A Subjective Study for the Design of Multi-resolution ABR Video Streams with the VP9 Codec

Abstract

Adaptive bit rate (ABR) streaming is one enabling technology for video streaming over modern throughput-varying communication networks. A widely used ABR streaming method is to adapt the video bit rate to channel throughput by dynamically changing the video resolution. Since videos have different ratequality performances at different resolutions, such ABR strategy can achieve better rate-quality trade-off than single resolution ABR streaming. The key problem for resolution switched ABR is to work out the bit rate appropriate at each resolution. In this paper, we investigate optimal strategies to estimate this bit rate using both quantitative and subjective quality assessment. We use the design of bitrates for 2K and 4K resolutions as an example of the performance of this strategy. We introduce strategies for selecting an appropriate corpus for subjective assessment and find that at this high resolution there is good agreement between quantitative and subjective analysis. The optimal switching bit rate between 2K and 4K resolutions is 4 Mbps.

Multi-Model Standard for Bitstream-, Pixel-Based and Hybrid Video Quality Assessment of UHD/4K: ITU-T P.1204

“…For example, in services applying HTTP-based adaptive streaming (HAS), such as Video on Demand (VoD) or live streaming, the different video representations are typically realized by encoding each sequence at different resolutions and bitrates, reflecting a balance between target screen resolution and expected channel bandwidth (also referred to as "bitrate ladders"), see e.g. [1,2,3]. Depending on its usage, a video bitrate ladder reflects aspects such as the optimal resolution and encoder setting for a given target bitrate, or the bitrate that is needed for a given resolution to reach a certain quality level.…”

“…As was shown in a number of studies, even in laboratory tests with highquality screens and controlled viewing conditions that follow recommendations such as those in [27,28], in many cases video quality can hardly be distinguished between HD and 4K UHD resolution, specifically depending on the initial quality of the source content used [29,30,31,32,33]. On the other hand, test contents in video quality tests 1 often are rather artificial and not representative of actual target applications such as VoD or live streaming. The role of content and its quality and representative character is discussed for example in [29,34,12,32].…”

Barriers Towards No-reference Metrics Application to Compressed Video Quality Analysis: on the Example of No-reference Metric NIQE

“…The videos were chosen from MSU video collection which consists of 15,833 videos. The collection was divided into 28 clusters by spatiotemporal complex-ity [7] and one sequence from each cluster, which was close to the cluster center, was chosen for the final testing set. Each video was encoded by x264 and x265 encoders.…”

Hacking VMAF with Video Color and Contrast Distortion

“…We chose 4 videos which represent different spatial and temporal complexity [8], content and contrast to test transformations which may influence VMAF scores. All videos have FullHD resolution and high bit rate.…”