JPEG XS Integration with SMPTE ST 2110 in IP Broadcast

This article examines how the JPEG XS codec integrates into SMPTE ST 2110 broadcast environments to deliver visually lossless, ultra-low-latency video over IP networks. As broadcast facilities transition to higher resolutions like 4K and 8K, uncompressed streams strain network capacity. By utilizing the SMPTE ST 2110-22 standard, broadcasters can transport lightweight, compressed video payloads while maintaining precise essence synchronization and low latency throughout the production workflow.

The Role of SMPTE ST 2110-22

The SMPTE ST 2110 suite traditionally handles uncompressed video under the ST 2110-20 standard. While uncompressed transmission ensures pristine quality, high-bandwidth demands—such as 12 Gbps for uncompressed 4K UHD—often necessitate costly network infrastructure upgrades.

To solve this, SMPTE established the ST 2110-22 standard, which defines parameters for transporting compressed essence streams over professional IP networks. ST 2110-22 treats compressed video as an isolated essence stream, preserving the fundamental ST 2110 architecture where video, audio, and ancillary data travel as separate, synchronized network flows.

JPEG XS Characteristics in Live Workflows

Developed under ISO/IEC 21122, JPEG XS (eXtra Speed, eXtra Small) is an intra-frame mezzanine compression format designed specifically for live broadcast operations. Unlike high-latency distribution codecs such as HEVC or AVC, JPEG XS operates with minimal processing overhead:

• Ultra-Low Latency: Latency is typically between a few lines of video to less than a frame (sub-millisecond delay), eliminating sync issues during live switching.
• Visually Lossless Quality: Operates efficiently at compression ratios ranging from 2:1 up to 10:1, maintaining visual fidelity across multiple encode/decode cycles.
• Resource Efficiency: Requires minimal compute power, allowing it to run efficiently on low-cost FPGAs, GPUs, and standard software platforms.

Mapping JPEG XS via VSF TR-08

The technical integration between JPEG XS and ST 2110 is standardized through the Video Services Forum (VSF) Technical Recommendation TR-08. VSF TR-08 outlines the specific payload format for carrying JPEG XS streams over the Real-time Transport Protocol (RTP), which ST 2110-22 directly adopts.

In this architecture, raw raster images are encoded via JPEG XS into codestream packets. These packets are tagged with RTP timestamps synchronized to the network’s Precision Time Protocol (PTP / IEEE 1588) clock as dictated by SMPTE ST 2110-10. This ensures that even though the video essence is compressed, its temporal alignment with uncompressed audio (ST 2110-30/31) and ancillary data (ST 2110-40) remains intact.

Signaling and Network Management

Integration within an ST 2110 system requires standard control plane signaling. The Session Description Protocol (SDP) defines stream configurations using parameters detailed in RFC 9134 (RTP Payload Format for ISO/IEC 21122 JPEG XS).

When establishing a stream, the SDP file advertises:

• Payload type and codec identification (video/jxsv).
• Sampling structures (e.g., 4:2:2, 10-bit).
• Quantization profiles and image buffer sizes.
• Source resolution and frame rate.

Network control systems using NMOS (IS-04 for discovery and registration; IS-05 for connection management) interpret these SDP definitions directly, enabling automatic routing and device interoperability across multi-vendor broadcast facilities.

Operational Advantages

Integrating JPEG XS with SMPTE ST 2110-22 enables broadcasters to carry 4K UHD signals inside a standard 10 Gigabit Ethernet (GbE) pipeline, or scale up to multiple UHD streams over 25 GbE or 100 GbE connections. It provides an efficient bridge for remote production (REMI), inter-facility connections, and constrained in-facility production routers, offering high-density stream transport without sacrificing image quality or introducing live production delay.