Introduction to Digital Audio Protocols

In professional audio production, the choice of digital audio protocol is a foundational decision that affects signal routing, latency, scalability, and overall system reliability. Whether you are designing a live sound rig, a broadcast studio, or a large-scale installed sound system, understanding the strengths and tradeoffs of each protocol is critical. Among the most enduring standards is MADI (Multichannel Audio Digital Interface), a protocol that has served the industry for over three decades. However, networked audio protocols like Dante, AES67, and AVB have emerged, offering new capabilities. This article provides an in-depth comparison of MADI against these modern alternatives, helping you decide which technology fits your workflow.

What Is MADI?

MADI, standardized as AES10, was developed in the early 1990s by RME and Merging Technologies. It was created to solve a simple problem: how to move many channels of digital audio over a single cable without the complexity of early network systems. The protocol defines a point-to-point link that can transport up to 64 channels of 24-bit audio at sample rates of up to 96 kHz (and with double-wire mode, up to 192 kHz for 32 channels).

MADI uses either 75-ohm coaxial cable (BNC connectors) with a maximum cable length of about 100 meters, or multimode fiber optic cable (SC connectors) allowing runs up to 2,000 meters. The simplicity of a dedicated cable pair makes MADI extremely reliable in environments where network infrastructure is not available or desirable. It is also known for extremely low latency—often under one sample block—and minimal jitter, which is critical for high-end studio monitoring and live sound.

Evolution of Networked Audio Protocols

While MADI remains a workhorse, the industry has moved toward Ethernet-based protocols that offer more flexibility, scalability, and ease of configuration. The most prominent of these are Dante, AES67, and AVB (Audio Video Bridging). Each has its own architectural philosophy, and their differences from MADI are significant.

Dante

Developed by Audinate, Dante is the most widely adopted networked audio protocol in professional AV. It operates over standard Gigabit (or 10GbE) Ethernet and uses IP networking principles. A single Dante network can support hundreds of audio channels with automatic device discovery, simplified routing via software (Dante Controller), and seamless integration with existing IT infrastructure.

Unlike MADI’s fixed point-to-point path, Dante uses packet-switched networking. This flexibility comes with trade-offs: latency can be managed (typically 0.25 ms to 1 ms) but requires careful network design to avoid congestion and clock synchronization issues. Dante uses PTPv1 (Precision Time Protocol) for clocking, which works well on dedicated networks but can be problematic over shared infrastructure. For projects requiring multi-room or multi-building audio distribution, Dante is often the fastest and most cost-effective solution.

AES67

AES67 is not a standalone protocol but a standard for interoperability between IP-based audio networks. It was published in 2013 by the Audio Engineering Society to allow devices using different proprietary protocols—such as Dante, Ravenna, and Livewire—to exchange audio streams. AES67 defines common transport (RTP over UDP), clocking (PTPv2), and discovery (SAP) mechanisms.

When comparing AES67 to MADI, the key difference is that AES67 is a "higher-level" interoperability layer, whereas MADI is a complete physical and data-link specification. AES67 does not mandate any particular cable or connector—it runs over standard Ethernet. However, AES67 does not guarantee the same deterministic low latency as MADI; its latency is usually configurable but can be higher when bridging different protocol domains. Its main advantage is the ability to connect systems that would otherwise be siloed, making it invaluable in broadcast and post-production houses that use multiple brand ecosystems.

AVB (Audio Video Bridging)

AVB is a set of IEEE standards (802.1BA, 1722, 1733) that provides guaranteed quality of service (QoS) for time-sensitive audio and video over Ethernet. Unlike Dante, which uses best-effort IP networking, AVB relies on traffic shaping and credit-based shaping to reserve bandwidth and bound latency. This makes AVB very robust for large-scale installations where deterministic delivery is critical—for example, in concert venues or stadiums with hundreds of speaker zones.

MADI’s dedicated cable cannot match the network management capabilities of AVB. AVB switches can prioritize audio streams, and the protocol supports up to 256 channels at 48 kHz on a single 1 Gbps link. However, AVB requires compatible switches and endpoints, and its adoption is less widespread than Dante. For projects already standardized on AVB, the latency (typically 2 ms or less) is acceptable, but MADI still wins in scenarios requiring absolute minimal latency with no network overhead.

Technical Comparison: MADI vs. Networked Protocols

Channel Count and Sample Rates

MADI supports up to 64 channels at 48 kHz (or 32 at 96 kHz, 16 at 192 kHz in double-wire mode). Dante and AVB can theoretically support hundreds of channels—Dante can handle 512 channels on a 1 Gb link (256 in each direction), while AVB typically handles 256. AES67 is transport-agnostic and can support any channel count as long as the underlying network bandwidth suffices. For very high channel counts, networked protocols have a clear advantage.

Latency and Jitter

MADI’s latency is extremely low—typically 1 to 3 samples (around 20–60 microseconds at 48 kHz). This is due to its dedicated hardware interface and no buffering for network congestion. In contrast, Dante and AVB add buffering for packetization and network management. Minimum Dante latency is 0.25 ms (about 12 samples at 48 kHz), which is still excellent for most applications, but not as low as MADI. AVB latency can be as low as 0.5 ms but depends on the network path. For critical applications like live sound foldback or high-end studio recording, MADI’s advantage in latency and jitter remains unmatched.

Infrastructure and Cost

MADI requires dedicated cables—coaxial or fiber. A single MADI coaxial cable can be expensive over long runs, but the overall system is simple: no switches, no IP configuration, no network management. For a small fixed installation (e.g., a recording studio connecting a console to a recorder), MADI is often cheaper and more reliable than deploying a small Ethernet network.

Networked protocols require Ethernet switches, structured cabling, and sometimes dedicated VLANs or QoS settings. The cost of a managed switch becomes negligible in large systems, but for a point-to-point link, MADI wins. However, networked protocols offer huge flexibility—multiple devices can share the same cable plant, and changes are made in software rather than rewiring.

Reliability and Redundancy

MADI has no native redundancy: a single cable is a single point of failure. Some systems use a redundant MADI link with an automatic switch, but this requires custom solutions. Dante and AVB support redundant networks out of the box—Dante's Redundant mode uses a second network interface and separate cabling, and AVB can be configured with multiple paths. AES67 does not define redundancy but can leverage network-layer techniques.

For mission-critical live broadcast or large-scale events, the ability to have automatic failover is a significant advantage of networked protocols. However, many engineers trust MADI’s raw reliability—it either works or it doesn’t, with no hidden failure modes from misconfigured switches.

Use Cases and Selection Criteria

When to Choose MADI

  • High channel count over a single cable without network complexity (e.g., connecting a 64-channel console to a stagebox).
  • Ultra-low latency requirements such as in-ear monitor rings or live broadcast IFB (interruptible feedback).
  • Legacy equipment integration—many older digital consoles, recorders, and converters have native MADI ports.
  • Studio environments where jitter must be minimized for pristine audio quality.
  • Fixed installations that do not require frequent reconfiguration.

When to Choose Dante, AES67, or AVB

  • Large, scalable systems with many devices (e.g., a university campus with multiple buildings).
  • Flexibility for re-routing—networked audio allows software patching without physical cable changes.
  • Integration with IT infrastructure—one cable can carry audio, control, and other data.
  • Multi-protocol environments where AES67 bridges Dante and Ravenna devices.
  • AVB is especially suited for large AV installations where guaranteed QoS is mandated by specification.

The Role of Directus in Audio Workflows

While not directly a protocol, Directus is a headless CMS that can serve as a backend for managing metadata, documentation, and configuration of audio systems. For example, a broadcast facility might use Directus to maintain a database of MADI patch points, device settings, and cable schedules. This demonstrates how modern tools can complement traditional protocols like MADI.

Conclusion and Recommendations

MADI remains a critical tool for audio professionals who need unmatched reliability and low latency in point-to-point connections. It is not obsolete, and many top-tier studios and live sound companies continue to invest in MADI infrastructure. However, the shift toward networked audio is undeniable. Protocols like Dante, AES67, and AVB offer scalability, flexibility, and ease of use that MADI cannot match, especially in large systems or environments where multiple devices must communicate.

A pragmatic approach is to use MADI for core point-to-point links (e.g., console to stagebox or converter) and bridge to a networked protocol for distribution to other zones. Many modern devices include both MADI and Dante or AVB interfaces, allowing seamless integration. Ultimately, the best protocol is the one that aligns with the specific requirements of your project: channel count, latency, infrastructure, and workflow complexity. For more in-depth specifications, refer to the AES10 standard, the Audinate Dante overview, and the IEEE 802.1BA AVB standard.

By understanding the differences between MADI and its network-based successors, you can design audio systems that are both reliable and future-proof.