On March 27, 1977, a pair of fully loaded Boeing 747s collided on the runway of Los Rodeos Airport on the island of Tenerife, killing 583 of the 644 passengers on board the two jets, and forever changing aviation industry views of air safety and communication.
Investigators drew several conclusions from the crash, but the main one highlighted communication. “Language needs to be very clear,” the investigator said, highlighting the source of the confusion between pilots and air traffic control that led to the catastrophic collision.
Voice communication is always mission critical, but rarely results in human tragedy. Bad communications, however, can be a sure way to lose business. Degraded communications can cause callers to misunderstand each other, in some cases forcing them to call back over and over again and, if aggravated enough, could force them to try an alternative company.
What’s going wrong?
In the business world, unified communication systems manage a fast increasing amount of voice-based communication. Vendors like Microsoft have made it extremely simple to set up unified communications, but often users underestimate the planning required to keep voice functioning at peak performance when it must compete with other data and users sharing the same network. Things go wrong.
For example, VOIP systems that worked like a dream during a trial often fluctuate when call volumes spike and or network traffic increases. Suddenly that expensive ‘nice to have’ QOS switch becomes essential and would have paid for itself. These types of switches prioritise VOIP traffic over less critical network data so as to guarantee high quality voice transmission.
However, ignoring the importance of QOS is just the beginning of a long list of problems awaiting adopters of so called plug-and-play VOIP systems. Here’s a list of other common mistakes:
- Incorrectly managed IP network switches
- Inexperienced and untrained IT staff
- Absent ITIL discipline
- Poor voice/DATA models that neglect DIFFSERV/TOS and 802.1P/Q protocols, and LOSSLESS and LOSSY compression conventions essential to voice/telephony network design
- Bandwidth assumptions
- Inferior hardware, including switches, termination patch panels and cables
Getting a measure of call quality
While call quality, or rather instances of poor call quality, are immediately apparent to callers, getting a more formal measure of quality helps network managers understand the performance they should expect, and diagnose weaknesses likely to degrade voice quality.
The most reliable predictor of quality is based on the MOS Standard (Mean Opinion Score) – a transmission planning tool based on the ITU-T G.107 standard that calculates expected voice quality based on the experience of a typical telephone user under conversational conditions.
A number of communication vendors, including Avaya, have adopted this MOS standard for VOIP quality measurement.
Tools like VIRSAE VQM and Prognosis, use the E-model (ITU-T Rec. G.107) to detect and remedy voice issues. In the case of VIRSAE VQM, RTCP packet data is captured and uploaded to the VIRSAE cloud, which provides the ability to analyse the MOS data remotely and also gives the ability to alert.
Without these tools the investigation of voice quality problems is difficult. Real-time built in tools like those on the AVAYA CM system could be used, but require special skills and additional software to decipher the information – a job that is further complicated when data is gathered across a WAN, ISPs and telco providers.
7 quick fixes
Before splashing out on new tools to improve call quality, network managers can do a number of things to make immediate improvements.
- Calm jitter
VoIP delivers voice information in packets, which must be evenly spaced and delivered in a constant stream to preserve voice quality. But configuration and network errors disorder voice packets, scrambling audio and making it difficult for callers to understand what’s being said. One way to fix jitter is to replace Ethernet cables with a Category 6 cable and use high-end switches with jitter buffers to accelerate the transfer of information. The combination of cable and switching stores and organises voice packets, keeping everything in the right order.
Like all services, VoIP communication competes with other applications and users on the network – and as competition intensifies, call quality suffers. The easiest way to free up bandwidth is with a contention ratio that measures the difference between available bandwidth and the possible maximum demand for service, and allocates bandwidth accordingly.
- Don’t overlook the simple things
Sometimes it’s the simple things that make the biggest difference to call quality – like headsets and cords. Cheap or aging IP telephones and headsets often have thin, poorly-insulated cables and reduced audio clarity design. High-standard approved IP telephones, headsets and cables are more expensive, but cost less in the long run.
- Monitor bandwidth usage
Limit downloads (and block YouTube) and other network activity during business hours. Reschedule patching, backups, and file transfers to afterhours, or during low business traffic periods. Consider running regular network speed tests on computers, servers, and network switches to identify patterns of network usage and the times of day best suited for outbound calls.
- Calibrate quality of service
Adjust your network’s QOS features to prioritise VoIP calls above other applications. Use your router to prioritise VoIP traffic to ensure calls always have sufficient bandwidth, regardless of other activity on the network. For example, when VOIP traffic is carried over switched Ethernet LANs (where the QOS is excellent and there is plenty of bandwidth for voice and data functions), use 64 kbps G.711 PCM voice coding rather than G.723.1 or G.729A to reduce the processing requirements of IP telephony. Should compression be required, the PBX (or Ethernet Hub) will do the job. This setup minimises DSP processing requirements and paves the way for a lower cost DSP, or even a single RISC processor core for the entire IP telephony system that does the work of the DSP/MCU combo. The RISC processor performs Host Signalling Processing (HSP) functions that handle basic voice and other processing requirements. However, a DSP could still be required to manage more intense processing functions, including acoustic echo cancellation, multiple lines, and conferencing features.
- Compression software
Like all systems that involve digital signals, VoIP networks compress data to eliminate as many data bits as possible. LOSSLESS compression eliminates redundant bits, leaving information essentially the same – perfect for sending faxes over your VoIP network. LOSSY compression eliminates all information deemed to be unnecessary, which makes it ideal for audio conversations, as it minimises bandwidth use and eliminates ambient noises unrelated to the VoIP conversation. However, compression software can eliminate too much information, making conversations unintelligible. In this case you should buy new compression software that preserves a larger portion of the original data.
- Control feedback
Turn down the volume on your speaker phone. Failing that, use a phone with a lower frequency (high frequency phones pick up more ambient sound, increasing the potential for feedback).
Mastering quality is harder than it looks
Thanks to packet voice networking, almost any network-enabled device can carry voice. However, implementing VOIP technology requires sophisticated real-time software to address QOS and network interoperability. Delivering consistently high quality voice is much harder than it looks. But with the right set of eyes and toolsets like Virsae VQM, network managers can keep voices crystal clear and avoid the pitfalls of crossed wires and miscommunication.
Contact Pyrios if you’d like help improving your VOIP call quality.
Author: Louis Barbosa