclusions

9
Hypothesis

• Low Motion
• The similarities among frames are high
• More macro-blocks are inter-coded
• High motion
• The similarities among frames are low
• More macro-blocks are intra-coded

• The y-intersect is determined by motion and scene
  complexity.
• High-motion video sequences starts with low
  quality, degrade slower.
• Low-motion video sequence starts with high
  quality, degrade faster.

10
Outline

• Introduction
• Background
• Hypothesis
• Methodology
• Results and Analysis
• Conclusions

11
Methodology

• Select a set of non-compressed video clips with a
  variety of motion content.
• All in YUV 420, CIF (352x288)
• Each video sequence contains 300 video frames
  with a frame rate of 30 fps.
• Change reference distances for each selected
  video sequence.
• Encode the video clips using H.264.
• Only P-slices or I-slices are used.
• Same bit-rate constraint is imposed for all
  experiments.
• Only one single reference frame is used
• Measure video quality using PSNR and VQM.
• Analyze the results.

12
Video Clips Used in the Experiments
Video Clip Motion Fraction of P-Blocks Description
Akyio Low 0.9666 A news reporter talking
Container Low 0.9246 A container ship moving slowly
News Low 0.8746 Two news reporters talking
Silent Medium 0.8637 A person demonstrating sign language
Mom Daughter Medium 0.8423 A mother and daughter talking
Foreman High 0.5947 A foreman talking
Mobile High 0.5722 Panning of toy train moving
Coastguard High 0.5225 Panning of a coastguard ship moving
13
Outline

• Introduction
• Background
• Hypothesis
• Methodology
• Results and Analysis
• Conclusions

14
PSNR vs. Reference Distance (1)

• The video quality for videos with high motions
  tends to degrade slower than that for those
  videos with low motion.

15
PSNR vs. Reference Distance (2)
Video Clips a b R-Squared
Akiyo -2.0116 47.965 0.9953
Container -1.9023 44.838 0.9948
News -1.8556 43.295 0.9984
Silent -1.5283 41.41 0.9929
Mom Daughter -1.4581 41.442 0.9904
Foreman -1.1681 38.511 0.9265
Mobile -1.1553 26.663 0.9754
Coastguard -0.8626 35.582 0.9975

• The relationship between PSNR and reference
  distance can be characterized using a logarithmic
  function

16
VQM vs. Reference Distance (1)

• We use (1-VQM) as the quality metric.
• The same trend as the ones with PSNR.

17
VQM vs. Reference Distance (2)
Video Clips a b R-Squared
Akiyo -0.0113 0.9847 0.9869
Container -0.0114 0.9766 0.9848
News -0.0115 0.9732 0.9931
Silent -0.0124 0.9606 0.9937
Mom Daughter -0.0085 0.9217 0.9821
Foreman -0.0068 0.9059 0.9779
Mobile -0.0022 0.8055 0.9076
Coastguard -0.0014 0.8423 0.9671

• The relationship between VQM and reference
  distance can be characterized using a linear
  function

18
Conclusion

• Both PSNR and VQM video quality degrade as
  reference distance increases.
• The degree of the video quality degradation is
  affected by the video content.
• High-motion video sequences starts with lower
  quality, degrade slower.
• Low-motion video sequences starts with higher
  quality, degrade more rapidly.
• Mathematical Characterization of the
  relationship between video quality and reference
  distance
• PSNR
• VQM

19
Future Work

• Include a broader set of video clips.
• Develop an analytical model to compare RPS NACK
  and ACK mode
• Use reference distance results!
• Packet loss rate
• Round-trip delay
• Bit-rate constraints
• Insight to other repair techniques.