Why Standard Ethernet Falls Short for Professional Audio

Ethernet is the backbone of modern data communication, yet standard switched Ethernet operates on a best-effort delivery model that is fundamentally incompatible with professional audio requirements. When a switch receives packets from multiple sources, they are queued in a buffer and transmitted sequentially. This introduces variable latency—known as jitter—which is especially damaging to audio streams. A microphone capsule requires a constant, stable clock; recovering that clock from a jittery packet stream forces large buffers into the design, increasing overall latency to unacceptable levels for live sound reinforcement. Even 1 millisecond of latency can be noticeable to a performer, and cumulative delays across multiple hops quickly ruin the experience.

Before Time-Sensitive Networking (TSN), engineers had to rely on proprietary networks such as early versions of Dante, CobraNet, or EtherSound, or on point-to-point digital connections like MADI and ADAT. These solutions often lacked interoperability or scalability, locking users into single-vendor ecosystems. The professional audio industry needed a standardized way to guarantee bandwidth, bounded latency, and sample-accurate synchronization over a common IP infrastructure. TSN provides exactly that, and its adoption is now essential for any large-scale audio deployment.

Defining Time-Sensitive Networking: The Technical Toolset

Time-Sensitive Networking is not a single protocol but a comprehensive set of IEEE 802.1 sub-standards designed to make Ethernet networks deterministic. For an audio professional, understanding these specific components is key to specifying and deploying a reliable TSN network. The standards relevant to audio are mature and widely adopted in modern AV-over-IP systems, and they are continuously refined by the Avnu Alliance.

IEEE 802.1AS-2020: Generalized Precision Time Protocol (gPTP)

This is the foundation of all TSN audio networks. gPTP synchronizes the clocks of every device on the network to a master clock with sub-microsecond accuracy—typically within 500 nanoseconds. Unlike standard NTP (which operates over UDP and suffers from variable latency due to software processing), gPTP works at the hardware level. Switches become transparent clocks, correcting for their own internal processing delays as they forward synchronization messages. This ensures that a microphone on one side of a campus network and a loudspeaker on the other can share the exact same sample clock without needing a dedicated word clock cable. gPTP also supports redundancy: if the grandmaster fails, another device automatically takes over without losing sync.

IEEE 802.1Qbv: Time-Aware Shaper (TAS)

Qbv is the mechanism that guarantees low and bounded latency. It divides network traffic into different priority queues and uses a “gate control list” (GCL) synchronized to the gPTP clock. The switch opens a gate for critical audio traffic at a precise time, ensuring it is transmitted immediately without waiting for a bulky file transfer to finish. This creates “scheduled traffic” slots that isolate audio streams from best-effort data. The result is deterministic latency measured in microseconds, not milliseconds, regardless of background network load. For an audio engineer, this means that moving faders on a console and streaming 128 channels of 96 kHz audio happens with predictable, consistent timing. Qbv requires all switches in the path to be configured with the same schedule, which is typically done automatically by the audio protocol controller.

IEEE 802.1Qbu & 802.3br: Frame Preemption

In cases where a large, non-critical Ethernet frame (e.g., a video file backup) is currently being transmitted, frame preemption allows a critical audio frame to interrupt the transmission, go through the switch, and then let the large frame resume. This further minimizes the worst-case latency for high-priority audio streams, ensuring that even under heavy network utilization, audio traffic gets through without needing dedicated cabling. Preemption is optional but highly recommended for converged networks that carry both multimedia and best-effort traffic.

Stream Reservation Protocol (SRP) / 802.1Qat

SRP allows an audio source (the Talker) to advertise its stream requirements—such as bandwidth, latency, and destination—to the network. Switches along the path reserve the necessary resources (buffer space and bandwidth) and send confirmation back to the source. If a switch cannot fulfill the request—perhaps a link is oversubscribed—the stream is refused. This is a fundamental shift from “probabilistic” delivery to “guaranteed” delivery. An engineer can trust that once a stream is established, it will not be dropped or interrupted by other traffic. SRP works in conjunction with gPTP to ensure that resource reservations are time-aware.

The Evolution to IP-Based Audio Networking

The transition from point-to-point digital audio (MADI) to packet-switched audio began with proprietary solutions. CobraNet and EtherSound showed the potential of Ethernet but were limited in channel count and topology. Audinate's Dante protocol became the first mass-market success, offering high channel counts, plug-and-play discovery, and low latency over standard switches. However, Dante originally relied on careful IT configuration and a network that was not congested to achieve its performance. The need for a fully standardized, IEEE-defined solution led to the formation of the Avnu Alliance and the development of the Milan protocol. Milan is an assurance profile built on AVB/TSN. It mandates specific TSN standards (gPTP, Qbv, SRP) and specific audio payload formats (AAF) to guarantee interoperability between devices from different manufacturers. This allows a Yamaha console to talk to a Neumann microphone or an L-Acoustics amplifier over a single standardized network without proprietary lock-in. The broadcast world, meanwhile, adopted AES67 and SMPTE ST 2110, which also rely on TSN principles for timing and synchronization.

How TSN Powers Modern Audio Protocols

TSN is the underlying transport mechanic for several major audio networking protocols. Understanding how each protocol leverages TSN is critical for system design and troubleshooting.

Audio Video Bridging (AVB) / Milan

AVB was the initial set of IEEE standards that evolved into TSN. Milan is the pro-audio profile that ensures strict adherence to these standards. In an AVB/Milan network, all devices must support gPTP for clock synchronization. Streams are established using SRP, which reserves bandwidth across the network. The switches use Qbv (or the earlier credit-based shaper, Qav) to prioritize AVB traffic. Milan ensures that the entire chain from source to destination is deterministic. This is the gold standard for interoperability in live sound and installed AV. The Milan protocol page offers detailed checklists for device certification.

  • Strengths: Guaranteed deterministic performance, strict interoperability profile (Milan), Layer 2 (no IP configuration needed), sample-accurate synchronization.
  • Use Cases: Large-scale live sound festivals, high-end installed corporate AV, recording studios requiring reliable multi-track streaming.

SMPTE ST 2110 and AES67

While often mentioned together, these protocols serve different purposes. AES67 is an interoperability standard that allows different audio-over-IP systems (Dante, Q-Lan, RAVENNA) to exchange audio. SMPTE ST 2110 is a broadcast suite for transporting professional media (video, audio, ancillary data) over IP. Both rely heavily on TSN principles, specifically Precision Time Protocol (PTP). AES67 mandates the use of PTP (IEEE 1588 or a profile of 802.1AS) to synchronize devices. It streams audio as uncompressed PCM over RTP. It is designed to coexist with standard data traffic, but it does not mandate the full TSN stack (like Qbv or SRP). Instead, it leverages PTP timing to achieve synchronization, relying on network capacity and careful configuration to manage latency. SMPTE ST 2110-30 (Audio) and -31 (AES3 Transport) are the standards for professional broadcast. They require a fully managed network with proper PTP configuration. While a full TSN switch infrastructure (Qbv) is beneficial for minimizing jitter in large ST 2110 systems, it is often implemented on standard high-performance switches with VLAN segmentation.

  • Strengths: Open standard, interoperability across thousands of manufacturers, integration with video and data (ST 2110), widely adopted in broadcast.
  • Use Cases: Radio and TV broadcast studios, master control rooms, remote production (REMI) trucks.

Dante and TSN Integration

Audinate, the creator of Dante, has been a leading voice in TSN adoption. Dante was originally designed to run on standard managed switches, but modern Dante devices now support a TSN profile (Dante AVIO, Dante Domain Manager). This allows a single network infrastructure to carry both traditional Dante streams and Milan/AVB streams simultaneously. This convergence is the ultimate goal: a unified, standards-based network where any device can communicate using the same timing and QoS mechanisms. Dante's massive installed base means that many existing systems can gradually migrate to TSN without a complete overhaul.

  • Strengths: Huge existing install base, mature software (Dante Controller), compatibility with standard and TSN networks, low cost of entry.
  • Use Cases: Corporate AV, fixed installations, transportation, live sound mixing, audio routing in post-production.

Critical Applications Driving TSN Adoption

The move to TSN is not an academic exercise; it is driven by real-world requirements in demanding environments where reliability and low latency are non-negotiable.

Live Sound Reinforcement

Large concerts and festivals require massive channel counts (128 to 512 inputs) over distances of hundreds of meters. TSN's deterministic nature eliminates the risk of network congestion causing dropouts during a performance. The ability to send audio, video (for IMAG), and control data over a single fiber backbone using TSN greatly simplifies cabling and reduces setup time. Monitoring solutions benefit immensely from sub-millisecond latency, allowing in-ear monitor mixes to be perfectly tight and phase-coherent. Staging boxes, console racks, and amplifiers can be spread across the venue and still maintain sample-accurate sync.

Broadcast and Production (SMPTE ST 2110)

The broadcast industry is aggressively migrating from SDI to IP. SMPTE ST 2110 relies on TSN timing (PTP) to ensure that video frames, audio channels, and ancillary data (subtitles, metadata) arrive at the production switcher in perfect sync. Without TSN, jitter in a packet-switched network would cause audio to drift out of sync with video, or cause video frames to be dropped. TSN allows broadcasters to build flexible, software-defined studios where resources can be allocated dynamically—for example, adding an extra microphone channel without rewiring. Remote production (REMI) trucks also depend on TSN to transmit high-channel-count audio over wide-area links with predictable delay.

Installed Sound and Transportation

In corporate boardrooms, airports, and stadiums, audio systems operate 24/7. Reliability is paramount. TSN provides predictable performance over dedicated infrastructure. Automotive audio is an emerging use case: modern luxury cars have dozens of microphones, speakers, and DSPs. The Automotive Audio Bus (A2B) handles in-cabin point-to-point audio, but TSN is used for the backbone network connecting the head unit, amplifiers, and telematics controller. TSN ensures that active noise cancellation and immersive audio formats (Dolby Atmos) perform correctly at any vehicle speed. Similar requirements apply to aircraft and trains.

Implementation Strategies for TSN Audio Networks

Key Principle: TSN is a Layer 2 technology requiring hardware support in network switches. Not all “managed switches” support TSN. Look for switches explicitly certified by the Avnu Alliance for Milan or AVB, such as those from Cisco, Netgear, or Extreme Networks.

When designing a TSN audio network, the physical and data link layers are critical. Use a dedicated switch fabric or VLAN for audio traffic. Ensure all switches in the audio path support 802.1AS and 802.1Qbv. Configuration is typically performed through a switch management interface or via the audio protocol’s controller software (e.g., Milan Manager or Audinate Dante Controller). It is also essential to disable energy-efficient Ethernet (EEE) on ports carrying audio, as it can introduce latency variations.

A common practice is to deploy a dual-redundant network (Primary and Secondary switches) to prevent a single point of failure. TSN’s 802.1CB standard also allows for seamless redundancy, where duplicate packets are sent over diverse paths, and the receiver discards duplicates. This provides immunity to link failures without audio interruption. For larger systems, consider using a spine-leaf topology to minimize hop count and latency.

Proper PTP configuration is non-negotiable. Designate a grandmaster clock (usually a switch or a dedicated PTP generator). Configure boundary clock and transparent clock functions on your switches. Ensure all audio devices use the same PTP profile (e.g., SMPTE ST 2110 profile for broadcast or AVB profile for live sound). Incorrect PTP configuration is the most common cause of synchronization issues—often showing up as intermittent pops or clicks. Use tools like Wireshark with PTP filters to verify sync accuracy.

TSN Configuration Best Practices

For those new to TSN, the following checklist can help ensure a successful deployment:

  • Select switches certified for AVB/Milan (e.g., Cisco 9200, Netgear M4300, Extreme VSP).
  • Enable gPTP (802.1AS) on all ports that carry audio.
  • Configure SRP (802.1Qat) if using Milan or AVB; allow talkers to reserve bandwidth.
  • Implement Qbv scheduling if your protocol requires it; otherwise, ensure sufficient bandwidth and low utilization.
  • Use a dedicated VLAN for audio traffic to isolate it from control and best-effort data.
  • Set up redundant paths (dual-homed devices or a redundant ring with 802.1CB).
  • Verify PTP grandmaster selection and clock class; monitor for drift.
  • Test with the audio controller software before final commissioning.

The Future: Unified Networks and Cloud Integration

The trajectory of TSN is toward complete convergence. The distinctions between “audio network,” “video network,” and “IT network” are disappearing. TSN provides the common language for them all. The Avnu Alliance continues to refine the Milan profile, expanding its capabilities for larger channel counts and emerging audio formats like immersive audio and object-based audio. Wireless TSN is on the horizon, promising deterministic performance for wireless microphones and in-ear monitors over Wi-Fi 6/6E, freeing performers from the constraints of RF coordination. Cloud-based audio processing (Audio over IP to the cloud) will rely on TSN to provide the predictable latency needed for real-time production workflows across wide-area networks. As the IP landscape evolves, engineers must master both audio engineering and network engineering. TSN is not a temporary trend—it is the foundational infrastructure upon which the future of professional audio will be built. For further reading, consider the IEEE 802.1 TSN Task Group page and the Audinate support site for Dante-specific TSN integration.