Trunk testing is a method for verifying that a bearer circuit is available for voice transmission from the calling party to the called party. It does nothing to validate that the circuit will offer a high grade of voice quality. To validate that, the bearer circuit is providing an acceptable level of voice quality. It is necessary to validate the circuit's performance using a reference test signal and one of the three main QoS algorithms: PSQM, perceptual analysis measurement system (PAMS), or perceptual evaluation of speech quality (PESQ).

Available Quality Schemes

• PSQM: The PSQM (ITU–T P.861) algorithm was originally developed to test voice quality through a voice encoder/decoder (codec). PSQM is an intrusive quality measurement system that involves the comparison of a voice sample after it has passed through the codec to the original sample. This standard is currently being used to test voice quality over end-to-end voice paths, even though the standard cannot account for errors associated with packet delay, packet jitter, lost packets, duplicate packets, noise, or other factors that might affect voice quality in a VoIP environment. This limitation was confirmed in a study conducted at the Czech Technical University¹.
• PAMS: Like PSQM, PAMS is designed to give an automated voice-quality score, or rating, based on the comparison of a received voice signal to the original signal. PAMS, however, is designed for network-wide testing and is capable of taking into account delay, jitter, and other events that would affect voice quality in a VoIP network.
• PESQ: PESQ (ITU–T P.862) is the latest standard for assessing voice quality and is expected to eventually replace PSQM. It builds on the PSQM and PAMS algorithms by adding additional processing steps to account for signal-level differences and the identification of errors associated with packet loss.

The Role of Network Statistics in Measuring Voice Quality
PSQM, PAMS, and PESQ are all tools to give a numerical rating to voice quality. As such, they mask the underlying network problems by producing a particular voice-quality score. To determine the reasons for a score, the network statistics associated with a voice-quality test must be analyzed.

¹Statistical events of interest are jitter, delay, packet loss, and duplicate packets. Analyses of these factors allow network operators to correlate voice-quality issues with busy-hour occurrence or to determine whether a media gateway involved in the call is the source of the voice-quality issues.