PLC stands for Packet Loss Concealment.
PLC is a technique used by audio decoders to mask the effects of lost packets by generating replacement audio that sounds natural to the listener. Rather than producing silence or audible glitches when a packet is missing, PLC algorithms synthesize audio that approximates what the lost packet likely contained.
How PLC works
PLC algorithms typically:
- Analyze the patterns in recently received audio (pitch, energy, spectral characteristics)
- Generate a synthetic audio frame that continues those patterns
- Gradually fade toward comfort noise if multiple consecutive packets are lost
Different PLC algorithms will offer different outcomes, varying in their complexity and quality.
The Opus codec used in WebRTC has good built-in PLC capabilities, which is one reason WebRTC audio quality remains decent even under moderate packet loss.
PLC vs other loss recovery
PLC is one of several packet loss mitigation strategies in WebRTC:
- PLC (concealment): Receiver generates replacement audio. Zero added latency
- NACK / RTX (retransmission): Sender retransmits lost packets. Adds one RTT of latency
- FEC (Forward Error Correction): Sender includes redundant data. Adds bandwidth overhead but no latency
PLC is always the first line of defense since it requires no network round-trip, making it the fastest recovery mechanism.
