Integrating AVB and TSN Protocols for Seamless Network Audio Delivery

Modern audio-visual (AV) installations face mounting challenges as they migrate from dedicated analog and digital point-to-point connections to converged Ethernet networks. Live sound reinforcement, broadcast studios, conference rooms, and industrial intercom systems all demand deterministic, low-latency audio transport alongside standard data traffic. Two complementary technologies have emerged to meet these requirements: Audio Video Bridging (AVB) and Time-Sensitive Networking (TSN). While AVB established the groundwork for streaming synchronized audio and video over standard Ethernet, TSN extends these capabilities into a broader, more robust framework for real-time communication across multiple industries. Integrating AVB and TSN protocols enables network architects to build scalable, highly reliable audio distribution systems that maintain sub-millisecond latency and sample-accurate synchronization even as the network grows.

This integration is not merely a matter of enabling both sets of features on a switch. It requires a deep understanding of how each protocol contributes to timing, traffic shaping, and stream reservation. When deployed correctly, the combination delivers a seamless audio experience that rivals traditional dedicated cabling while leveraging the flexibility and cost-efficiency of IP networking. This article explores the technical foundations of AVB and TSN, the compelling reasons for their integration, practical implementation strategies, and best practices for ensuring a robust deployment.

Understanding AVB and TSN Protocols

Audio Video Bridging (AVB)

Audio Video Bridging is a collection of IEEE standards (primarily 802.1BA, 802.1AS, 802.1Qat, and 802.1Qav) ratified in the early 2010s to address the specific needs of professional audio and video transport over Ethernet. The core premise is simple yet powerful: provide guaranteed bandwidth, bounded latency, and precise timing synchronization for time-sensitive streams, all while coexisting with best-effort traffic.

At the heart of AVB is the generalized Precision Time Protocol (gPTP), defined in IEEE 802.1AS. gPTP synchronizes all devices on an AVB network to a common time reference with nanosecond accuracy. This clock synchronization is essential for multi-channel audio playback where multiple speakers, amplifiers, or digital signal processors must stay sample-aligned. Without it, listeners would hear comb filtering, echo artifacts, or drift over time.

To manage bandwidth, AVB uses the Stream Reservation Protocol (SRP) (802.1Qat). Each talker (source) advertises a stream, specifying required bandwidth and latency class. Bridges (switches) along the path reserve resources, ensuring that total reserved traffic never exceeds the configured budget. This prevents congestion-induced jitter for audio streams. The Forward and Queuing for Time-Sensitive Streams (FQTSS) (802.1Qav) component then shapes the egress queues on switches, prioritizing AVB traffic over regular data. FQTSS uses credit-based shaper logic to allow bursts while maintaining average rate limits, guaranteeing bounded delivery times.

AVB initially targeted professional AV environments, automotive infotainment, and residential systems. Its major limitation is that the credit-based shaper (CBS) works best with isochronous traffic patterns like audio, but it may not provide the ultra-low latency required by some industrial or automotive control applications. Additionally, the protocol stack was designed with a maximum of 75% of bandwidth reserved for time-sensitive traffic, leaving headroom for best-effort data but capping the maximum deterministic throughput.

Time-Sensitive Networking (TSN)

Time-Sensitive Networking represents the evolution and generalization of AVB. Beginning around 2012, the IEEE 802.1 TSN Task Group took the core concepts of AVB and expanded them into a suite of standards suitable for a wider range of time-critical applications, including industrial automation, automotive autonomous driving, 5G fronthaul, and professional audio. TSN is not a single protocol but a toolkit of mechanisms that can be combined to meet specific latency, reliability, and synchronization requirements.

The most significant additions in TSN include:

  • Time-Aware Shaper (IEEE 802.1Qbv): Instead of credit-based shaping, Qbv uses time-division multiple access (TDMA) gates on egress ports. Each queue has a gate that opens and closes according to a predefined schedule. This allows time-critical traffic to be transmitted in precise windows, eliminating queuing delays entirely. For audio, Qbv can enforce that all audio packets of a single sample block leave the switch within the same guard interval, achieving deterministic latency as low as 30–50 microseconds per hop.
  • Frame Preemption (IEEE 802.1Qbu / 802.3br): Low-latency frames can preempt the transmission of a lower-priority frame. This reduces worst-case delay when large best-effort frames block small audio packets.
  • Per-Stream Filtering and Policing (PSFP) (IEEE 802.1Qci): Ingress policing ensures each stream does not exceed its reserved bandwidth or send at the wrong time, providing security and reliability.
  • Frame Replication and Elimination for Reliability (FRER) (IEEE 802.1CB): For mission-critical streams, duplicate frames can be sent over disjoint paths and eliminated at the receiver, guaranteeing zero packet loss even if a link fails.
  • Asynchronous Traffic Shaping (ATS) (IEEE 802.1Qcr): For networks where global time synchronization is not feasible, ATS provides bounded latency without a common clock.

TSN retains backward compatibility with AVB: gPTP remains the core synchronization protocol, and SRP can be used for stream reservation. However, TSN introduces more flexible resource allocation (e.g., per-stream shaping versus per-class) and supports much larger network topologies and tighter latency bounds. The AVnu Alliance has defined certification profiles that specify exactly which TSN features must be implemented for professional AV, automotive, and industrial domains.

Benefits of Integrating AVB and TSN

Merging AVB and TSN capabilities yields a network that is both backward-compatible with existing AVB equipment and forward-compatible with emerging industrial and automotive standards. The combined framework unlocks several distinct advantages for audio delivery.

  • Precise Sample Synchronization Across Large Networks: AVB’s gPTP provides the highly accurate clock distribution needed to keep all nodes sample-locked. When TSN’s time-aware shaping is layered on top, the jitter in packet delivery is reduced to near zero, so receivers can reconstruct audio with sub-microsecond inter-channel delay. This is essential for immersive audio systems with dozens of speakers or for distributed recording setups where multiple microphones must be phase-aligned.
  • Ultra-Low and Bounded End-to-End Latency: AVB’s credit-based shaper can achieve latencies around 250 microseconds per hop for audio streams. By replacing or augmenting CBS with TSN’s time-aware shaper, the per-hop latency can drop to less than 100 microseconds, with absolute bounded worst-case delay known at design time. This makes integrated AVB/TSN suitable for live sound reinforcement in large venues where signal chain delay must stay below 1 millisecond or for remote production where low latency is critical for talent monitoring.
  • Inherent Scalability Without Resource Exhaustion: AVB alone limits reserved bandwidth to 75% of link capacity. TSN allows for more granular reservation per stream and can operate at up to 100% reservation when using time-aware shaping with proper guard bands. Additionally, FRER and PSFP prevent a single misbehaving stream from affecting others, allowing thousands of audio streams to coexist on a single backbone.
  • Robustness to Network Congestion and Faults: The combination of AVB’s stream reservation with TSN’s frame preemption and per-stream policing ensures that even under heavy network load, audio packets are delivered within their deadline. FRER provides redundancy without complex reconfiguration protocols; a lost frame is simply replaced by a duplicate arriving via a different path. This level of resilience is vital for public address systems, broadcast interruption, and emergency communication networks.
  • Interoperability Across Domains: Many modern audio endpoints already support AVB, but they increasingly need to coexist with industrial controllers, cameras, or generic TSN data streams. An integrated network can speak both protocols: AVB talkers and listeners can reserve streams using SRP, while TSN-capable switches prioritize those streams using the appropriate shaping method. This allows a single Ethernet infrastructure to handle audio distribution, control traffic (e.g., AES70 or Dante control), and time-sensitive machine data from the same switch.

Implementation Considerations

Integrating AVB and TSN is not a plug-and-play activity. It requires careful selection of hardware, consistent configuration of bridging software, and rigorous testing. The following subsections outline the key technical components that must work together.

Hardware Requirements

All network switches and endpoint devices in the audio path must support the required IEEE standards. For a full AVB/TSN integration, the minimum switch capability should include:

  • IEEE 802.1AS – gPTP grandmaster clock selection and boundary clock functionality.
  • IEEE 802.1Qat – Stream Reservation Protocol (SRP) to register and maintain stream paths.
  • IEEE 802.1Qbv – Time-aware shaper with configurable gate schedules.
  • IEEE 802.1Qbu / 802.3br – Frame preemption (strongly recommended for sub-100 μs latency).
  • IEEE 802.1Qci – Per-stream filtering to enforce traffic contracts.

Endpoint devices (audio interfaces, amplifiers, microphones) need an AVB/TSN-compatible network interface controller (NIC) and a protocol stack that implements gPTP, SRP, and the selected traffic class shaping. Many commercial products from companies like Dante (Audinate), Milan (Avnu Alliance), and RAVENNA/AES67 already incorporate AVB or TSN features; verify the exact profile supported.

Software and Configuration

Network management plays a crucial role. Most TSN features are configured via YANG models and NETCONF/RESTCONF, though some vendors offer simplified web interfaces. Key configuration steps include:

  • Clock hierarchy: Designate a gPTP grandmaster (usually a dedicated switch or the primary audio controller). Ensure all bridges and endpoints are set as boundary clocks to segment the timing domain and reduce cumulative jitter.
  • Stream reservation policy: Decide whether to use AVB-style SRP (static or dynamic) or TSN’s centralized reservation model (802.1Qcc). For smaller installations, SRP is simpler; for large, multi-vendor networks, a centralized network controller may be preferable.
  • Schedule construction: Define the time-aware gate cycles on each egress port. The schedule must account for propagation delay, link speed, and guard band lengths. Several scheduling algorithms exist (e.g., cyclic executive, ILP-based optimization). For audio-only networks, a simple repeating cycle of 125 μs (matching 8 kHz audio frame rate) works well.
  • Traffic prioritization: Map audio streams to a high-priority queue that is opened during the time-sensitive window. Best-effort traffic uses lower queues whose gates open only when audio queues are idle.

Time Synchronization Using IEEE 802.1AS

Accurate time synchronization is the foundation of both AVB and TSN. gPTP uses peer-to-peer delay measurement between adjacent nodes to compute link delay, then propagates the grandmaster’s clock using sync and follow-up messages. In an integrated network, all devices must run the same profile; the standard profile (802.1AS-2020) is sufficient for most audio deployments. It is essential that the network does not include non-802.1AS bridges that can delay or corrupt timing messages. Using a boundary clock at each switch ensures timing accuracy does not degrade as the network scales.

Traffic Shaping and Scheduling

The most impactful difference between AVB and integrated AVB/TSN is in traffic shaping. When both TSN and AVB equipment coexist, switches must be able to classify streams into different queues. Audio streams from AVB-only talkers will likely be mapped to a credit-based shaped queue (CBS), while streams that support TSN can use the time-aware shaped queue. Hybrid switches can hold queues with different shapers simultaneously, but care must be taken that a CBS queue does not delay a time-aware queue beyond its scheduled window. The usual practice is to assign TSN audio to a higher-priority queue with preemption enabled, while CBS queues serve AVB-only legacy devices.

Addressing and Stream Reservation

AVB uses multicast MAC addresses derived from the stream ID for stream identification. TSN retains this concept but also supports source-specific multicast and VLAN-based tagging. During integration, ensure that SRP can propagate reservation requests across the entire network. If some switches do not support SRP, they will forward reservation messages as ordinary data but will not reserve bandwidth—potentially causing oversubscription. Thus, all bridges in the path must be fully SRP-capable or operate in a “static reservation” mode where bandwidth is pre-allocated.

Best Practices for Deployment

To ensure a smooth integration of AVB and TSN in a production audio network, follow these guidelines derived from real-world installations.

  • Plan the topology carefully: Use a tree or star structure to minimize hop count. Avoid rings and redundant paths unless using FRER; standard STP/RSTP can cause large delays during reconfiguration.
  • Perform a timing budget analysis: Calculate the worst-case latency for each audio stream based on hop count, link speed, schedule cycle length, and guard bands. Use this budget to determine maximum network size.
  • Use certified equipment: The AVnu Alliance provides certification programs for AVB and TSN devices. Choose switches and endpoints that carry the “Milan” (professional AV profile) or “AVnu-TSN” certification to ensure interoperability.
  • Test with realistic traffic loads: Simulate peak audio stream counts plus background data (file transfers, video streaming) while monitoring jitter, clock offset, and packet loss. Expand the test to include link failures and cable swaps.
  • Monitor and maintain: Deploy network monitoring tools that can track gPTP clock offset, stream reservation status, and queue depths. Regular firmware updates are critical as IEEE standards are still evolving (e.g., 802.1AS-2020, 802.1Qcz).
  • Consider future scalability: Even if initially deploying with AVB-only endpoints, install TSN-capable switches. This allows a seamless upgrade path when endpoints are replaced with TSN-native ones, without rewiring.

Future of AVB/TSN Integration

As the IEEE refines TSN standards and as more industries adopt deterministic Ethernet, the integration of AVB and TSN will become the default for professional audio. The professional AV market is moving toward the Milan protocol (based on TSN), which specifies a common feature set for interoperability. Meanwhile, the automotive and industrial sectors drive demand for lower-latency and higher-reliability network components, which in turn benefits broadcast and live event applications.

One notable development is the convergence of AES67 (audio over IP) with TSN. AES67 already operates over standard Ethernet, but when combined with TSN-optimized networks, it can deliver clock stability and latency that rival raw AVB. This opens the door for mixing AES67 streams with AVB streams on the same integrated backbone.

Another trend is the use of TSN in cloud and edge computing: remote music production and remote broadcast trucks can use TSN tunnels across wide area links (e.g., via TSN-aware SD-WAN). While early implementations exist, widespread adoption will depend on further standardization and increased bandwidth availability.

Integrating AVB and TSN is not just a technical upgrade—it represents a paradigm shift in how we think about audio networking. Instead of separate infrastructures for audio, video, and data, a single Ethernet fabric can carry all time-sensitive streams with deterministic performance. For organizations that require high-quality synchronized audio in a dynamic environment, investing in this integrated approach today prepares them for the increasingly demanding applications of tomorrow.