This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. IEEE TRANSACTIONS ON BROADCASTING 1 Towards a General Parametric Model for Perceptual Video Quality Estimation Jos´ e Joskowicz, IEEE Senior Member, Rafael Sotelo, IEEE Senior Member, and J. Carlos Lopez Ardao, IEEE Member Abstract —During the last few years, different parametric models were proposed for video quality estimation. Each model uses different parameters as inputs, such as bit rate, frame rate and percentage of packet loss, and each model was designed and tested by their authors for a particular codec, display resolution and/or application. This paper presents a review of the parametric models published by ten different groups of authors. Each model is briefly described, and the relevant parametric formulas are presented. The performance of each model is evaluated and contrasted to the other models, using a common video clips set, in different coding and transmission scenarios. Based on the results, a new and more general parametric model is presented, which takes into account bit rate, frame rate, display resolution, video content and the percentage of packet loss. Index Terms—Video perceptual quality, video quality parametric models, video signal processing I. Introduction I N RECENT years different evaluations and standardized efforts have been made, and are currently ongoing, in order to obtain objective models and algorithms to predict the perceived video quality in different scenarios. The video quality models can be classified into FR (Full Reference), RR (Reduced Reference) and NR (No Reference) models [1]. In the first class, FR models, the original and the degraded video sequences are directly compared. In the RR models, some reduced information about the original video is needed, and is used along with the degraded video in order to estimate the perceived video quality. NR models are based only on the degraded video in order to make an estimation of the perceived video quality. Based on the work of VQEG (Video Quality Experts Group) and other contributions, ITU-T has standardized some FR and RR models. Among them the Recommendation ITU-T J.144 [2] in 2004, the Recommendations ITU-T J.247 [3] and ITU-T J.246 [4] in 2008 and the Recommendation ITU-T J.249 in 2010 [5]. Manuscript received June 13, 2012; revised July 25, 2013; accepted July 26, 2013. J. Joskowicz is with the Facultad de Ingenier´ ıa, Universidad de la Rep´ ublica, Montevideo, Uruguay, (e-mail: josej@fing.edu.uy). R. Sotelo is with the Facultad de Ingenier´ ıa, Universidad de Montevideo, Montevideo, Uruguay, (e-mail: rsotelo@um.edu.uy). J. C. L. Ardao is with ETSE Telecomunicaci´ on, Campus Universitario, Vigo 36310, Spain (e-mail: jardao@det.uvigo.es). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TBC.2013.2277951 Parametric models predict the perceived video quality based on a reduced set of parameters that are related to the en- coding process, video content and/or transmission process (i.e. network information). These models typically present a mathematical formula, representing the estimation of the perceived video quality as a function of different parameters. Parametric models are easy to implement since there is no need to full access to the original video source. They may be applied to network design, network assessments and/or to real time monitoring. The quality estimation is easily computed as the result of a direct mathematical formula. Many different parametric models have been proposed, with different scopes and applicable to different scenarios, and a parametric model has been standardized in the Recommenda- tion ITU-T G.1070 in 2007 [6]. Each of the proposed paramet- ric models has been evaluated by their authors. However, they usually study them in a particular use case. Nevertheless, a general parametric model that would apply for a wide range of applications, encoding parameters and transmission scenarios has not been developed yet. In this paper we present a review of parametric models published in the last years by ten different groups of authors. The model’s parameters and performance are evaluated and compared. The strengths and weaknesses of each model are remarked and are employed towards the development of a general parametric model for video quality estimation. The paper is organized as follows: Section II describes main published parametric models. In Section III the performance of each model is presented. In Section IV the results of the previous section are analyzed, and based on them, a more general model is proposed. Section V summarizes the results and main contributions. II. Parametric Models Since this work aims to contribute towards a general para- metric model, this section presents relevant parametric models proposed in last years. Each one is just briefly described, while its parametric formula is thoroughly detailed. The presentation order is arbitrary and does not respond to any order of relevance. A. Kazuhisa Yamagishi et al.: ITU-T G.1070 Model ITU-T has published a model for predicting the video quality in video telephony applications, based on measurable parameters of the encoding process and the IP transport network. The Recommendation ITU-T G.1070 [6] describes a 0018-9316 c  2013 IEEE