Godkända
Prestandaanalys av adaptiva streaming algoritmer för en låg-latens miljö
Antonio Kevo () och Albert Sjölund ()
Start
2022-01-11
Presentation
2022-06-09 11:15
Plats:
E:3139
Avslutat:
2022-06-20
Examensrapport:
Sammanfattning
Axis Communications is a world leading company in the market of surveillance and security solutions. Their products include network cameras, video encoders etc. Axis devices are used in many different kinds of security installations. An increasingly common use case is streaming of live video over the Internet, and across wireless networks. Live video puts higher demands on low latency, compared to streaming of surveillance video for storage (video stored for future inspection in case of an incident). The live use case gives higher demands of adapting the video bitrate to the currently available bandwidth, adjusting the amount of data for keeping low latency and avoiding frame drops. This thesis will focus on researching various means of adapting the bitrate when streaming live over networks with varying latency and bandwidth in Axis devices. This includes finding various techniques for bitrate adaptation similar to Transport-Wide Congestion Control (TWCC). We will also research methods of testing adaptive streaming under various simulated networks conditions, and ways to measure the effect of the bitrate adaptations. Finally, we will implement and evaluate a prototype which will demonstrate the improved live streaming when the network bandwidth and latency varies on real devices.
Handledare: Emma Fitzgerald (EIT)
Examinator: Maria Kihl (EIT)