PerCeption of impairments

Description

PerCeption of impairments (PC) refers to a framework and set of methodologies defined in 3GPP for assessing the quality of experience (QoE) for multimedia services, particularly how users perceive degradations like packet loss, delay, or jitter. It involves objective models that estimate subjective quality metrics, such as Mean Opinion Score (MOS), based on measurable network and media parameters. These models are essential for network operators to monitor, control, and optimize service quality without requiring constant subjective user testing.

Architecturally, PC is implemented through quality estimation algorithms that can be deployed in network probes, user equipment, or core network functions. These algorithms take inputs like codec type, bitrate, packet loss rate, delay, and jitter to compute a predicted perceptual quality score. In 3GPP systems, PC models are specified for various services including voice (e.g., using P.863 POLQA), video streaming, and conversational video. The outputs feed into QoS management systems, policy control, and network optimization engines.

How it works involves several steps: first, media and transport parameters are collected during a service session. These parameters are then processed by a standardized perceptual model (e.g., defined in ITU-T or 3GPP) that mimics human auditory or visual perception. The model outputs a score, often on a scale like 1-5 MOS, indicating the expected user satisfaction. This data can be used in real-time for adaptive bitrate streaming (e.g., in DASH) or post-analysis for network planning. In 3GPP, specifications detail parameter mapping, measurement requirements, and reporting procedures.

Its role in the network is integral to end-to-end quality assurance. PC enables proactive quality monitoring, helps in troubleshooting service degradation, and supports dynamic resource allocation. For example, in LTE and 5G, PC metrics can influence PCRF (Policy and Charging Rules Function) decisions to prioritize traffic or trigger handovers. It is a key component of Self-Organizing Networks (SON) and network automation, allowing operators to maintain high QoE efficiently.

Purpose & Motivation

PC exists to bridge the gap between objective network measurements and subjective user experience. Traditional network KPIs (e.g., throughput, latency) do not directly translate to how users perceive service quality. PC solves this by providing standardized models that predict perceptual quality, enabling operators to manage networks based on actual user satisfaction rather than just technical metrics.

Historically, quality assessment relied on expensive and slow subjective testing with human panels. The development of objective perceptual models, often from ITU-T (e.g., P.800, P.910), provided a scalable alternative. 3GPP adopted and specified these models for mobile environments to address the unique impairments of wireless networks, such as variable bandwidth and handover interruptions. This was motivated by the growth of multimedia services (VoIP, video streaming) where quality directly impacts customer retention.

The integration into 3GPP standards ensures consistency across vendors and operators, facilitating interoperability and benchmarking. It addresses limitations of earlier approaches that lacked standardized perceptual metrics, allowing for automated, real-time QoE management in complex mobile networks.

Key Features

• Objective estimation of subjective quality (e.g., MOS)
• Models for audio, video, and audiovisual services
• Inputs from network parameters (packet loss, jitter, delay)
• Standardized algorithms (e.g., based on ITU-T recommendations)
• Integration with QoS and policy control mechanisms
• Support for real-time monitoring and reporting

Evolution Across Releases

Defining Specifications