Simulcast Format

Description

In 3GPP multimedia telephony and conversational video services, simulcast refers to the simultaneous transmission of multiple encoded versions of the same media content, typically at different resolutions, bitrates, or quality levels. A simulcast format is the specific encoded representation used by one of these parallel streams. Each simulcast format is described in the Session Description Protocol (SDP) as a distinct media format, including the codec, profile, level, and any format-specific parameters such as resolution, frame rate, and bitrate. The SDP format line and its associated attributes fully characterize the encoding. In the Real-time Transport Protocol (RTP), the payload type (PT) field in the RTP header identifies which simulcast format a given packet belongs to, allowing the receiver to demultiplex and decode the correct stream.

The concept was introduced in 3GPP Release 13 as part of the evolution of the IP Multimedia Subsystem (IMS) multimedia telephony service (MTSI) and the related WebRTC interworking work. It is specified in 3GPP TS 26.114 (IMS Multimedia Telephony; Media handling and interaction) and TS 26.980 (Multimedia Telephony over IP Multimedia Subsystem (IMS); Media handling and interaction for WebRTC). Simulcast is particularly relevant for multiparty conferencing and WebRTC-based services, where a sender transmits multiple spatial or temporal layers to accommodate heterogeneous receivers with different capabilities and network conditions. Each simulcast format corresponds to one of these layers or versions.

The architecture relies on the Session Initiation Protocol (SIP) and SDP for negotiation. During session setup, the offerer proposes multiple simulcast formats, each with its own payload type and attributes. The answerer selects which formats it can receive and may indicate preferences. The sender then transmits the selected simulcast formats as separate RTP streams, each with a unique Synchronization Source (SSRC) and payload type. The receiver can switch between formats dynamically, for example, based on available bandwidth or display size, without renegotiating the session. This is a form of receiver-driven adaptation.

Key components include the SDP format description, the RTP payload type mapping, and the RTP streams themselves. The SDP attributes may include parameters like 'max-fr' (maximum frame rate), 'max-fs' (maximum frame size), 'max-br' (maximum bitrate), and 'rid' (RTP stream identifier) as defined in IETF RFC 8851 and RFC 8853. The payload type field in the RTP header is a 7-bit value that maps to a specific SDP format through the 'a=rtpmap' attribute. In simulcast, multiple payload types are used, one per format. The receiver uses the payload type to route packets to the appropriate decoder instance.

In the network, simulcast formats are transported over the user plane using RTP/UDP/IP. The IMS core, including the Proxy-CSCF, Serving-CSCF, and Interrogating-CSCF, handles SIP signaling but does not inspect or manipulate the media streams. Media resource functions (MRF) may be involved in conferencing to mix or switch streams. The access network (LTE, 5G NR) provides the bearer with appropriate QoS to support the multiple streams. The simulcast format concept is also used in 5G MBS (Multicast/Broadcast Services) and in WebRTC interworking with IMS.

Overall, the simulcast format is a fundamental building block for adaptive multi-stream media delivery in 3GPP services, enabling efficient use of network resources and improved user experience across diverse devices and conditions.

Purpose & Motivation

The simulcast format concept was created to address the need for adaptive media delivery in IMS multimedia telephony and WebRTC interworking. Before its introduction, a single media stream with a fixed encoding was used, which could not efficiently serve receivers with different capabilities or varying network conditions. If the sender used a high-quality encoding, low-end devices or congested links would suffer; if it used a low-quality encoding, high-end devices would receive suboptimal quality. Simulcast solves this by allowing the sender to transmit multiple encoded versions simultaneously, each tailored to a different quality level. The receiver can then select the most appropriate format based on its capabilities and current network conditions.

The simulcast format provides a standardized way to describe each of these versions in SDP, ensuring interoperability between different vendors and implementations. It also enables receiver-driven adaptation without the need for complex transcoding or renegotiation. This is particularly important in multiparty conferences where a central server (e.g., a conference bridge) may need to forward different streams to different participants. The format concept allows the server to identify and forward the appropriate stream to each participant.

Historically, previous approaches like scalable video coding (SVC) attempted to solve similar problems by encoding multiple layers into a single stream. However, SVC has limited support and complexity. Simulcast offers a simpler, more robust alternative by using independent streams, which can be encoded with standard codecs like H.264 or VP8. The 3GPP specification in Release 13 aligned with IETF efforts (e.g., RFC 8853 for simulcast) to ensure global interoperability. Since then, the concept has been maintained and refined in subsequent releases, supporting new codecs and use cases.