Real-time Transport Protocol

RTP (Real-time Transport Protocol) is the network protocol that carries the actual voice audio during VoIP calls. While SIP handles call signaling (setup, teardown, and negotiation), RTP handles the media — the encoded audio stream between endpoints using the negotiated codec. Each active call uses a pair of UDP ports for sending and receiving RTP packets.

How It Works in VICIdial

In VICIdial’s Asterisk configuration, the RTP port range is defined in rtp.conf — typically ports 10000-20000. Each concurrent call consumes two RTP ports (one for audio, one for RTCP control), so a server handling 500 concurrent calls needs at least 1,000 available ports. The SIP or PJSIP signaling negotiates the specific RTP ports during call setup via the SDP (Session Description Protocol) embedded in SIP messages.

RTP packets are time-sensitive — they must arrive in order and with minimal delay (under 150ms one-way) for clear audio. VICIdial servers should be configured with QoS (Quality of Service) rules that prioritize RTP traffic over other network traffic. Jitter buffers in Asterisk help smooth out packet timing variations.

For recording, Asterisk captures the RTP streams from both legs of the call and mixes them into a single audio file stored on disk.

Why It Matters

RTP configuration issues are the most common cause of audio problems in VICIdial. One-way audio (where only one party can hear the other) is almost always a NAT traversal issue affecting RTP port connectivity. If a firewall blocks the RTP port range, calls will connect (SIP signaling works) but have no audio.

Ensuring your firewall allows UDP traffic on the full RTP port range, configuring proper NAT settings in SIP/PJSIP, and placing VICIdial servers on public IPs when possible are the key steps to reliable RTP media delivery. See the VICIdial setup guide for network configuration best practices.