Improving DASH Encoding with Scenes andDownscaling Techniques for VoD Streaming

Abstract

Video content represents a high percentage of thetraffic on Internet with an increasing number of platformsoffering live and on demand video. Most platforms use HTTPAdaptive Streaming (HAS) to deliver their content. In HAS, theprovider splits the videos in segments and each segment is offeredin multiple representations. The player can dynamically requestthe appropriate representation for the next segments dependingon the varying conditions. This alleviates problems such as initialdelay or stalling, thus providing a better quality of experienceto users. These segments can have a fixed duration or can beadapted to match scenes up to a maximum duration of typically10 seconds. In this work, we study the effects in time, quality andsize of the encoded videos when downscaling the video to obtainthe scenes. The experiments are performed using two codecs:H.264 and VP9, and using 10 videos with 4K resolution witha duration of more than 150 seconds. The videos are encodedusing fixed segments and variable segments based on scenes(obtained from the downscaled video). Results show that the useof downscaling to obtain the scenes has a small impact on thefinal quality, while reduces the total time, the consumption ofcomputational resources, and the size of the encoded video.