Early Media

Description

Early media refers to media streams—most commonly audio like ringback tones, network announcements, or interactive voice response (IVR) prompts—that are transmitted to the calling party before the communication session is formally established (i.e., before the final 200 OK response in SIP signalling). This concept is defined and managed within the 3GPP IMS (IP Multimedia Subsystem) framework, specifically across the supplementary service specifications TS 24.428 (Communication Diversion), TS 24.528 (Communication Hold), and TS 24.628 (Common Basic Communication procedures). Early media allows the network or the called party to deliver meaningful audio to the caller while call setup is still in progress, enhancing user experience by avoiding dead silence during the ringing phase.

Architecturally, early media operates within the IMS session establishment flow. When a caller initiates a session via SIP INVITE, the network may begin sending media before the final session acceptance. This is signalled using SIP provisional responses (such as 183 Session Progress) that carry SDP (Session Description Protocol) information to negotiate early media parameters. The media itself is transported via RTP (Real-time Transport Protocol) over UDP, using the bearer established during the provisional phase. The IMS core elements—P-CSCF, S-CSCF, and the MRF (Media Resource Function)—coordinate to ensure that the early media path is properly set up and that QoS is maintained where possible.

The Media Resource Function (MRF), comprising the MRFC (Media Resource Function Controller) and MRFP (Media Resource Function Processor), plays a key role in generating and injecting early media such as announcements, tones, and prompts. For instance, when a call is diverted or when the called party is unreachable, the MRF may generate an announcement that is played to the caller as early media. The specifications TS 24.428, 24.528, and 24.628 define how these supplementary services interact with early media, ensuring consistent behaviour across different service scenarios including call forwarding, call hold, and multiparty sessions.

From a protocol perspective, early media involves coordination between SIP signalling and bearer-level resource reservation. The 183 Session Progress response is critical because it indicates that the callee's device or the network is ready to send media. The caller's UE must be prepared to receive and render this media. In some cases, multiple provisional responses may carry updated SDP to modify the early media session. The 3GPP specifications also address interactions with PSTN interworking, where early media from the PSTN side (e.g., a carrier's ringback tone) must be relayed to the IMS caller.

Early media is important because it enables a richer, more informative call experience. It supports customised ringback tones (CRBT), network announcements, and pre-answered media, all of which are widely deployed in commercial networks. The specifications ensure interoperability across vendors and operators, enabling consistent early media handling in multi-vendor IMS deployments. As IMS has evolved through Releases 7 to 19, the early media framework has remained stable, with the definition text carried over essentially unchanged, reflecting its maturity and foundational role in IMS service delivery.

Purpose & Motivation

Early media exists to solve the problem of silence during the call setup phase. In traditional circuit-switched telephony, callers hear ringback tones generated by the network while the call is being set up. In IP-based systems like IMS, there is no inherent mechanism to provide such audio unless explicitly designed. Early media was created to bridge this gap, allowing the network or the called party to send audio (or other media) to the caller before the session is fully established. This is essential for providing ringback tones, network announcements (e.g., 'The number you have dialled is not available'), and IVR interactions that begin before answer.

Without early media, IMS callers would experience an unnatural silence after dialling, which could lead to confusion or premature call abandonment. The introduction of early media in 3GPP Release 7, alongside the maturation of IMS, addressed this by defining how provisional SIP responses and SDP negotiation can be used to establish a one-way or two-way media path before the final answer. The specifications 24.428, 24.528, and 24.628 were created to ensure that supplementary services like Communication Diversion and Communication Hold could correctly interact with early media, preventing issues such as media clipping or incorrect tone generation.

The motivation also included interworking with legacy networks. When an IMS caller reaches a PSTN subscriber, the PSTN generates ringback tones; early media allows these tones to be relayed seamlessly to the IMS caller. Similarly, customised ringback tones offered by operators rely on early media to deliver subscriber-specific audio to callers. Thus, early media is not just a convenience but a critical enabler for service parity between IMS and legacy telephony, and for value-added services that operators monetise.