This Recommendation1 describes algorithmic models for measuring the visual quality of IP-based video services.

The models are hybrid no-reference encrypted (Hybrid-NRe) models, which operate by analysing packet header information and video image data captured at the video player. The models operate without parsing or decoding the packet payload. Thus, these models can be used with encrypted bitstream data as well as non-encrypted bitstream data.

As output, the models provide an estimate of visual quality on the [1,5] mean opinion score (MOS) scale, derived from five-point absolute category rating (ACR) as in [ITU-T P.910]. The models address low-resolution (VGA/WVGA) application areas, including services such as mobile TV, as well as high-resolution (HD) application areas, including services such as IPTV.

This Recommendation is to be used with videos encoded using ITU-T H.264 and media payload encapsulated in RTP/UDP/IP packets for the low resolution and encapsulated in MPEG-TS/RTP/UDP/IP for the high resolution.

The models in this Recommendation measure the visual effect of spatial and temporal degradations as a result of video coding, erroneous transmission or video rescaling. The models may be used for applications such as to monitor the quality of deployed networks to ensure their operational readiness or to benchmark service quality. The models in this Recommendation can also be used for lab testing of video transmission systems.

The models identified in this Recommendation have limited precision. Therefore, directly comparing model results can be misleading. The accuracy of models has to be understood and taken into account (e.g., using [ITU-T J.149]).

The validation test material consisted of video encoded using different implementations of [ITU-T H.264]. It included media transmitted over wired and wireless networks, such as WIFI and 3G mobile networks. The transmission impairments included error conditions such as dropped packets, packet delay, both from simulations and from transmission over commercially operated networks.

The following source reference channel (SRC) conditions were included in the validation test:

• 1080i 60 Hz (29.97 fps);
• 1080p (25 fps);
• 1080i 50 Hz (25 fps);
• 1080p (29.97 fps);
• SRC duration: HD: 10 s, VGA/WVGA: 10 s or 15 s (rebuffering);
• VGA at 25 and 30 fps;
• WVGA at 25 and 30 fps.

¹ This Recommendation includes an electronic attachment containing test vectors, including video sequences, bitstream files and predicted objective model scores.