Comparison of representation switching number and achieved bit-rate in DASH algorithms
| Autor: | Drago Zagar, Jelena Vlaovic, Luka Filipovic, Snjezana Rimac-Drlje |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Test case
Hypertext Transfer Protocol Computer science computer.internet_protocol Encoding (memory) MPEG DASH adaptive streaming spatial and temporal information representation switching Dash Bandwidth (computing) Quality of experience Representation (mathematics) Algorithm Streaming algorithm computer |
| Zdroj: | 2020 International Conference on Smart Systems and Technologies (SST). |
| Popis: | MPEG Dynamic Adaptive Streaming over Hypertext Transfer Protocol (MPEG DASH) is adaptive streaming standard developed to unify previously developed adaptive streaming technologies. Over the years various MPEG DASH based adaptive streaming algorithms have been developed, mainly hybrid algorithms that take into account the buffer occupancy, available bandwidth and other parameters like video segment size. Adaptive streaming algorithms are used to determine the next video representation in order to ensure the best possible quality of experience in varying network conditions. They must balance well between the initial video delay and the number and depth of representation switching, as well as avoid video stalling events. This paper compares three different algorithms: Rate Adaptation Algorithm (LIU), Segment-Aware Rate Adaptation (SARA) and Content-Aware Rate Adaptation (CARA) in five network test cases with three video sequences. Obtained results show that on average LIU has the most representation switching events regardless of the spatial (SI) and temporal information (TI) of video sequence and it has the lowest achieved bitrate values. Although CARA and SARA achieve similar results, due to the fact that SARA has four phases of deciding on the representation of the next segment, it has the least number of switching events and a larger value of switching depth. CARA achieves highest average values of achieved bitrate in cases when video sequences Elephants Dream and Big Buck Bunny with lower SI and TI are used. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|