QoE-driven Smart Adaptive Video Streaming
SSIMWave’s QoE-driven smart adaptive video streaming technology is the engine that drives SSIMWave’s Smart Video Streaming (SVS) product.
An increasingly popular approach in state-of-the-art over-the-top (OTT) video-on-demand (VoD) applications is the adoption of adaptive video streaming techniques, where each source video content is encoded/transcoded into multiple streams of different bitrates and resolutions. The video streams are divided into time segments in the order of seconds. When a client requests the video content online, it can instantly pick one of the many streams for each time segment in an adaptive manner, typically based on network bandwidth, buffer size, playback speed, and perhaps other parameters. Such an adaptive streaming framework puts the burden at the server side due to increased CPU power (for repeated encoding/transcoding demand) and increased storage space (for the storage of multiple streams of the same content). However, it allows to serve users of large variations in terms of their connections to the network without changing the existing video delivery infrastructure, with the potential to provide the best possible service to each individual user on a moment-by-moment basis.
A major problem with the current adaptive streaming techniques is not properly taking the viewer’s quality-of-experience (QoE) into account. Since the ultimate goal of video delivery service is to provide the clients with the best possible video in terms of their visual QoE, properly assessing visual QoE and using such assessment as the key factor in the design and optimization of the video delivery systems is highly desirable. Unfortunately, this is exactly what is missing in the current adaptive video streaming implementations. Real-world systems typically use bit rate as the key factor, equating it to a visual quality indicator, but using the same bit rate to encode different video content could result in dramatically different visual quality. Even worse, the actual user QoE varies depending on the device being used to display the video, another factor that cannot be taken into account by bit rate-driven streaming strategies.
SSIMWave’s smart adaptive video streaming technique solves the problem in a scientific way by optimizing the adaptive streaming system for the best user experience. It is built upon SSIMWave’s automatic video QoE assessment technology, the most advanced visual QoE assessment technology so far. The key is to first let the client be aware of the QoE assessment, and then to use the QoE assessment as a critical factor to drive the decision making process in adaptive video streaming at the client side.
In SSIMWave’s smart streaming technology, a matrix of viewer QoE for each segment of each video stream is first created, and the decision on the selection of the next segment of the video is made by combining SSIMWave’s QoE estimation with other important factors, including the bitrates of video streams, the resolutions of the video streams, the available bandwidth of the network, the decoding speed, display speed, buffer size and power of the receiver device.
SSIMWave’s smart streaming technology constitutes of a series of advanced QoE-driven decision making strategies. The benefits of adopting the SSIMWave technology include reduced bandwidth usage, reduced probability of video playback freezing/rebuffering, improved customer visual QoE, improved smoothness of customer QoE, reduced customer data usage, and reduced customer power usage.