Mikko Uitto

Dynamic Cross-Layer Adaptation of Scalable Video in Wireless Networking

The recent trend in Internet traffic indicates the proliferation of usage of multimedia services where a substantial part is related to some sort of video transmission. Moreover, an increasing number of Internet users employ wireless access technologies. High-quality video streaming over wireless access in unison with great mobility brings challenges to sustain the mobile user perceived video quality high. Capacities of wireless links vary due to, for instance, coverage area limitations, multipath propagation, and fading. However, novel video codecs utilize a layered encoding/decoding mechanism, which conveniently allows adapting the video quality, and thus the bitrate, by adjusting the number of layers transmitted. In this study, we exploit an extensive cross-layer signaling framework for a dynamic scalable video adaptation in varying network capacity. We focus on comparing a fast and fair MAC-layer packet scheduling with a relatively slow and long-term adaptivity taken place already at the application layer using a real H.264/SVC video. Our results attest the advantages of adaptation through the use of feedback signaling, which enables continuing the use of the current network access despite its capacity variation.

Spatial enhancement layer utilisation for SVC in base layer error concealment

The Scalable Video Coding (SVC) has been recently added as an extension to H.264/AVC standard. This extension allows both bit rate and device capability adaptation which are desirable features especially in error-prone wireless heterogeneous networks. This paper investigates the spatio-temporal error concealment techniques for packet losses in wireless IP networks, where scalable video coding (SVC) can be used. Three methods are introduced: pixel-value interpolation, frame copy and a new method, which utilises the correctly received spatial enhancement layer information if the corresponding base layer is missing. Unlike the new method, the traditional methods discard the corresponding enhancement layer data in the case when the base layer is missing. The simulation results indicate that enhancement layer utilisation provides better results in the case of a missing base layer than the traditional error concealment methods improving the image quality on average 2 dB.

Towards ubiquitous video services through scalable video coding and cross-layer optimization

Video content as one of the key features of future Internet services should be made ubiquitously available to users. Moreover, this should be done in a timely fashion and with adequate support for Quality of Service (QoS). Although providing the required coverage for ubiquitous video services, wireless networks, however, pose many challenges especially for QoS-sensitive video streaming due to their inadequate or varying capacity. In this article, we propose a cross-layer video adaptation solution, which may be used for optimizing network resource consumption and user experienced quality of video streaming in wireless networks; thus improving the availability of video services to mobile users. Our solution utilizes the flexibility of the Scalable Video Coding (SVC) technology and combines fast and fair Medium Access Control (MAC) layer packet scheduling with long-term application layer adaptation. The proposed solution both improves the usage of network resources by dropping video data based on its priority when the network is congested but also reduces efficiently the number of useless packet transfers in a congested network. We evaluate our solution with a simulation study under varying network congestion conditions. We find that already application layer adaptation gains over 60% less base layer losses, momentous for SVC video decodability and quality, than in the case without any adaptation. When our MAC layer scheduling is enabled, nearly a zero loss situation with respect to packet losses carrying base layers can be attained, resulting in peak-signal-to-noise ratio values very close to the original.

Quality assessment and error concealment for SVC transmission over unreliable channels

Scalable video coding (SVC) has aroused a wide interest in the areas of video coding and transmission technology, since it provides desirable features for heterogeneous error-prone network environments. The layered video structure allows not only adaptation to the available bandwidth but also a device adaptation capability via multiple decodable sub-streams. In this paper, we focus on SVC Medium- Grain Scalability (MGS) and investigate the end users quality of experience (QoE) in error-prone transmission conditions. We describe our error detection and concealment implementation on the JSVM 9.15 reference decoder and assess its effectiveness through several quality indicators. Results show an improvement of the proposed concealment strategy when compared to the standard Frame Copy (FC) and Interpolation, in terms of Peak Signal-to-Signal Noise Ratio (PSNR) and Structural Similarity Index Metric (SSIM) quality metrics. This improvement is evident in the case of concealment on quality enhancement layers, since video fluidity is preserved assuring an acceptable QoE to the end users.

QoE-based management of medical video transmission in wireless networks

The prevalence of broadband wireless networks offering nearly ubiquitous Internet connectivity today has opened new possibilities for utilizing wireless and mobile services also in eHealth applications. However, such services often involve transmission of critical multimedia data, thus having stringent requirements for network Quality of Service (QoS) not supported in current networks. In this paper, we propose means for implementing enhanced QoS support for medical video transmission in wireless networks. Our solution extends existing wireless network QoS architectures with support for differentiating multiple flows and packets of the same service type based on their impact on the Quality of Experience (QoE). The paper also presents an experimental evaluation of the solution in a simulation environment. The simulation results indicate that our solution enhances the QoE significantly and therefore improves the availability of critical medical video services.

Hierarchical management architecture and testbed for mobile video service optimization

There is demand for solutions capable of managing the constantly increasing video traffic loads over wireless networks while ensuring sufficient QoS and QoE for mobile video services. This paper presents a hierarchical network management architecture for optimized mobile video streaming in heterogeneous multi-access networks. The architecture jointly utilizes the access diversity of multi-access networks and video adaptation capabilities in the optimization, and relies on novel cognitive algorithms in implementing the automated decision-making for the management. To facilitate the development of such algorithms as next steps, the paper presents a testbed including the required components. The paper also includes a first-pass experimental evaluation on the testbed and preliminary optimization algorithms.

SAND-assisted encoding control for energy-aware MPEG-DASH live streaming

Adaptive HTTP video streaming has gained popularity in providing live feed from the events to mobile consumers. For reaching the widest possible audience, video service providers usually need to generate multiple video representations from the input stream suitable for different clients, which can increase the server-side energy consumption and processor load. HTTP streaming enables client adaptation against network fluctuations, but the server is usually unaware of the client(s) network capacity and device characteristics. With this knowledge, the server could control and allocate the encoding resources better according to the client(s) needs and reduce the server energy consumption and load. In this paper, we provide energy-aware dynamic encoding control for MPEG-DASH live streaming, which can benefit especially when representations are unpopular or network is under congestion. To achieve this, we utilize messages between client and encoding server based on the upcoming MPEG standard initiative Server and Network Assisted DASH (SAND). The results indicate that our dynamic encoding control signaling can reduce energy consumption in the server, create storage savings and therefore decrease also server costs.

Dynamic Subscription-Based QoS Service for Traffic Optimization

Real-time Internet services such as video transmission and online gaming often require privileged treatment in the Internet routers and wireless links in order to meet their real-time QoS requirements. The challenge is to know which are the services requiring higher priority and when this special treatment is needed. In this study, we introduce mechanisms to intelligently classify different traffic flows and to utilize this information in a dynamic adjustment of traffic priorities. In the presented solution, mobile stations can dynamically subscribe to special treatment for their network services. We capitalize on a deployment of the solution in the context of scalable video and employ fixed WiMAX to provide broadband quality of service enabled wireless connection. Beyond introducing and demonstrating the feasibility of such a deployment, our measurement results attest that dynamic traffic priority change can provide sufficient mechanism to sustain the quality of the real-time service despite the unreliable wireless link.

Energy consumption evaluation of H.264 and HEVC video encoders in high-resolution live streaming

Green computing and energy efficiency has become an important issue in todays rapidly growing multimedia transmission. Meanwhile users want high-quality content to be viewed anywhere, content producers must think alternatives for cutting expenses in the server. Usually multiple sub-streams from the original video feed are needed for clients with unique device characteristics, which can notably load the video encoding server leading also to increased energy consumption. Furthermore, social sharing applications using live video streaming can consume battery notably, which also drives developing energy-efficient video coding solutions. Therefore, it is essential to find ways for optimizing and selecting the proper tools for energy-efficient content production already in the server. In this paper, we evaluate the energy and power consumption for the existing top-rated open source video encoders in proportion to video quality and bit rate. The selected video coding formats include current dominant H.264 Advanced Video Coding (H.264/AVC), latest standard High Efficiency Video Coding (HEVC) and Googles royalty free VP9. The results show that the selected H.264 encoder (x264) has the lowest energy consumption, but the worst compression efficiency. On the contrast, x265 for HEVC has the best compression efficiency, but suffers from increased energy consumption. For our experiments, VP9 provides the best trade-off between compression efficiency and energy consumption.

Impact of Slice Size for SVC Coding Efficiency and Error Resilience

Scalable Video Coding (SVC) as an extension to the H.264/AVC standard enables adaptive video transmission, where several types of sub-streams can be decoded from a single encoded stream. The same video content can be streamed to low bit rate mobile phones with low quality as well as high bit rate televisions with extremely high quality. However, strict real-time requirements and unreliable transmission channel can cause packet losses, which means that sufficient error protection and concealment methods are needed. One of the error protection techniques in the encoder enables division of a picture into slices. Introduction of slices will decrease the coding efficiency but at the same time it will improve resilience against transmission errors. In this paper, we try to find a trade-off between the number of slices and coding efficiency for H.264/SVC to be used for video streaming in error-prone networks. In addition to coding efficiency, the selected slice size will affect on error resilience and error propagation inside the video stream. This means that the optimal slice size should not only provide sufficient coding efficiency but it should also provide a good error concealment ratio in relation to packet losses. This paper evaluates the trade-off between the coding efficiency and error resilience of H.264/SVC. The simulation results presented in the paper indicate that a minor increase of the amount of slices per picture greatly improves the error resilience but does not reduce the coding efficiency greatly. The best trade-off in our simulations is achieved by using three slices in a layer.