Major Criteria for Monitoring Quality of Service in Video Streaming
Four parameters to consider when evaluating the QoS in video streaming
Nowadays, video streaming is becoming increasingly popular. The current COVID19 situation and the wide-spreading format of a home office, online conferences, and web events accelerate live video streaming penetration in our everyday life. Video streaming is no longer a fancy innovation but a necessity. People are no longer satisfied with whatever quality the video output is, but demand high-quality streaming. They expect online streaming where video and audio run smoothly in high definition without interruptions and constant “buffering”. This is where the quality of user experience comes into play. To achieve the desired QoE (Quality of Experience), one must provide sufficient Quality of Service (QoS). In Part 1 of the article, we will discuss the main parameters of QoS. In Part 2, we will outline ways to monitor these parameters.
Quality of Service
Quality of Service is defined as the measurement of the customer experience of a given service. Concerning the video streaming service, the QoS consists of a series of measurements of each component in the streaming service chain. Starting from the receiving point of the streaming system, where the multimedia, either being live stream or VOD (Video On Demand), is played, going through the multimedia transcoding modules, then to the system’s network. In each point, the multimedia should be transformed in a way so it can reach its audience in good audio and video quality.
An important remark here is that the quality of a multimedia output can be measured by objective and by subjective factors. Often, we are more interested in the Perceived Video Quality, which analyzes how the Human Visual System assesses the quality of the playback. To correctly monitor the QoS for video streaming, we have to examine four main parameters. These are video quality, audio quality, adaptive bitrate streaming, and network delivery.
Video quality is one of the main parameters of every transcoding or streaming service. In most cases, there are two groups of measurements. The first group compares the video received at the end-user side with the one at the streaming system’s input. The second group comprises mechanisms for detecting different anomalies of the video at the receiving side.
The measurements of the video quality from the first group are most of the time expressed in MSE (Mean Squared Error), PSNR (Peak Signal Noise Ratio), or SSIM (Structural Similarity), where each has multiple variations. Still, all of them deal with the alteration of the video signal going through the streaming system. The streaming system’s transcoding modules are responsible for transforming the input video streams into a set of streams compatible with the delivery network and receiving devices’ capabilities. In the transcoders, the never-ending battle between lowering the bandwidth requirements and keeping good video quality occurs, a struggle with the highest impact on the overall QoS.
The second group of measurements deals with video anomalies at the receiving side, such as blank video, frozen frame, compression artifacts (e.g. blockiness, blur, etc.), where the measurement signals each anomalies’ presence.
Audio quality assessment is not as easy as the video quality measurement and can be only partially automated. The automatic measures involve the detection of audio volume peaks and loss of audio. At the same time, the actual quality assessment requires the involvement of real people voting into a predefined audio quality scale, unless machine learning algorithms are applied. Last but not least, audio and video synchronization are of high importance when it comes to audiovisual multimedia.
Adaptive bitrate (ABR) is a multimedia streaming technique for dynamically changing (adapting) the multimedia quality, depending on the network bandwidth and system resources. The presence of a sufficient number of encoded layers with adequate diversity in resolutions and bitrates is the main factor directly related to the QoS. The quality measurements related to the ABR are mainly focused on the synchronization between the different quality layers and validity of the streams with respect to the corresponding network delivery standards and consumer devices.
The QoS level can drastically decrease in case of bad network delivery, even with multimedia streams with perfect quality and excellent ABR layers distribution. The multimedia delivery highly depends on the CDN (Content Delivery Network) provider and its capability to handle all the consumer’s requests to access the multimedia. Another important QoS related parameter is the consistency of media delivery, which needs to be maintained at a reasonable level, no matter the system’s conditions. Most QoS network delivery related measurements are performed on CDN’s side, where the network load is generated.
Precise monitoring and correct data interpretation of the main QoS parameters are essential for achieving high video streaming performance. It is not just about distributing valuable content anymore but about delivering it in an exquisite format bringing superb QoE. Video quality, audio quality, adaptive bitrate streaming, and network delivery are the main parameters that we monitor to provide the desired QoS. In the second article, we will discuss the ways to monitor these parameters.
About the author
Tsviatko Jongov is a specialist in the area of digital multimedia technologies with more than 20 years of experience in the field. He started his career as a software developer, designing and building software solutions for the television broadcast industry, serving broadcasters around the globe. In 2008 he started his own software company providing digital media analysis and television automation solutions with customers such as HBO, Microsoft, Adobe, SpaceX, Dolby, and more. Now he is part of Bianor to help in the growth of the company.