Network Quality Tests

Tests to assess network quality can be divided into two distinct groups:

Non-intrusive tests - Measurements are made on in-service trunks, so service is not affected or disrupted. These tests can accurately measure many aspects of network performance, and allow the statistical distribution of aspects such as call levels, noise levels, echo path loss and delay to be derived. However, in circumstances where the network contains equipment such as low bit-rate speech codecs and DCME, non-intrusive measurements may not provide a reliable indication of overall voice quality.

Intrusive tests - Intrusive measurements require the trunk or circuit to be temporarily taken out of service, and so are particularly appropriate during commissioning of new circuits or networks. Parameters of interest are measured directly. As the trunks are not in service it is not possible to make statistical measurements of aspects such as call levels. It is, however, possible to use techniques such as Perceptual Analysis Measurement System (PAMS) analysis to make an objective assessment of overall voice quality. PAMS will provide a reliable quality indication even if low bit-rate speech coding equipment or DCME is present in the network.

Non-intrusive tests
These are the most common tests employed to assess PSTN quality.

The unique feature of DSPG's non-intrusive test method is the use of voice recordings on selected 2Mbit/s trunks. The voice recordings are then used for detailed off-line measurement and analysis. This approach is more reliable for the purposes of overall quality analysis than making instantaneous measurements at the source.

Recorded wave-files can be retained for future reference if required, or deleted.

To be in compliance with the data protection act the recordings are made very short (approx. 30sec). This time is more than sufficient for all voice quality measurements to be performed.

Typically, two types of recordings are made:

Semi-automated voice quality recording on all channels of selected 2Mbit/s trunk. Since all channels are recorded and analysed, these recordings and follow-up measurements provide good statistics regarding the overall voice quality of the trunk under test.
The Voice Snapshot Recorder is used to record the data.

Manual voice quality recording on selected channels of selected 2Mbit/s trunk. These recordings are performed manually, and usually only channels carrying poor quality calls are selected. Measurement and analysis is then performed, and a corrective action applied if possible.
The Universal Channel Analyser is used to record the data under control of the operator.

The two types of recordings can be conducted concurrently for improved efficiency. All recordings are made digitally and stored on the test equipment's internal hard disk.

Voice Quality analysis software
The data from both types of recordings, stored in a special wave-file format, can be analysed using a software package developed by DSPG to automate the process. The analysis software resides on the same disk as the wave-file recordings, which are analysed file by file to produce the results.

All measurement methodology has been thoroughly verified.

The following parameters are measured:

  • Echo path Delay
  • Echo Path Loss
  • Speech Level
  • Noise level
  • ERL

    The accuracy of the calculations has been verified against results of various audit measurements performed by DSPG.

    To calculate the Call Clarity Index (CCI) other measurements have to be made, as follows:

  • Activity factor
  • PCM threshold violation
  • Double-talk
  • One way transmission
  • Front-end clipping

    These measurements are necessary to allow proper derivation of CCI, as specified in ITU Rec. P.562.

    Intrusive tests
    In certain cases, it may be required to evaluate the end-to-end connection quality between specific types of access technology. In these cases, and in the case of new networks, intrusive testing may be appropriate.

    Intrusive measurements are usually single-channel measurements. Defined test signals are injected at one network location, and measurements of the desired parameters are made at another location (or at the same location in the case of echo and delay). Loopback test methods may be suitable for some purposes.

    The Universal Channel Analyser may be connected at virtually any network access point, and fulfils the roles of both a signal injection device and a measurement and analysis device. Examples of parameters that can be examined are signal levels within the network, echo levels and ERL, and end-to-end delay across a network or any sub-section of that network. Measurements to derive the Call Clarity Index are also possible.

    Intrusive testing may also be desirable if voice compression systems (Voice Codecs), DCME and other similar equipment areused in a network, especially if such items are connected in tandem.

    In these circumstances, the traditional non-intrusive quality measurement methods described in section 2.1, cannot accurately describe the overall voice quality of a connection. Only by employing particular intrusive test methods can more accurate results be achieved. Perceptual Analysis Measurement System (PAMS) analysis is a suitable technique.

    DSPG equipment can perform PAMS tests in two ways:

    Trunk testing
    This is most commonly used during commissioning of 2Mbit/s trunks, when there is no live traffic. DSPG's Universal Channel Analyser is connected to the trunk under test and Perceptual Analysis (PAMS) measurements are made. The results are supplied in Mean Opinion Score (MOS) scale format. Where DCME or other form of competitive access equipment forms part of the network, DSPG's Network Loader equipment can be used to reproduce any desired loading, so that connection quality can be evaluated in either typical or worst-case conditions.

    Channel testing
    This test can be performed on any channel within a 2Mbit/s trunk, without a need to take the 2Mbit/s trunk out of service. The ability to leave the 2Mbit/s trunk in service can be a considerable advantage. The Universal Channel Analyser is connected to both the 2-wire and 4-wire parts of the network. The call is made on the 2-wire connection and a reference signal fed into the line. The PAMS measurements are made on the 4-wire (both Transmit and Receive).

    A DSPG Audit
    When commissioned to perform a network audit, DSPG will use the equipment and adopt the techniques described in the earlier sections of this document. The precise nature of the tests will depend on the customer requirements and objectives, and whether intrusive or non-intrusive testing is required. Overall goals are to be as objective as possible, to extract as much quantitative information from the measurements as practical, to present this information in a consistent and easily interpreted fashion, and to make recommendations for improvement where appropriate. A typical sequence of operations would be as follows:

  •  

    • Liase with customer and site personnel regarding site access, measurement logistics, and timescales
  •  
    		• Perform the appropriate network measurements
  •  
    		• Analyse results and make any proposals considered necessary for improvement
  •  
    		• Assist in the implementation or bread-boarding of proposed solutions, where appropriate
  •  
    		• Repeat measurements and analysis, as necessary, to evaluate effectiveness of solutions
  •  
    		• Provide measurement data and full report on findings to the customer

    Where network improvements are found to be necessary and effective, a network operator may wish to duplicate the improvements at other network locations in order to achieve a consistent, high quality network. Once a good quality network is achieved, the audit results may prove useful in the event of customer complaints. If the complaints are regarding connection quality, the measurement information can be used to demonstrate that the problem arises outside the operator's network.

    •