ITU-T P.1202 PDF

This Recommendation describes recommended objective models for non-intrusive monitoring of the video quality of IP-based video services based on packet-header and bitstream information. This Recommendation addresses two application areas:

– [ITU-T P.1202.1] specifies the model algorithm for the lower resolution (LR) application area, including services such as mobile TV.

– [ITU-T P.1202.2] specifies the model algorithm for the higher resolution (HR) application area, which includes services such as IPTV. This application area is currently under study but not yet completed.

These models are restricted to information contained in packet headers, the packet bitstream (payload information), prior and static knowledge about the media stream and dynamic buffering information from the client.

[ITU-T P.1202.1] consists of one model. [ITU-T P.1202.2] consists of two models corresponding to two modes: mode 1 and mode 2, which both are no-reference (i.e., non-intrusive) models. Mode 1 refers to a parsing mode; the model operates by analysing information in the video bitstream without fully decoding the bitstream (i.e., no pixel information is used) for MOS estimation. Mode 2 refers to a full decoding mode, in addition to the bitstream information which mode 1 uses, the model can also decode parts or all of the video bitstream (i.e., pixel information is used) for mean opinion score (MOS) estimation. Further client specific information, such as concealment type, is provided to the algorithm out-of-band, for example in the form of stream specific side information.

These models predict video quality in terms of MOSs on a five-point ACR scale (see [ITU-T P.910]).

The primary application for these models is the monitoring of transmission quality during service operation or for maintenance purposes. The ITU-T P.1202 model may be deployed both in end-point locations and at mid-network monitoring points. The location of the model and the location of the measurement probe together determine the mode of operation, as described in more detail in clause 6.1.

The primary quality prediction made by such models is based on the payload of the stream being analysed. Therefore, this Recommendation can provide a comprehensive evaluation of quality as perceived by a particular end-user because its scores can reflect the impairments on the coding and the Internet Protocol (IP) network being measured, which differ from user to user. This Recommendation cannot provide a comprehensive evaluation of video quality as perceived by a particular end-user, because its scores reflect the impairments due to encoding and the subsequent IP network being assessed, which may only be one part of the end-to-end connection. An explicit inclusion of processing steps such as content contribution from e.g., satellite networks, display properties etc. are not considered. See Table 3 for more information. Further, the quality-impact due to a specific video encoder implementation or a specific decoder-side packet loss concealment implementation is not explicitly addressed. Instead, the models have been developed for a set of dedicated service implementations, which are assumed to be meaningful representations of today's IP-based streaming video services. As a consequence, however, in case of significant deviations of a given service being assessed from the service configurations used for developing this standard, it is possible to obtain high quality scores with this Recommendation and yet to have a poor quality of the stream as it is perceived by actual users.

As a consequence, this Recommendation can be used for applications such as:

– in-service quality monitoring for specific IP-based audiovisual services, as specified in more detail in Tables 4 to 7;

– benchmarking of different service implementations. However, it cannot be used for direct benchmarking of different decoder implementations. The implementations that can be assessed with ITU-T P.1202 include the encoding aspects and potential packet loss.