This work studies the jitter buffer management algorithm for Voice over IP in WebRTC. In particular, it details the core concepts of WebRTC’s jitter buffer management. Furthermore, it investigates how jitter buffer management algorithm behaves under network conditions with packet bursts. It also proposes an approach, different from the default WebRTC algorithm, to avoid distortions that occur under such network conditions. Under packet bursts, when the packet buffer becomes full, the WebRTC jitter buffer algorithm may discard all the packets in the buffer to make room for incoming packets. The proposed approach offers a novel strategy to minimize the number of packets discarded in the presence of packet bursts. Therefore, voice quality as perceived by the user is improved. ITU-T Rec. P.863, which also confirms the improvement, is employed to objectively evaluate the listening quality.

Objective speech quality assessment techniques, which use the perceptual models to emulate the human listening perception, have seen several revisions in the recent years. This study investigates the evolution of POLQA and ViSQOL models and scrutinise their latest versions. Prior work had identified weaknesses in both prediction models when presented with speech containing imperceptible playout adjustments. This study follows up the experiments to evaluate the progress and report the progress and the current issues, benchmarked against subjective listening quality scores. The assessment is conducted for all published versions of the POLQA and ViSQOL models and the evolution and improvement offered is analysed. We can conclude that the models have been improved in terms of imperceptible jitter buffer adjustments highlighted in prior work. This study also explores the performance of objective quality models and intelligibility (STOI and POLQA Intelligibility) models for a data set produced with realistic but extreme WebRTC scenarios using a standard and novel WebRTC jitter buffer strategy. An expert listening test was conducted to subjectively evaluate the WebRTC data set. It is observed that the standard WebRTC jitter buffer strategy produces more natural speech while the novel approach offers better intelligibility. The subjective and objective quality results suggest that the speech quality for standard jitter buffer were lower but more consistent than for the novel jitter buffer. The objective intelligibility results were conflicting. A followup study will conduct independent subjective evaluations of quality and intelligibility to further explore the relationship between the objective intelligibility and quality results.

This paper proposes an innovative, flexible, and easily reproducible test bed that facilitates large scale multimedia (MM) quality centric experiments. It makes use of Linux container technologies and several other utilities that collectively facilitate ease of deployment, minimal resource utilisation and better reproducibility. We outline the significant advantages of this approach over physical test beds, virtual machines and simulation and showcase it in an experiment dealing with speech quality aspects of WebRTC using the state-of-the-art quality prediction model, i.e. POLQA. However, it can be easily extended for any type of real time communication (RTC) and MM content. The particular test bed involves a single host on which multiple We-bRTC based VoIP endpoints run within Linux containers. The test bed also comprises a native network emulator running be-tween the endpoints. Several other tools are additionally incorporated in the test bed so that speech quality of the transmitted signal can be evaluated using objective quality prediction models, such as PESQ, POLQA, etc.

Real-time communication (RTC) applications like VoIP ideally require networks that support the necessary quality of service (QoS) whereas the reality is that network impairments such as latency, jitter and packet loss exist. In order to cope with jitter and delay, some VoIP applications employ time-scale modification or warping in the jitter buffer that adjusts the rate of playout while controlling the pitch to minimize Mouth-to-Ear (M2E) delay whilst preserving speech intelligibility and quality. In this paper, we firstly investigate the extent to which time-scaling occurs using WebRTC [1] VoIP clients over Wi-Fi networks with different levels of congestion. We then assess the impact of such time-scaling, both subjectively via expert listening test and objectively using POLQA, on quality experienced by the end user, and review the correlation between scores.

Time synchronisation plays a critical role in time-sensitive distributed applications. While a variety of such applications exist across many domains, one particular set of applications where improved time synchronisation can lead to significant benefits, particularly with respect to QoE (Quality of Experience), is multimedia applications. While time synchronisation is not a new challenge, advances in wireless technologies have drastically transformed network infrastructures. 802.11 wireless networks increasingly represent the last hop within the ever expanding Internet and whilst users expect the same levels of multimedia QoE as exist over wired networks, the reality of moving back to contention based access leaves many disappointed. This transformation of networks has also proven problematic for time synchronisation protocols that were designed for wired infrastructures. Wireless networks, particularly contention based networks, can be the source of very significant non-deterministic packet latencies. In certain scenarios, such latencies can greatly degrade the performance of time synchronisation. This work details and validates a technique that can be used to determine the latency of time messages in real-time as they traverse an 802.11 wireless link. Knowledge of these latencies can be used to greatly reduce the error in a dataset employed by time synchronisation protocols such as NTP and, thus, improve their performance. Experimental results confirm error reductions of up to 90% in a dataset and prove that the use of this technique can deliver time accuracies akin to those achievable over wired networks. This in turn can greatly benefit users by enabling multimedia applications to benefit from the continued use of time synchronisation for QoE management. We outline two such scenarios, one where time synchronisation is used to prioritise VoIP traffic within an Access Point and a second where the aim is to use time synchronisation to optimise jitter buffer strategies for WebRTC.

WebRTC is an open-source platform for real-time communications over the web and has been experiencing widespread adoption in recent years. WebRTC clients employ the technique of time scaling of packets to cope with the impact of network jitter and/or clock skew. A black-box study presented in this paper focuses on two aspects, namely time scale modification behaviour under different packet arrival interval and its impact on the listening quality perceived by the end user. Specifically, we examine the MOS scores predicted by the POLQA speech quality prediction model. Our tests involve both iSAC and Opus codecs, two of the widely used WebRTC codecs. In the experiment, a speech file played from one client, is directed through a network simulation before reaching the receiving client. Our results surprisingly show that the extent of time scaling is consistently higher for Opus producing shorter speech files. Regarding the consequent impact on quality, we also find that there are many cases where POLQA is reporting MOS predictions that contradict expert listener assessments.

We expect that WebRTC will experience high adoption rate as a peer-to-peer realtime communication standard for browsers. WebRTC allows direct media and data transport between browsers without having to go through a web server. In this study, we use a black-box testing technique to evaluate, via PESQ, the voice quality of WebRTC sessions under varying network delay and jitter. Network emulators are employed to implement the delay and jitter variations. Our results highlight the dangers of black-box testing, whereby test-bed issues can result in very misleading results. This is especially the case when executed on a single machine. This paper also provides an extendable baseline methodology for WebRTC centric research.