Exploring AV1 Encoder Potentials for Priority-Driven Wireless Multimedia Services

Abstract

World wide internet data usage is growing at a rapid rate and one of the main reasons for this high data usage is the increasing use of media streaming services. According to Cisco, video streaming will take up 82% of the world’s internet traffic by this year, 2022. Media streaming services use a large amount of data due to videos being inherently such large files which, for streaming services, need to be transferred over the internet. For this reason, we are researching methods to impact the use of a new generation open source video encoder, AV1, with wireless communications. Though the encoder is already more efficient at compressing video files than previous encoders, our goal was to research and understand AV1 in order to customize wireless protocols that optimize its usage in streaming scenarios. In our first attempt at understanding this encoder, we went through processes of tweaking settings built into the encoder in order to find optimal video quality when compared to file size. The file size is an important metric that correlates with how much data would have to be streamed. After this we went on to investigate how we could possibly estimate the per frame quality contribution of each frame transmitted. This goal was inspired by that of previous H.264/H.265 encoders’ I, P, and B frames, which each have their own distinguishable amount of quality contribution that can be split between 3 levels: high, medium, and low. We believe that the method provided could have a good indication of quality contribution similar to that of I, P and B frames with high, medium, or low quality contribution.