I-Lab: Quality of experience
The quality of experience (QoE) is often used in the information technology and consumer electronics domain to indicate the overall satisfaction with the service users receive.
I-Lab is a leading research centre which has longstanding experience in 3D audio-visual QoE. Its research portfolio extends over fundamental research on human visual perception, modelling 3D audio-visual experience, investigating quality of business (QoB), and modelling the impact of ambient environmental factors on perceived quality.
Traditionally, communication service providers assumed service level parameters such as quality of service (QoS) to quantify the level of service users experience. However, service level measures do not provide sufficiently accurate picture on the user experience on ever growing number of different services operated on state-of-the-art communication service platforms. QoE, in contrast, provides more comprehensive measure which incorporates not only QoS but also content quality and service quality such as cost, reliability, availability, usability, and fidelity. Hence, this is a rather subjective measure, which is immensely user dependent and hence modelling QoE objectively is quite challenging. Nevertheless, many industries, including audiovisual broadcasting industry, would be benefited from an objective model, which can assess the QoE on the service they provide reliably, quickly and cost effectively. The main focuses of the research carried out in the I-Lab are:
- Analysing sensitivity of the HVS to depth perceived by different cues
- Modelling 3D audio-visual experience
- Investigating quality of business (QoB)
- Modelling the impact of ambient environmental factors on perceived quality.
Several different cues are made use by humans to perceive the depth of different objects in a scene. The depth cues can be classified mainly in to two categories, namely, oculomotor cues and visual cues. 3D video provides an additional experience of depth while watching 3D video as compared to 2D video. 3D display systems provide additional cues to its viewers that enhance the viewers’ perception of depth in a video scene.
The most important one of these additional cues is the binocular disparity, which is obtained by providing two views of the same scene, from slightly different perspectives, to the each eye of a viewer. The eye positioning and simulated depth levels are explained in figure 1.
In order to understand the sensitivity of the human visual system (HVS) to fundamental depth cues, I-Lab has been conducted psyco-physical analysis of depth perception in 3D vision.
In this research the sensitivity of the HVS for different depth cues in 3D video is theoretically analysed and subjectively validated. The aim of the experiments was to model how much sensitive are the humans for depth cues such as binocular disparity, retinal blur and relative size.
Some important results obtained from these experiments are shown in figure 2, 3 and 4.
Figure 1: Eye positioning while 3D viewing.
Figure 2: Average of just noticed difference in depth at various testing disparity levels (Viewing distance = 2m).
Figure 3: Subjective results for identifying blur as a depth cue.
Figure 4. Sensitivity of relative size as a depth cue.
The high level view of a 3D audio-visual QoE model is depicted in figure 1. I-Lab is actively engaged in research on modelling spatial audio experience and 3D visual experience. The spatial audio experience is modelled using four key performance indicators (KPI).
These are the basic audio quality, front and surround spatial fidelity, and correlation between audio and video. In contrast, the visual experience can be modelled using two KPIs, which are the image quality (IQ) and depth perception (DP). These KPIs are modelled using the knowledge on psycoaccustic and psycovisual information gained from fundamental research. Subsequently, the theoretical models are verified using extensive subjective experiments carried out in the state-of-the-art laboratories readily available in the I-Lab.
A theoretical model developed for predicting the perceived depth distortion due to depth map compression is shown in figure 4. Results obtained for audio-visual congruency experiment is shown in figure 4. Finally, 18 shows some results obtained through subjective experiment to assess the impact of the quality of audio and video components on overall audio-visual perception.
Figure 5: 3D audio-visual QoE model.
Figure 6: Mean disparity distortion model (MDDM) for perceived depth distortion due to depth map compression colour-plus-depth 3D video.
Figure 7: Means and 95 per cent confidence intervals of QoE scores for different angular deviation of auditory scene from the video.
Figure 8: MOS vs. audio quality results for different video quality settings (LQ – low quality, MQ – medium quality, HQ – high quality).
Ambient environment has a significant impact on audio-visual experience. I-Lab has conducted extensive studies for quantitatively analyse the impact of ambient illumination on visual perception.
In summary these studies reveal that the video quality perception ratings of the viewers also increase with the ambient illumination. This is because the human eye only receive relatively low amount of light produced by the display when it is seen under bright ambient illumination conditions.
Therefore, human visual system cannot easily detect visual artefacts (due to compression, transmission, rendering, etc) under bright environment as much as they are noticeable in a dark environment since. Subjective experimental results shown in 9 illustrate this effect.
Figure 9: Subjective mean opinion score (MOS) for perceived image quality vs. the channel bandwidth, at 4 distinct ambient illumination levels.
The “direct” impact of customer satisfaction on the payment they make for receiving a specific multimedia service is emphasized by the quality of business (QoB). This active role of QoB aims at maintaining the revenue of SPs and guaranteeing a better quality of experience and satisfaction for users.
The current multimedia services delivery models can be illustrated by reviewing a projection of the general triple-Q model on an IPTV service delivery model. 10 depicts an IPTV service delivery model with emphasis put on the components where QoS and QoE are considered. Observing the foregoing model, although it is believed to provide promising QoE to the user, it does not expose QoB precisely and literally. From a business point-of-view, this model focuses on the importance of user satisfaction trusting it will result in better customer loyalty, which is believed to generate more profit.
From this vision, it is clear that there is no direct link between profit and customer satisfaction. Instead, it is an indirect link emerged as a result of customer loyalty. Consequently, there is a gap between customer satisfaction (QoE) and business profit (QoB). Therefore, the research on QoB at I-Lab is aimed a bridging this gap to identify comprehensive models that employ QoB as a major player.
Figure 10: An IPTV service delivery model illustrating the components where QoS and QoE are considered.