ITU members have approved a new set of algorithmic models to monitor the quality of streamed audiovisual media. The model algorithms set out in the new ITU-T P.1200 series provide objective models for non-intrusive monitoring of the quality of User Datagram Protocol (UDP)-based video services, addressing both lower and higher resolution applications such as mobile TV and Internet Protocol TV (IPTV).
A major challenge in the management of IP-based networks is the provision of adequate Quality of Experience (QoE) and Quality of Service (QoS) for new multimedia services and applications such as IPTV, 3D video, and Internet media including over-the-top (OTT) video.
ITU-T Study Group 12 (Performance, QoS and QoE) is the expert group responsible for the ITU-T P.1200 series which offers tools to assess video quality in a scalable manner; proposing quality estimation models which operate by analysing packet-header and bitstream information as available from respective packet-trace data.The ITU-T P.1200 models are ‘no-reference’ or ‘non-intrusive’ in that their primary quality prediction does not require prior knowledge of the audiovisual stream being transported, being based solely on information as it is available on the network or inside terminals.
Recommendation ITU-T P.1201 “Parametric non-intrusive assessment of audiovisual media streaming quality” provides an overview of algorithmic models for non-intrusive monitoring of the audio, video and audiovisual quality of IP-based video services based on packet-header information. ITU-T P.1201.1 and ITU-T P.1201.2 specify model algorithms for lower and higher resolution application areas, respectively.
Recommendation ITU-T P.1202 “Parametric non-intrusive bitstream assessment of video media streaming quality” complements ITU-T P.1201 with an overview of bitstream-based algorithmic models for non-intrusive monitoring of video quality based on packet-header and bitstream information. ITU-T P.1202.1 and ITU-T P.1202.2 specify model algorithms for lower and higher resolution application areas, respectively.
The ITU-T P.1201 models cover audio, video and audiovisual quality estimates while the ITU-T P.1202 models focus on video quality, but the key difference between the two Recommendations is that ITU-T P.1201 models are restricted to the use of packet-header information whereas ITU-T P.1202 models exploit further bitstream information such as coding-related information.
Consequently, ITU-T P.1202 models can be more accurate in their quality predictions but in turn require non-encrypted streams to enable access to payload information. In addition, ITU-T P.1202 models are more complex and thus demand more computational power.
The model algorithms specified by the ITU-T P.1200 series provide individual estimates of audio, video and audiovisual quality in terms of the 5-point absolute category rating (ACR) mean opinion score (MOS) scale (described in Recommendation ITU-T P.910). The models are also capable of producing diagnostic information on the causes of quality degradations, as different types of performance parameters are derived during model calculations.
The series will in the future be extended to include transmission control protocol (TCP) and OTT video (considering both adaptive and non-adaptive streaming), and another natural progression of this work will be towards the monitoring of the video and audiovisual quality of 3D video.
More information on the work of ITU-T Study Group 12 can be found here.