ITU-T P.1204.3 PDF

This Recommendation¹ describes a bitstream-based video quality model that can be used: stand-alone as a video quality prediction model; or together with audio and integration modules to form a complete model to predict the impact of audio and video media encodings and observed Internet protocol (IP) network impairments on quality experienced by the end-user in multimedia streaming applications. The streaming techniques addressed comprise progressive download and adaptive streaming, for both mobile and fixed network streaming applications.

This model is defined to cover a range of use cases, from monitoring bitstreams where the video payload is fully encrypted, unencrypted bitstreams and where deep packet inspection is possible or where the bitstream is available at the encoding premises, e.g., from the client side. The model thus has a wide range of application, from encoding optimization over client-side quality of experience (QoE) assessment to network or service optimization or benchmarking purposes. The model in this Recommendation is bitstream based.

The model described here is applicable to progressive download and adaptive streaming or other streaming applications with reliable transport, where the quality experienced by the end user is affected by video degradations due to coding, spatial re-scaling or variations in video frame rates. Quality assessment of adaptive streaming includes aspects of media adaptation that may be handled in integration modules such as those of [ITU-T P.1203.3] and not in the video modules in this Recommendation. This Recommendation is able to handle various video codecs (i.e., H.264, H.265/high-efficiency video coding (HEVC) and video payload type 9 (VP9), resolutions up to 4K/ultra-high definition-1 (UHD-1) and frame rates up to 60 frames/s. In contrast to the video-quality module Pv of [b-ITU-T P.1203], i.e., [ITU-T P.1203.1], only addresses ITU-T H.264 and full high definition (HD) with up to 30 frames/s.

The model predicts a mean opinion score (MOS) on a five-point absolute category rating (ACR) scale (see [ITU-T P.910]) as an overall video quality MOS (5 s to 10 s). In addition to the overall quality score, this video quality model produces a per-one-second quality score, suitable for diagnostics or integration into an integral quality score for longer sessions (see, for example [ITU-T P.1203.3] for 1 min to 5 min duration sessions).

The model associated with this Recommendation cannot provide a comprehensive evaluation of the video quality as perceived by an individual end-user because the scores reflect the perceived impairments due to coded video media data being transmitted over an IP connection with certain performance and do not include specific terminal device or user-specific information. The scores predicted by such a general quality model necessarily reflect average perceptual quality.

Effects due to source generations, such as signal noise, video shake, certain colour properties (and other similar video factors) and other impairments related to the payload, are not reflected in the scores computed by this model.

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 clause 6.1;

– performance and quality assessment of live networks (including codecs) considering the effect due to encoding bitrate, encoding resolution and encoding frame rate;

– laboratory testing of video systems;

– benchmarking of different service implementations.

In particular, targeted applications are progressive download streaming and adaptive streaming (using reliable transport), which includes the following.

– Over-the-top (OTT) services, as well as operator-managed video services (over the TCP).

– Video over both mobile and fixed connections.

– The streaming protocols HTTP live streaming (HLS) or dynamic adaptive streaming over HTTP (DASH) used with the hypertext transfer protocol (HTTP) or HTTP2 over TCP/IP or quick user datagram protocol internet connections (QUIC), or real-time messaging protocol (RTMP) over TCP/IP. Note that the model is agnostic to the specific application or transport layer protocol, with the exception that it assumes reliable delivery of video packets.

– Video services typically using container formats based on the ISO/IEC base media file format such as Moving Picture Experts Group-4 (MPEG-4) Part 14 (MP4), or other container formats such as audio video interleave (AVI), Matroska video (MKV), WebM, Third Generation Partnership (3GP), and MPEG-2 transport stream (MPEG2-TS). Note that the model is agnostic to the type of container format.

¹ This Recommendation includes an electronic attachment with the Trees for final prediction announced in clause 8.2.