Users describe call experience as ‘good’ ‘ok’ ‘poor’ ‘bad’ ‘terrible’. And call experience is defined by the elements of voice quality and the network factors that affect them. Elements of voice quality are loudness, distortion, noise, fading and cross talk whereas network factors that affect them are latency, jitter, packet loss, voice activity detection, echo and echo canceller performance.
How do these factors affect voice quality?
- Latency: is the delay or the time it takes for the speech to get from one designated point to another. Very long latency results in delay of hearing the speaker at the other end.
- Jitter: is the variance of inter-packet delay. When multiple packets are sent consecutively from source to destination and if there are delays in the network like queuing or arriving through alternate routes, the arrival delay between packets is the jitter value. For delay-sensitive applications like VoIP a jitter value of 0 is ideal.
- Packet loss: occurs when data packets are discarded at a given moment when a device is overloaded and unable to accept incoming data. You need to keep packet loss to the lowest value possible. For VoIP, packet loss causes parts of the conversation to be lost.
- Voice Activity Detection: is used in VoIP to reduce bandwidth consumption. When this technology is used, the beginnings and ends of words tend to be clipped off, especially the "T" and "S" sounds at the end of a word.
- Echo: is the sound of the speaker's voice returning to and being heard by the speaker. Echo is a problem of long round-trip delay. The longer the round-trip delay, the more difficult it is for the speaker to ignore the echo.
- Echo Canceller Performance: The echo canceller remembers the waveform sent out and, for a certain period of time, looks for a returning waveform that it can correlate to the original signal. How well the echo is cancelled depends on the quality of the echo canceller. If the return signal (echo) arrives too late, the echo canceller won't be able to correlate and cancel it properly.
- CODEC: stands for coder-decoder, converts an audio signal into digital form for transmission and then back into an audio signal for replay. CODECs also compress the packet to gain maximum efficiency from the network. How well the CODEC converts speech to digital packets and back again affects voice quality. Choosing the right codec for the network depends on the required sound quality, available network bandwidth and so on. Some networks use more than one codec but this again may impact call quality.
The index to measure the call quality using network data is called the Mean Opinion Score (MOS).
Mean Opinion Score (MOS)
MOS is a benchmark used to determine the quality of sound produced by specific codecs and the opinion scores are averaged to provide the mean for each codec sample. It is always preferable to have a MOS score of 4 or 5 for your VoIP calls. When the MOS decreases to 3.5 or below, users find the voice quality to be unacceptable. It is used to assess the performance of codecs that compress the audio.
Measuring VoIP performance using MOS
Test Infrastructure Readiness for VoIP Traffic
When implementing VoIP, it is a good practice to test the network for its readiness to carry voice traffic. But how can you accomplish this without spending CAPEX on VoIP infrastructure? Cisco devices use IP SLA to generate synthetic VoIP traffic and collects data to measure metrics like latency, jitter, packet loss and MOS. The MOS score is an indication of what voice quality to expect. You can start by troubleshooting network devices in the route of VoIP calls and those with an MOS score lower than 3. However, configuring IP SLA operations requires good knowledge of CLI (Command Line Interface).
Troubleshoot Poor VoIP Performance In the Network
Manually troubleshooting VoIP issues involves collecting performance metrics like jitter, latency, packet loss, etc., from various nodes in the network such as switches, routers or call managers. But this does not provide a standard to compare and understand the VoIP call quality in the network. MOS acts as a standard for measuring call quality. For a network experiencing poor VoIP performance, you can pin point the root cause depending on the MOS scores measured for a specific codec or at a particular time or for a particular department, location.
In summary, MOS scores are good indicators for troubleshooting VoIP performance issues in the network. Tools like SolarWinds VoIP and Network Quality Manager (VNQM) lets you enable IP SLA operations on your devices without knowledge of CLI commands, as well as helps avoid time consuming manual configurations on multiple devices.
SolarWinds VNQM also provides automated reports with locations and MOS scores, comparison of MOS scores per codec for each call, comparison of call performance metrics between departments or call managers and notification via email in case of bad calls.
Reduce time needed to evaluate network readiness and troubleshoot VoIP performance issues from hours to minutes!