Advanced NDI Concepts
In traditional AV systems, connections between devices are straightforward, using physical cables like HDMI or SDI. However, in NDI’s IP-based ecosystem, video, audio, and metadata are transmitted over a network. This shift introduces complexities that require advanced knowledge to manage and optimize. Advanced users must understand not only the principles behind these workflows but also how to resolve issues and implement strategies that keep systems running efficiently.
For example, in a live production setup, where dozens of NDI-enabled cameras and devices are interconnected, network congestion can cause delays and interruptions. This ability to navigate the complexities of NDI distinguishes advanced users from beginners, empowering them to deliver seamless, professional results.
NDI Discovery and Registration
NDI discovery allows devices to identify and communicate with each other automatically. When an NDI-enabled transmitter comes online, it announces its presence using mDNS by broadcasting service advertisements on the local subnet. Receivers listen to these advertisements to build a registry of available sources.
This passive listening approach reduces the need for manual configuration and accelerates setup in environments where mDNS traffic flows freely. For instance, a production team might deploy a new NDI camera just minutes before going live, and the camera will immediately appear in the list of available sources on NDI receivers, assuming the network supports mDNS properly.
However, challenges arise in more complex network environments. Firewalls might block mDNS traffic (UDP port 5353), or the network may be segmented across multiple subnets that don't support multicast routing. In such cases, devices might not discover each other automatically.
To address these limitations, NDI also supports NDI Discovery Server, a centralized discovery method suitable for larger or segmented networks. Instead of mDNS, NDI devices are configured to communicate with a specified Discovery Server over unicast. Transmitters register themselves with the server, and receivers query it for the list of available sources. This approach ensures reliable operation across routed networks, data centers, or when tighter network control is required.
Advanced users should be familiar with both discovery methods. Verifying subnet configurations, enabling appropriate multicast settings, and using tools like NDI Access Manager to specify preferred discovery servers or manually add remote sources are all key to maintaining a robust and scalable NDI environment.
Deep Dive into NDI Protocols
NDI has evolved through several major versions, each introducing transport-layer improvements to meet real-time video production demands across various network topologies.
- NDI 1 – Single TCP
The initial implementation of NDI used a single TCP. While this approach guaranteed reliable delivery, it proved less scalable in environments with a large number of NDI devices. TCP's connection-oriented nature and retransmission mechanisms can lead to excessive network overhead, making it less suitable for high-density production workflows where hundreds of streams may coexist simultaneously.
- NDI 3 – UDP with FEC and Multicast
NDI 3 significantly evolved by adopting UDP for video and audio transport, alongside Forward Error Correction (FEC) to compensate for packet loss. This shift significantly reduced latency and improved real-time performance in typical LAN environments.
Importantly, multicast support was introduced, allowing a single NDI stream to be distributed efficiently to multiple receivers without duplicating traffic, which is ideal for large-scale or multi-viewer setups.
However, using UDP with FEC has trade-offs: it typically increases overall bandwidth usage compared to TCP due to the additional error correction data. Furthermore, FEC requires extra processing resources on the receiver side, as packets must be reconstructed from redundant information to recover lost data. This overhead can impact performance in scenarios with limited CPU capacity or constrained network links.
- NDI 4 – Multipath TCP
NDI 4 introduced Multipath TCP, allowing devices to leverage multiple network interfaces in parallel. This boosted bandwidth and provided redundancy, but its effectiveness could vary depending on the balance and configuration of connected interfaces (e.g., mixing 10GbE and 1GbE links).
- NDI 5 – Reliable UDP (RUDP)
NDI 5 adopted Reliable UDP (RUDP) as the standard transport mechanism, blending UDP's low-latency benefits with enhanced reliability. RUDP includes packet sequencing, retransmission, and congestion control, ensuring consistent media delivery even under challenging network conditions.
Working in Complex Environment
Deploying NDI in complex environments, such as broadcast facilities, large corporate venues, or multi-floor event locations, requires precise network design to ensure stability, scalability, and predictable performance.
A key foundation is IP address management. Using static IPs or DHCP reservations avoids source name conflicts and discovery issues, especially when dealing with dozens or hundreds of devices. Segmenting the network using VLANs in larger setups helps isolate NDI traffic from unrelated data.
Since NDI relies heavily on low-latency, high-throughput unicast transmission, network switches must be configured with proper buffering and flow control. While NDI can work on general-purpose networks, in professional environments, it’s critical to ensure:
• All devices are connected through non-blocking gigabit or 10GbE switches.
• The switch fabric is properly dimensioned to handle peak aggregate throughput
• There is minimal hop count and consistent latency between sources and receivers
Although NDI supports multicast starting from version 3, it is disabled by default. IGMP snooping and querier configuration are relevant when multicast is explicitly enabled, which is covered in a dedicated chapter.
Advanced users also extensively use tools like NDI Access Manager to manually configure routing between subnets or setup devices to register with an NDI Discovery Server, which becomes essential in routed or segmented environments where mDNS is not viable.
In these scenarios, proactive design decisions, like reserving bandwidth per stream, ensuring dedicated switching paths, and mapping device flows, are critical. This avoids hidden bottlenecks and allows for seamless operation even under pressure.
Ultimately, building a high-performance NDI deployment in a complex environment means treating the network as a part of the production chain, applying the same rigor used for video signal paths, and ensuring deterministic behavior across all components.
Utilizing NDI Tools
NDI Tools is a fundamental software suite for anyone working with NDI. It includes essential utilities that help users get started with NDI, explore its capabilities, and manage more complex infrastructures. While some tools serve as entry points for discovering and testing NDI sources, others are critical in maintaining and controlling large-scale, multi-device environments.
NDI Studio Monitor is a powerful application that enables real-time viewing of NDI streams on the local network. It's ideal for basic source verification, remote preview, and confidence monitoring.
NDI Access Manager is a key tool for configuring how a system interacts with other NDI devices. Available for Windows and macOS, users can modify the local NDI configuration file through a user-friendly interface. Its main functions include:
•Managing NDI groups: Users can define send and receive groups to separate devices and logically control source visibility. A device will only discover or be visible to other devices in the same group, making this feature essential for multi-team productions or installations with overlapping workflows.
•Manual source configuration: It allows manual connection to specific NDI transmitters located on different subnets or across networks where mDNS discovery is impossible, ensuring reliable access to remote sources.
•Multicast settings: Access Manager can enable or disable multicast transmission for each device, allowing multicast-aware networks to distribute streams more efficiently.
•Discovery Server setup: It allows configuring NDI Discovery Server as an alternative to mDNS. This is critical in routed or segmented networks where multicast traffic is restricted or disabled.
Access Manager is not available on Linux; all these settings must be applied manually by editing the configuration file, as documented on the NDI Docs & Guides website.
For advanced users, Access Manager is more than a utility, it is a central component for controlling NDI behaviour in structured, high-performance IP production networks.
The NDI Analysis Tool, which will be covered in detail later, complements this workflow by offering advanced bandwidth, stream integrity, and system performance diagnostics, which are crucial during setup, testing, and troubleshooting phases.
OSI Model Recap
While the OSI model is a theoretical framework, understanding its key layers helps advanced users troubleshoot and design stable NDI-based systems. NDI operates primarily at:
- Layer 2 – Data Link: Relevant when diagnosing switch-level issues, MAC address conflicts, or when dealing with VLAN segmentation.
- Layer 3 – Network: NDI relies on IP addressing for discovery and communication. Problems like duplicate IPs, missing routes, or incorrect subnet masks can prevent devices from seeing each other or streaming correctly.
- Layer 4 – Transport: NDI uses different protocols (TCP, UDP, RUDP) depending on the version and configuration. Understanding how these behave, especially concerning packet loss, retransmission, or congestion, is key in complex environments.
Practical example: If a source is not visible, an advanced user may first check the IP configuration (Layer 3), then verify VLAN or switch isolation (Layer 2), and finally inspect packet flow using tools like NDI Analysis or Wireshark to evaluate transport-level issues (Layer 4).
By mapping NDI behaviours onto the OSI model, professionals can approach problems methodically, saving time during critical workflows and ensuring consistent system performance.
Designing a network infrastructure for NDI
Designing a network infrastructure for NDI involves more than just connecting devices to a switch; it requires choosing the right hardware, configuring it attentively, and applying best practices developed through real-world experience.
Network switches used in professional NDI deployments must meet specific requirements:
• Non-blocking architecture with sufficient backplane bandwidth to support high-throughput, low-latency traffic.
• Gigabit, 10-Gigabit, or even 25/40-Gigabit ports, depending on the number, resolution, and expected concurrency of NDI streams.
• Support for flow control, VLAN configuration, and detailed traffic monitoring.
Because managed switches contain increasingly complex processing logic and firmware, it's essential to keep switch firmware up to date. Vendors often release updates that address performance issues, fix compatibility problems, and improve reliability, which is especially important in demanding environments like video production. Modern switches are passive conduits and active computing devices that play a key role in the network's behaviour.
A few configuration considerations specific to NDI include:
• Flow Control:
Enabling flow control on switch ports allows for better handling of sudden traffic bursts. It reduces the risk of buffer overflows and packet drops by letting overloaded devices pause traffic momentarily, which is particularly useful in full-duplex streaming environments.
• QoS (Quality of Service):
Since NDI does not interpret or use QoS tags, enabling QoS can introduce unnecessary complexity or mis prioritization. It's generally recommended that QoS be disabled for networks dedicated to NDI.
• Jumbo Frames: Although often promoted for high-performance networking, jumbo frames are not recommended in NDI workflows. Many devices (including embedded cameras and NICs) may not handle them consistently. Furthermore, if a jumbo frame is lost, the impact on video quality is more severe due to the larger amount of data per packet. Standard MTU values (typically 1500 bytes) offer better interoperability and fault tolerance.
• VLANs: Using VLANs to separate NDI traffic from unrelated data (e.g., office traffic, internet access) helps avoid unnecessary congestion and minimizes latency. This is especially beneficial in mixed-use networks or when multiple production teams share the same infrastructure.
Takeaways
Mastering NDI in professional environments requires more than knowing how to connect sources and viewers; advanced users must understand how discovery, protocols, and network infrastructure interact to form a stable and scalable system.
In this module, we explored the internal mechanics of NDI discovery (mDNS and Discovery Server), the evolution of its transport protocols (from TCP to RUDP), and the network configurations necessary for high-performance deployment.
We also examined the role of tools like NDI Access Manager and Studio Monitor and the importance of switch-level features such as flow control, firmware management, and appropriate bandwidth planning.
These foundations are critical for tackling real-world scenarios, from OB vans to multi-floor installations, and from tightly segmented VLAN topologies to multicast configurations.
In the next modules, we will explore additional advanced aspects of NDI infrastructure, such as multicast deployment, VLAN configuration and usage, and techniques for debugging packet loss in NDI networks.
Keep building on this foundation to unlock NDI's full potential in complex, mission-critical environments.
