Packet Loss — Causes, Effects, and Measurement
Learn about packet loss — causes, effects, and measurement
Packet loss is the most directly destructive impairment for voice quality. Latency makes conversations awkward; jitter makes audio choppy at the edges. But packet loss literally removes pieces of the conversation. Understanding the types, causes, and measurement of packet loss is fundamental to NOC operations.
Random Loss vs. Burst Loss
This distinction changes everything about the impact on audio quality.
Random (Isolated) Loss
Individual packets dropped independently — packet 50 is lost, then packets 51-80 arrive fine, then packet 81 is lost. At 1% random loss, you're losing one packet every 100, which means one 20ms gap every 2 seconds.
Impact: Modern PLC algorithms handle this well. Opus can conceal isolated losses almost perfectly because it has good surrounding context. The listener may not notice 1-2% random loss at all.
Burst Loss
Multiple consecutive packets lost — packets 50-55 all drop (120ms of audio gone). This happens when a buffer overflows and drops a queue's worth of packets, or when a link flaps briefly.
Impact: Devastating. PLC can interpolate one missing packet from its neighbors, but it can't reconstruct 100ms+ of audio from nothing. A 6-packet burst at 20ms packetization = 120ms gap — clearly audible as a "hole" in the conversation.
The math: 2% random loss produces occasional barely-audible concealment. 2% burst loss (concentrated in bursts of 5-10 packets) produces periodic, clearly audible dropouts even though the aggregate loss rate is the same.
💡 NOC Tip: en reviewing loss metrics, always look at the loss pattern, not just the percentage. RTCP reports aggregate loss over the reporting interval — they can't distinguish random from burst. You need packet captures to see the actual pattern.
Causes of Packet Loss
🔒 Full Course Content Available
This is a preview of the Telephony Mastery Course. The complete lesson includes:
- ✅ Detailed technical explanations
- ✅ Advanced configuration examples
- ✅ Hands-on lab exercises
- ✅ Real-world troubleshooting scenarios
- ✅ Practice quizzes and assessments
Preview shows approximately 25% of the full lesson content.