In brief: Poor VoIP call quality is the most common complaint in business phone system deployments. This guide covers the key quality metrics, the tools available to measure them, and the diagnostic process for identifying and fixing audio issues in SA VoIP deployments.
What Makes VoIP Call Quality Good or Bad
VoIP call quality is determined by how well the network transports real-time audio packets between two endpoints. Unlike data transfers where packets can arrive in any order and be retransmitted if lost, voice calls require packets to arrive in order, on time, and with minimal variation. The key metrics are:
Latency (delay): The time for a voice packet to travel from sender to receiver. ITU-T G.114 recommends a one-way delay of less than 150ms for acceptable call quality. SA domestic VoIP calls typically achieve 5–20ms one-way delay on good fibre connections. Delay above 300ms one-way produces noticeable conversational awkwardness; above 400ms is typically unacceptable.
Jitter: The variation in packet arrival timing. If packets consistently arrive with 20ms spacing but occasionally arrive 50ms or 5ms apart, the jitter buffer in the phone or softphone must compensate — enlarging the buffer adds delay; shrinking it risks audio dropouts. Jitter below 20ms is generally transparent to users; above 40ms causes audible quality degradation. Jitter is most commonly caused by network congestion and poor QoS.
Packet loss: Packets that never arrive. Even 0.5% packet loss is noticeable as audio “clipping” or momentary silence. At 3% packet loss, call quality is significantly degraded. Good business networks achieve <0.1% packet loss. The causes of packet loss include congested links, failing network hardware (cables, switches, routers), and wireless interference.
Codec: The algorithm used to encode voice into digital packets. G.711 (64 kbps, uncompressed, highest quality) is the standard for SA SIP trunks. G.729 (8 kbps, compressed, lower bandwidth but slightly lower quality) is used where bandwidth is constrained. opus codec is increasingly used for WebRTC calls. The codec contributes to the theoretical maximum quality; network factors then determine whether that quality is achieved.
VoIP Quality Testing Tools
MyBroadband Speed Test (mybroadband.co.za/tools/speedtest): SA’s best-known speed test, useful for measuring available bandwidth but doesn’t directly test VoIP-relevant metrics like jitter and latency. Good for initial connectivity verification.
PingPlotter: A network diagnostic tool that sends continuous packets to a target server and displays latency, jitter, and packet loss over time. Invaluable for identifying intermittent quality issues that don’t show up in single-point speed tests. Run PingPlotter against your SIP trunk provider’s SIP server or the hosted PBX server for 30–60 minutes during business hours to capture peak load behaviour.
3CX Quality of Service Report: 3CX V20’s management console includes a call quality reporting feature that scores each call based on MOS (Mean Opinion Score — a standardised voice quality measure from 1 to 5). Calls scoring below 3.5 MOS are flagged for review. This built-in reporting allows proactive quality monitoring without external tools.
Yeastar CDR Quality Reports: Similar to 3CX, Yeastar P-Series logs call quality metrics in its call detail records. The P-Series’s quality report view allows filtering by date, extension, or queue to identify patterns in poor-quality calls.
VoIPmonitor: Open-source VoIP monitoring software that captures and analyses SIP/RTP traffic on the network, providing per-call quality metrics, codec information, and detailed audio quality statistics. More powerful than built-in PBX reporting for deep diagnostics.
SIPP (SIP Performance testing): For load testing — generating large numbers of simulated calls to verify that a system performs correctly under maximum expected load. Relevant for contact centre deployments and high-volume environments before go-live.
Diagnosing Common Quality Issues
Echo on calls: Caller hears their own voice echoed back. Usually caused by acoustic echo (sound from the speaker being picked up by the microphone) or electrical echo in the network. Acoustic echo: ensure headset is being used rather than speaker/microphone, or enable echo cancellation in the IP phone/softphone settings. Electrical echo: may indicate a misconfigured SIP trunk with improper echo cancellation settings.
Choppy/robotic audio: Audio that cuts in and out or sounds distorted. Almost always a jitter or packet loss problem. Check PingPlotter results for jitter peaks and packet loss events. If jitter is high during specific times (peak internet hours), the internet connection’s contention ratio is the likely cause. If packet loss is present consistently, check physical network components (switch, cables, WiFi).
One-way audio: One party can hear the other but not vice versa. Common cause: NAT traversal issue where the RTP audio stream can be initiated but not returned. Check STUN settings in the PBX, verify UDP ports for RTP are not blocked by a firewall, and confirm the PBX’s external IP is correctly configured.
Calls dropping after exactly 30 seconds: A highly specific symptom almost always caused by a SIP ALG (Application Layer Gateway) on a router or firewall intercepting and corrupting SIP packets. The 30-second timeout corresponds to the SIP re-INVITE cycle. Fix: disable SIP ALG on the router (check router settings under “SIP helper,” “SIP ALG,” or “application firewall”).
The Diagnostic Process
When a client reports call quality issues, work through this sequence:
- Collect evidence: when does it happen, on all calls or specific numbers, affecting all staff or specific locations/devices?
- Run PingPlotter against the SIP provider for 30+ minutes while the issue is occurring
- Check PBX quality reports for affected calls — MOS score, packet loss, jitter reported at PBX level
- Test with a different internet connection (tethering to mobile) to isolate ISP vs. network issue
- Test with a different endpoint (different phone, softphone, or location) to isolate device vs. network issue
- If network-specific: check for SIP ALG, review QoS settings on router, check for duplex mismatch on switch ports
What This Means for SA Telecoms Resellers
Building systematic quality testing into your deployment process — running a 30-minute PingPlotter test, reviewing the first week’s call quality reports, and documenting the baseline metrics — gives you the evidence to diagnose future issues quickly and demonstrate that quality has been maintained over time.
Call quality is the ultimate measure of a VoIP deployment’s success. Everything else — features, apps, reporting — matters less if the actual call quality is poor. SA resellers who develop strong quality diagnostic skills and fix problems quickly build the kind of reputation that generates referrals and protects against competitive displacement.
About the Author
Mechelle Gindra
Mechelle Gindra is the managing editor and lead content strategist at Telecoms-Channel, South Africa's dedicated intelligence platform for telecoms resellers, ICT distributors, and channel partners. She oversees the editorial team's AI-assisted content production pipeline, ensuring vendor news, market analysis, and regulatory updates are accurate, timely, and actionable for the SA telecoms channel community. With deep expertise in the South African telecoms reseller ecosystem, Mechelle directs coverage of vendor partner programmes, ICASA regulatory developments, wholesale market dynamics, and channel business strategy. She is responsible for maintaining Telecoms-Channel's editorial standards, fact-checking processes, and corrections policy.
