RTP (Real-time Transport Protocol)
RTP carries the actual audio of a VoIP call in small timestamped UDP packets. Learn what it does, how it relates to SIP and why NAT often breaks it.
Published 8 November 2026 · 1 min read · By SpaGreen Creative
Short answer
RTP is the protocol, defined in RFC 3550, that carries encoded audio and video in sequence-numbered, timestamped packets, usually over UDP, after SIP has set up the call.
SIP sets up a call; RTP carries the voice. Defined in RFC 3550, RTP packages each short chunk of encoded audio, typically 20 milliseconds, into a packet with a header that includes a sequence number, a timestamp, a payload type that identifies the codec, and a synchronization source identifier. The packets travel over UDP because a late audio packet is useless and retransmission would only add delay.
The receiver uses the sequence numbers to detect loss and reordering and the timestamps to play the audio out at a steady rate through a jitter buffer. A companion protocol, RTCP, reports statistics about the stream.
The IP addresses and ports for RTP are negotiated in the SDP body of the SIP messages. Each side tells the other where to send audio. This is why NAT causes so many problems: if a client advertises its private address, the far end sends audio to a place it cannot reach, and the result is one-way or no audio. STUN, ICE and PBX-side NAT handling exist to correct the advertised addresses.
Plain RTP is unencrypted. SRTP adds encryption and authentication on top of the same packet format.
Sources and further reading
- RFC 3550: RTP: A Transport Protocol for Real-Time Applicationswww.rfc-editor.org
This article is published by SpaGreen Creative, the maker of SIPdia. Product names mentioned belong to their owners. Where it describes other software, treat the steps as general guidance and check the official documentation linked above.
