Academia.eduAcademia.edu

Figure 2 – uploaded by Jeroen van der Hooft

See full PDFdownloadDownload figure
Figure 2. An example live video scenario for HTTP/1.1 (top) and HTTP/2 (bottom), where the client requests m available segments to ramp up the buffer. If the last released segment has index n, the first segment to play isn—m+l1. Note that r; denotes the release of segment i at server side, while s; denotes its request for download by the client. Furthermore, quality g; indicates that the server should change the quality of pushed segments to j.

Figure 2 An example live video scenario for HTTP/1.1 (top) and HTTP/2 (bottom), where the client requests m available segments to ramp up the buffer. If the last released segment has index n, the first segment to play isn—m+l1. Note that r; denotes the release of segment i at server side, while s; denotes its request for download by the client. Furthermore, quality g; indicates that the server should change the quality of pushed segments to j.

Related Figures (5)

Figure 1. The concept of HTTP Adaptive Streaming. Jeroen van der Hooft, Student Member, IEEE, Stefano Petrangeli, Student Member, IEEE, Tim Wauters, Member, IEEE, Rafael Huysegems, Patrice Rondao Alface, Senior Member, IEEE, Tom Bostoen, and Filip De Turck, Senior Member, IEEE
Figure 1. The concept of HTTP Adaptive Streaming. Jeroen van der Hooft, Student Member, IEEE, Stefano Petrangeli, Student Member, IEEE, Tim Wauters, Member, IEEE, Rafael Huysegems, Patrice Rondao Alface, Senior Member, IEEE, Tom Bostoen, and Filip De Turck, Senior Member, IEEE
Figure 3. A car travelling from north to south (left), along with the measured throughput (right). When travelling from (1) to (2), large townhouses on the right side impede the client’s connection. Arriving at (2), the client switches to a new antenna with better coverage. Once an open area is reached in (3) and a new antenna is again selected, throughput improves significantly.
Figure 3. A car travelling from north to south (left), along with the measured throughput (right). When travelling from (1) to (2), large townhouses on the right side impede the client’s connection. Arriving at (2), the client switches to a new antenna with better coverage. Once an open area is reached in (3) and a new antenna is again selected, throughput improves significantly.
Figure 4. Obtained video bit rates for the different quality representations and a GOP length of 8, 16, 30, 60 and 120 frames.
Figure 4. Obtained video bit rates for the different quality representations and a GOP length of 8, 16, 30, 60 and 120 frames.
Figure 5. Impact of the RTT on the video quality, both for HTTP/1.1 and HTTP/2 with an initial buffer size of 10 seconds.
Figure 5. Impact of the RTT on the video quality, both for HTTP/1.1 and HTTP/2 with an initial buffer size of 10 seconds.
Figure 6. Impact of the buffer size on the video quality and freeze time, both for HTTP/1.1 (GOP 120) and HTTP/2 (GOP 30) with an RTT of 300 ms.
Figure 6. Impact of the buffer size on the video quality and freeze time, both for HTTP/1.1 (GOP 120) and HTTP/2 (GOP 30) with an RTT of 300 ms.
Related topics:
Computer ScienceDistributed ComputingReal Time ComputingQuality of ExperienceVideo StreamingVideo QualityDynamic Adaptive Streaming over HTTP (DASH)Technology and EngineeringElectrical and Electronic EngineeringServer Push
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved