Martin Slanina

A Survey on Quality of Experience of HTTP Adaptive Streaming

Changing network conditions pose severe problems to video streaming in the Internet. HTTP adaptive streaming (HAS) is a technology employed by numerous video services that relieves these issues by adapting the video to the current network conditions. It enables service providers to improve resource utilization and Quality of Experience (QoE) by incorporating information from different layers in order to deliver and adapt a video in its best possible quality. Thereby, it allows taking into account end user device capabilities, available video quality levels, current network conditions, and current server load. For end users, the major benefits of HAS compared to classical HTTP video streaming are reduced interruptions of the video playback and higher bandwidth utilization, which both generally result in a higher QoE. Adaptation is possible by changing the frame rate, resolution, or quantization of the video, which can be done with various adaptation strategies and related client- and server-side actions. The technical development of HAS, existing open standardized solutions, but also proprietary solutions are reviewed in this paper as fundamental to derive the QoE influence factors that emerge as a result of adaptation. The main contribution is a comprehensive survey of QoE related works from human computer interaction and networking domains, which are structured according to the QoE impact of video adaptation. To be more precise, subjective studies that cover QoE aspects of adaptation dimensions and strategies are revisited. As a result, QoE influence factors of HAS and corresponding QoE models are identified, but also open issues and conflicting results are discussed. Furthermore, technical influence factors, which are often ignored in the context of HAS, affect perceptual QoE influence factors and are consequently analyzed. This survey gives the reader an overview of the current state of the art and recent developments. At the same time, it targets networking researchers who develop new solutions for HTTP video streaming or assess video streaming from a user centric point of view. Therefore, this paper is a major step toward truly improving HAS.

“To pool or not to pool”: A comparison of temporal pooling methods for HTTP adaptive video streaming

Current objective video quality metrics typically estimate video quality for short video sequences (10 to 15 sec) of constant quality. However, customers of video services usually watch longer sequences of videos which are more and more delivered via adaptive streaming methods such as HTTP adaptive streaming (HAS). A viewing session in such a setting contains several different video qualities over time. In order to express this in an overall score for the whole viewing session, several temporal pooling methods have been proposed in the related work. Within this paper, we set out to compare the performance of different temporal pooling methods for the prediction of Quality of Experience (QoE) for HTTP video streams with varying qualities. We perform this comparison based on ground truth rating data gathered in a crowdsourcing study in the context of the NGMN P-SERQU project. As input data for the models, we use objective video quality metrics such as PSNR, SSIM but also very basic inputs such as the bitrate of the clips only. Our results show that certain pooling methods perform clearly better than others. These results can help in identifying well performing temporal pooling methods in the context of HAS.

A Novel Metric for H.264/AVC No-Reference Quality Assessment

The article desribes a metric for H.264/AVC compressed video quality evaluation, which is being developed by our group. The metric operates on the encoded bitstream only. It uses the data directly present in the bitstream (or just entropy coded). No pixel decoding should be necessary, which will result in low computational complexity, and cosequently the possibility to analyze several bitstreams simultaneously. In this article, we describe one part of the system which has already been implemented -the feature extractor. A simple analysis is done to demonstrate the variability of the selected parameters used for evaluation.

Estimating PSNR without reference for real H.264/AVC sequence intra frames

The paper describes a novel metric for video quality measurement. The metric is capable of estimating peak signal-to-noise ratio (PSNR) values of pictures in compressed bit streams conforming to the H.264/AVC standard. It is designed and has been tested for intra predicted pictures with varying encoder settings. The metric uses an artificial neural network to estimate PSNR examining the compressed bit stream only. In other words, we do not need to decode the actual pixels within the pictures being evaluated.

Impact of test condition selection in adaptive crowdsourcing studies on subjective quality

Adaptive crowdsourcing is a new approach to crowdsourced Quality of Experience (QoE) studies, which aims to improve the certainty of resulting QoE models by adaptively distributing a fixed budget of user ratings to the test conditions. The main idea of the adaptation is to dynamically allocate the next rating to a condition, for which the submitted ratings so far show a low certainty. This paper investigates the effects of statistical adaptation on the distribution of ratings and the goodness of the resulting QoE models. Thereby, it gives methodological advice how to select test conditions for future crowdsourced QoE studies.

Coexistence Between DVB-T/T2 and LTE Standards in Common Frequency Bands

In the last decade, fulfilling the growing users’ expectations on the quality of provided multimedia services is one of the biggest challenges for different wireless communication systems. However, many times, different kinds of wireless systems can operate in the same (shared) frequency band. Such coexistence generally negatively affects the performance of the systems colliding. This is especially true for the widely used Digital Video Broadcasting-Terrestrial (DVB-T) or upcoming DVB-2nd Generation Terrestrial (DVB-T2) and Long-Term Evolution (LTE) standards, operating in Europe. Hence, the aim of this paper is to explore the impact of interfering mobile telecommunication services on the mobile TV services provided by LTE and DVB-T/T2 systems in the same frequency band. For the simulation of the coexistence between the LTE and DVB-T/T2 services on the physical layer level, an appropriate method is proposed and complex simulation tools are coupled. The results, presented as a dependence of modulation error ratio at the receiver input and bit error ratio at the channel decoder input on the spectral level ratio, show comparable performances for DVB-T and DVB-T2 systems. Results from simulations are verified by laboratory measurements and should help to better understand the influence of co-channel interferences from LTE mobile networks on DVB-T/T2 networks for mobile transmission scenarios.

A comparison of H.265/HEVC implementations

This paper deals with the most recent video coding standard H.265/HEVC. HEVC introduces new coding tools compared to older standards. Such a change has to influence the encoder complexity and will have influence on the encoding speed. In this paper, we focus on the performance of different implementations of the HEVC. The encoders are compared on the basis of final video quality and encoding speed. As the results showed, the differences in coding speed may be very significant among the encoders.

Subjective quality assessment in scalable video for measuring impact over device adaptation

Multiplicity of devices used for the reproduction of video is increasing nowadays. The characteristics of the displayable video are specific for each type of device. In order to avoid using separate source streams for the different devices and use optimal delivery bit rate, it is desirable that the encoded and transmitted original is identical for all the display devices and adapted in terms of video characteristics and bit rate with techniques such as Scalable Video Coding (SVC). In this paper, an analysis of perceived quality for different devices and different video characteristics is described, consisting in subjective tests with a group of observers and analysis of results. A collection of sequences were presented to the observers at different qualities, receiving a score on the Mean Opinion Score (MOS) scale (five grades from worst to the best quality). The devices considered included tablets and mobile phones, and complemented previous studies developed with high definition televisions. The results of the tests are presented and conclusions are drawn.

Reference free SSIM estimation for full HD video content

This paper proposes a reference-free video quality estimation method for full high definition video services based on a structural similarity index. The design of our estimator is based on an artificial neural network. To achieve this, the neural network was trained with a set of video statistical parameters extracted from the most representative video contents. Moreover, estimations with neural networks allow higher applicability and require lower processing power as known reference based methods. Finally, the achieved correlation between the calculated and the estimated structural similarity index shows a very good fit.

Fast method for reconstruction of 3D coordinates

The paper deals with the reconstruction of spatial coordinates of an arbitrary point from a scene using two monochromatic images which are in general relation. These two images can be created for example by a stereo camera in practical applications. Calculation is based on the perspective geometry. Search of the point features in each of these images is a fundamental building step to find image correspondence in this procedure. An algorithm for fast elimination of false correspondences which could devastate the scene reconstruction was designed. Well known mathematic relation is used to obtain the fundamental matrix and then reconstruct the 3D coordinates of feature points. Acquisition of 3D coordinates of points in the scene for which it is difficult to find correspondences is more complicated. In the paper, we have proposed a method based on utilization of the relation between the selected point and near features points. The performed experiments confirmed usability of designed procedure.