Analysis of FEC Function for Real-Time DV Streaming

1. Analysis of FEC Function for Real-TimeDV Streaming Kazuhisa Matsuzono, Hitoshi Asaeda, Kazunori Sugiura, Osamu Nakamura, and Jun Murai Keio University Graduate School of Media and Governance

2. HD Camera Introduction • Due to the widespread dissemination of high speed DSL and FTTH, the demands of transmitting high quality audio and video data have been increased. • DVTS (Digital Video Transmission System) is used in various research communities. • International symposium, network collaboration game, etc. studio studio HD Camera

3. buffer buffer Digital Video Streaming • Simply equipment • International symposium, network collaboration game, and so on.. • High quality and low delay and low cost • Resolution: 720×480,1280×720, 1440×1080 • average consumption bandwidth • About 30Mbps Receiver Sender Internet

4. The problem of quality reduction Receiver Sender Internet bursty • According to network condition, pktloss happens • Physical bandwidth or available bandwidth • Σ(DVTS traffic + other traffic ) > available bandwidth • Congestion • Quality reducing DV/RTP packet

5. buffer buffer Real-Time Streaming • Play quality vs Timeliness (interaction) • Packet reachable Dead Line is restricted • No utilizing the large amount of buffer • Congestion  reducing play quality • Adjusting sender’s transport method to changing network condition is very difficult • Rate control • FEC Receiver Sender

6. Motivation • Providing seamless and high interactive streaming on End-to-End model • taking into account high quality streaming, timeliness and high bandwidth-delay network • Supportive packet loss avoidance mechanism for real-time streaming application with minimal quality loss • use the network resource effectively • Providing best possible video quality • help reducingquality losses • Hopefully collaborate within other packet @ Internet sender Internet Video frame data Changing each rate FEC data

7. The relation between quality and FEC Video quality Quality is good? Max Situation A Optimal Point Situation B need Rate Control!! FEC rate

8. Related Work • Rate Control • TFRC (TCP Friendly Rate Control) • DCCP (Datagram Congestion Control Protocol) • QAFEC (quality adjusted FEC) MPEG source Change of MPEG quality Internet MPEG sender Receiver Change of FEC encoding rate Video Frame of pattern 1 Video Frame of pattern 2 FEC data

9. Design and Implementation of FEC function for DVTS • Static FEC using Reed-Solomon Code PAD RTP Header DV data FEC data Receive buffer FEC group ID = X FEC group ID = X+1 FEC group ID = X+2 DV part FEC part discard Waiting for DV data and FEC data Packet loss Waiting for FEC data Play buffer

10. Experiments • Test bed network Sender Receiver Dummynet DV streaming • Hardware in our experiment

11. Experimental Analysis • Costs of encoding and decoding FEC redundancy

12. DV Frame Rate 100%

13. DV Frame Rate 50%

14. Conclusion • Analyzing FEC function with DV streaming • It is very effective to provide the best possible streaming quality • Using the appropriate decided frame rate and FEC encoding rate • The further disruption of video and audio could occurs • deciding non-proper FEC encoding rate and frame rate toward the current bandwidth condition

15. Future Work • Verify and define the algorithm • Adaptive rate control and FEC for DVTS • Rate Control • Applied only when the network congestion is highly observed • FEC • Protect the quality to provide best possible streaming quality • Bandwidth estimation with packet error torelance • Probe the best combination between FEC rate and frame rate according to the current network condition • Packet loss rate, non-recovery rate, and the amount of receiver buffer • Evaluate our proposed mechanism on a heterogeneous communication enviroment

20. Thank you !!