In this paper the quality of adaptive bit rate video streaming is investigated and two state-of-the-art models, i.e., the NTT audiovisual quality-estimation and ITU-T P.1203 models, are considered. This paper shows how these models can be applied to new conditions, e.g., 4K ultra high definition (4K-UHD) videos encoded using H.265, considering that they were originally designed and trained for HD videos encoded with H.264. Six subjective evaluations involving up to 192 participants and a large variety of test conditions, e.g., durations from 10sec to 3min, coding-quality variation, and stalling events, were conducted on both TV and mobile devices. Using the subjective data, this paper addresses how models and coefficients can be transferred to new conditions. A comparison between state-of-the-art models is conducted, showing the performance of transferred and retrained models. It is found that other video-quality estimation models, such as VMAF, can be used as input of the NTT and ITU-T P.1203 long-term pooling modules, allowing these other video-quality-estimation models to support the specificities of adaptive bit-rate-streaming scenarios. Finally, all retrained coefficients are detailed in this paper allowing future work to directly reuse the results of this study.

Since the quality of video streaming services is degraded due to the encoding, network congestion, and lack of a playout buffer, the normality of services needs to be monitored by gathering the quality measured at the end clients. When measuring quality at the end clients, the computational power should be sufficiently low, the bitstream information cannot be accessed for the quality estimation due to the encryption, and reference video cannot be used at the end clients. Therefore, metadata-based models have been developed and standardized that take metadata such as the resolution, framerate, and bitrate, and stalling information as input and calculate the quality. However, calculated quality for linear TV and video on demand (VoD) services cannot be compared because metadata-based models cannot calculate the impacts of codec strategies (e.g., H.264/AVC, H.265/HEVC, and AV1) and configurations (e.g., 1-pass encoding for linear TV or 2-pass encoding for VoD) on the quality. To take into account the impact of codec strategies and configurations, coefficients of metadata-based model need to be optimized per codec strategy and configuration using subjective quality. However, extensive subjective assessment tests are difficult to frequently conduct because they are expensive and time consuming and need to be conducted by video quality experts. Therefore, to monitor the quality of several types of video streaming services (e.g., linear TV and VoD) and to compare these qualities, a derivation procedure is proposed for obtaining coefficients of metadata-based models using a full-reference model. To validate the procedure, extensive subjective assessment tests were conducted. Finally, it is shown that three metadata-based models (i.e., the P.1203.1 mode 0 model, extended P.1203.1 mode 0 model, and model proposed by Yamagishi et al.) based on the proposed procedure using the video multimethod assessment fusion (VMAF) algorithm estimate quality accurately in terms of root mean squared error.