Video walls come in various shapes and sizes for many different uses. From Control Rooms to Collaboration Rooms and from Education to Digital Signage, video walls show information at a large scale to convey important data and help solve problems or, in the case of digital signage, influence people. Small video walls can be driven by a single device, but also cannot display much information. Larger video walls need to be driven by multiple graphics pipelines in some sort of scalable architecture. This article explores how sources and other content are ingested and displayed in scalable video wall systems. While the concepts are general to the field, specific examples will use the Hiperwall scalable architecture and its video wall software components. Because of Hiperwall’s focus on Command and Control as well as many forms of Collaboration, the examples will be taken from those domains rather than Digital Signage, though Hiperwall is often used for large-scale Digital Signage video walls.
In this post, I will:
- Define scalable video wall systems,
- Explain the advantages of scalable video wall systems, including their reliability,
- Explore the types of content and sources used in these systems,
- And describe how sources are delivered to the scalable video wall.
Scalable Video Wall Systems Defined
A scalable video wall system means that multiple input channels and graphics pipelines drive content on the video wall displays and that their number and capability grow as the system grows or the display size increases. Whether the large video wall display is many LCD panels arranged together or a direct-view LED system consisting of many tiles driven by several LED controllers, multiple graphics pipelines are required to deliver content to the screens. Many video wall systems consist of more than one video wall connected to the system and sometimes even satellite displays distributed throughout the facility. The ability to handle many (hundreds, at least) sources of information and display them simultaneously on video walls comprised of tens or even hundreds of displays is a critical requirement for large scale video wall systems, and only scalable systems can meet it.
In the figure below, a scalable video wall system is driven by PCs running Hiperwall video wall software. The figure shows the video wall displays driven by a pair of “HiperView” computers that decode and draw the content on the screens, in this case LCDs, but LED controllers are just as applicable. Some people call these “players,” but that vastly understates their capability, so I will stick with “display computers” or HiperView computers. Several “HiperSource” input computers capture many kinds of source content for display on the video wall. Much more information about this approach will be explored in a later section of this article. Finally, several control points, including fault tolerant HiperController PCs and the powerful, easy-to-use HiperOperator PCs provide customers the ability to manipulate content in real-time and control the video wall system. All these components, each running on commodity PCs, are tied together by a powerful commercial-quality Ethernet network switch. Because all content in a Hiperwall system is “just data,” whether it be video streams, screen captures, or high-resolution web content, the full bisection bandwidth of a properly configured network switch is sufficient to deliver hundreds of content and source items to hundreds of displays or LED controllers.
Advantages of Scalable Video Wall Systems
Cost-effectiveness is one of the primary advantages of scalable video wall systems, for several reasons:
- The system can be scaled to meet customer needs. This means the customer gets a system that does what they want without having to over-provision (and overpay for) a less-scalable video wall controller.
- Linear costs – because a scalable system is designed to scale up and down, costs grow in a fairly linear way as the customer explores their needs. Adding displays, computers to drive them, and licenses add small and predictable marginal costs, with no need to expand to a different chassis or a totally different design, as with many non-scalable systems. Even the Ethernet switch can grow to fit the system’s needs with stacking and other approaches.
- Scalable systems can expand as customer needs grow. If a customer needs to expand by adding another wall or satellite displays or new inputs, a scalable system can grow to accommodate the new requirements with only an incremental growth in cost.
- Upgrades – scalable systems are comprised of commodity PCs, so it is easy to replace and upgrade individual PCs without having to replace the entire system. A newer PC may add more capacity or greater reliability, thus providing a cost-effective piecemeal upgrade path. This means never getting left behind as technology improves.
- Reliability – by design, truly scalable video wall systems avoid bottlenecks that slow performance. This has the positive side effect that it also reduces single points of failure. Using multiple commodity components together can add fault tolerance and recover capability to help the system maintain reliability and content integrity.
Another significant advantage of scalable video wall systems over non-scalable ones is capability/capacity. By adding computing power and graphics processing as the system scales, this increases the capacity of the system to accept and display content. If a video wall display system is large enough to display hundreds of video streams or other content, then the processing power driving that display system needs to be up to the task. Scalable video wall designs like Hiperwall fit the bill because their capability grows as the system grows. Each new HiperView computer adds capacity to decode more streams and drive more pixels on the displays. Using HiperView Quantum, the playback of all that content is synchronized between all the computers and, more importantly, between the displays, so even the largest LED video wall will not show tearing between LED controller boundaries.
Content Types for Scalable Video Wall Systems
A previous Hiperwall article describes the many uses of control room video wall systems and the content types they show. I will not repeat all that information here, but in summary, we differentiate between two content types: stored and dynamic.
- Stored content is pre-rendered content, including images, text, and video files. These content items are imported into the system so they can be shown on the video wall.
- Dynamic content includes IP camera streams, whether direct or via a VMS like Milestone or Genetec, captured video content via HDMI capture cards for example, web pages, or display/application capture streams. This content is “live” and must be made available to the video wall system, ingested to make it compatible, and then displayed on the video wall.
Scalable video wall systems have exceptional advantages and some challenges with displaying content. The largest advantage is that each display computer (HiperView in the Hiperwall example) only needs to render the content that is visible on the pixel space (displays or LED controllers) it is driving. This is an extreme advantage over other video wall systems when many content items are present. Instead of rendering the tens or hundreds of video streams on the entire video wall, a single display computer may only need to render 4 or 8 or 16 and then display the result. This is a significant component of the scalability advantage. This also comes into play with extremely large imagery. Instead of having to decode and render every byte of a gigapixel image, a scalable display computer only needs to decode and render just what is visible in its pixel space. This allows Hiperwall systems to interactively show, move, and manipulate gigapixel images with ease.
A challenge scalable distributed video wall systems have is because content that is shared between the pixel space of two different display computers is rendered and drawn by both those computers. For many reasons ranging from how busy each computer is to out-of-sync vertical refresh timing, the appearance of tearing could occur at the boundary if one computer renders the content later than the other one. With LCD displays, this effect is typically not very visible, partially because the bezel separation makes it less obvious. On LED systems, it can be distracting, so Hiperwall’s HiperView Quantum technology uses software and hardware to synchronize content playback to prevent tearing.
Reliability of Scalable Video Wall Systems
Another recent Hiperwall article explores architectural approaches to take advantage of the distributed nature of Hiperwall to greatly enhance reliability of the video wall system. Summarizing, various levels of partitioning the system into multiple regions and then adding inexpensive redundancy greatly enhances system reliability and fault tolerance. Two HiperControllers means if one gets disconnected, the other takes over. Partitioning the video wall into segments and driving each with separate display computers (and LED controllers if an LED wall) means only a small portion of the video wall fails if something happens to that computer or its connection. Driving different inputs of the displays or LED controllers with hot standby computers means the video wall keeps running even in the face of a display computer failure. Even the network switch can be redundant. These and other techniques make reliability and content integrity of a Hiperwall system cost effective to engineer from the start.
Source Input and Display in Scalable Video Wall Systems
So far, we have seen how a scalable video wall system can be used to show lots of content, including live or dynamic content, but what is this dynamic content and how do we get it into the system? Dynamic content is not stored content, so it is being generated in real-time by something and needs to be shown with minimal delay to be relevant. Depending on the type of content, that delay can be measured in milliseconds or maybe a few seconds, but timeliness is a key constraint on dynamic content usage.
Source Types
to come from an external source, such content is often called a “source” which is the term used for the rest of this article.
If the video wall is in a security operations center (SOC), the most common source type is a surveillance camera feed. These feeds can come directly from IP cameras, or via a VMS system like Genetec or Milestone, or even a raw video feed digitized or converted by a video capture device. In the first two cases, the source likely arrives at any video wall system as a compressed stream such as H.264 and must be decompressed to use it. For the capture device case, unless the capture device does compression in hardware, the video will arrive as frames, requiring high memory throughput to manage and use.
Other types of control rooms often need the output from specialized applications like SCADA systems or other monitoring systems to be shown on the video wall. Since these applications run on a computer, a screen capture type of source can be used to capture the state of the application and show it on the video wall. In other cases, the output of the computer can be fed via HDMI or other video cable into a capture device for display on the video wall.
Many control rooms are seeing a rise in the use of web applications, including GIS systems, maps, social media and news feeds, and web-based dashboards that show the current status of a process or system, such as data center monitoring. Web sources are different because they have no predefined shape or frame rate. A social media feed may only update every few seconds or even minutes, while traffic maps may update even less frequently. Maps showing emergencies may need to update within a second of data coming in. Unlike video streams or HDMI captures, web sources may have very unusual aspect ratios, possibly being very tall or wide, or can be much higher resolution than any typical display. This means the video wall system must be able to display web sources at sizes and shapes that match the source material, or else the view will be distorted or unreadable.
While the topic here is “video” walls, audio often plays an important role. News and weather feeds, as well as camera streams, may have audio that can provide information in a crisis situation. Therefore, the video wall system needs mechanisms to play that audio, as required, and adjust the volume of each source or content item as well as the output volume of each display computer. Therefore, if a control room operator needs to hear the audio associated with a specific source on the wall, the operator can unmute the content and turn up the volume on the display computer so the room can hear the audio.
Source Ingestion for Display
Getting all these sources into a video wall system is an important component of the system design. Some of them, like IP camera feeds, are on the network, thus many of them can come into the system via a simple network cable. Others, like analog and digital video signals, require capture devices with ports to accept the signal and a connection to the rest of the video wall system so the video feed can be shown. Web page sources also arrive via the network but may require complicated rendering before they can be shown on the video wall. In all cases, some level of processing and data transfer is required to show live sources on a video wall. Designing a scalable approach to accept these sources is a significant part of specifying a scalable video wall system.
Hiperwall uses several different “HiperSource” applications running on commodity PCs to ingest or create source feeds to show on the video wall. Each of these applications connects to the HiperController to identify the available sources, including their resolution, encoding, and more, so they can be listed in the content list to be selected and displayed on the video wall. Each HiperSource application also provides periodic thumbnails for each of its connected sources, so system operators can preview what the source shows before putting it on the video wall. While this can be a computationally expensive operation, this preview feature is an essential component of Hiperwall’s powerful but friendly HiperController software UI. The latest HiperOperator content list can show multiple such source previews at once, all periodically updating, so the user knows exactly what they will see when the content is sent to the video wall.
HiperSource IP Streams is the application that handles incoming IP camera feeds (including ONVIF support), encoder box streams, and many other video streams. Each HiperSource IP Streams application can manage up to 75 streams on a single commodity computer, and multiple such IP Streams computers can work together, so a Hiperwall system supports hundreds of simultaneously available streams. And because of Hiperwall’s scalable design, those streams can all be shown at once, up to the limits of the hardware. This means even small Hiperwall systems can easily have hundreds of streams available and show tens or hundreds at once.
HiperSource Streamer and the new Streamer+ handle high resolution, high frame rate video captures and deliver them to the Hiperwall system. While Streamer and Streamer+ have different commodity hardware requirements, both can capture and send the screen of their PC at high frame rate to the video wall. Both can also use readily available capture cards to capture incoming video feeds to send to the system. They take advantage of hardware-based video compression to deliver high quality, low bandwidth source streams to the video wall. Multiple screen captures and video captures can be used simultaneously. Both also offer KVM (Keyboard, Video, Mouse) control of the Streamer PC when capturing the screen, so operators on the HiperControllers or HiperOperators can interact with the PC.
HiperSource Browser is a powerful Chromium-based web browser that delivers web content to the Hiperwall system at very high resolution (over 200 million pixels) for extremely high-quality maps, dashboards, social media feeds, PDFs, and more. The application supports Hiperwall’s patented Multi-Active browser tabs, so a single Browser PC can deliver several high quality web pages to the system simultaneously. HiperSource Browser also supports KVM control so HiperController and HiperOperator users can interact with the web pages as needed.
HiperSource Sender is a cross-platform screen capture application. It captures all or part of a computer’s desktop and delivers it to the video wall. It can send multiple portions of the desktop at once, to allow a user to only share data that is needed. It uses CPU-based compression and runs on Windows, Mac, and Linux, even in Virtual Machines, so is perfect for SCADA applications, small dashboards, and proprietary applications. Sender supports KVM control, so HiperController and HiperOperator users can control the Sender PC, if needed.
The HiperSource IP Streams and Streamer(+) applications take advantage of network features only provided within the LAN, so they must be on the Hiperwall subnet and connected to the Hiperwall switch. HiperSource Browser and Sender, however, can deliver their feeds via TLS-encrypted and authenticated connections, thus they can be outside the LAN and even outside the country. This remote source mechanism of Sender and Browser enables a powerful HiperSource sharing capability called HiperCast. HiperCast allows multiple HiperSource Sender and Browser sources to be shared with several Hiperwall video wall systems around the world. This facilitates data sharing and a common operating picture for organizations distributed across geographic locations.
Scalability and Design for Reliability
Having multiple source applications running on a few separate PCs seems natural for scalability, but it can also enhance system reliability as well. For example, if a system needs 50 IP camera streams, a single HiperSource IP Streams computer can handle sending all 50 streams to the video wall, but if someone accidentally kicks the plug on that PC, all 50 streams disconnect. Using 2 HiperSource IP Streams PCs adds only a small incremental hardware cost and no additional Source license cost, yet if one fails, only the streams connected to it go offline. With one of the VMS plugins that work with the most popular VMS systems, it is quick and easy to transition those disconnected streams to the other IP Streams computer while the failed one is repaired. Such transfers can also be done via the HiperInterface web services-style control interface.
Similarly, HiperSource Streamer and Streamer+ work with multi-port capture cards, and Streamer+ is regularly used to stream 4 or 8 video captures from a single PC, but if that PC fails, those 8 streams go offline. Splitting the capture capacity between several Streamer+ PCs adds a little to the hardware cost but provides easier failure recovery and lessens the impact if something does fail, again with no additional source license cost.
The Hiperwall software also has some built-in features to help with source reliability. First, if a source disconnects unexpectedly, that event is logged. Also, the locations where instances of that source were visible on the video wall are remembered, so if the source reconnects, it will be automatically shown in the same places again. Another important reliability feature is HiperFailSafe content. This content type is useful if there is an extremely important content item that must be shown even with failures present. As an example, if the weather report is critically important to the operation of a control room, a stream of a weather channel can be shown via HiperSource IP Streams. But what if the cable goes out or the capture device fails? Well, a different feed of the channel through a different provider can be shown via Streamer. What if that fails too? A web page showing the weather can be shown via HiperSource Browser. HiperFailSafe content lets an operator define which sources or other content items are related along with their priority, so the video streams have higher priority than the web page version of the weather report. Then, the Hiperwall system shows the highest priority item in that list that is available, so if one fails, it will automatically switch to the next one. Of course, multiple HiperFailSafe content items can be defined and used simultaneously. So together, hardware redundancy and software features can greatly enhance video wall system reliability.
Security and Integrity Approaches
Security and Integrity are important constraints in many mission critical environments, particularly because Hiperwall uses software applications to capture source feeds and deliver them to the video wall. It may be that security constraints prevent certain sensitive source computers from connecting to the Hiperwall LAN, or perhaps the application being captured is sensitive and integrity requirements prevent other software from running with it. The easiest way to solve those problems is to use a capture card or an encoder box that simply captures the HDMI output of that PC so it can be turned into a source for the Hiperwall video wall. This provides complete separation of the sensitive PC and the Hiperwall system. Since Hiperwall software does not store captured source streams, no sensitive data can be retained by the system when the video wall is not in operation. This means the integrity of the source PC is maintained, the security of the data on the PC is not impacted, and yet its output can be shown on the video wall. The Hiperwall team and partners have white papers and other guidance for system security and integrity.
Robust Support
The Hiperwall software and system architecture is designed to be easy to install and manage, but properly specifying a system takes practice. Our closest partners have exactly that expertise and are able to help design and specify a scalable Hiperwall system to meet the needs of just about any customer. The Hiperwall Technical Services team is readily available to apply their experience to scalable Hiperwall system designs as well. It is important to us that our customers know they do not have to do things alone – they will have support in designing and specifying the system, installing and configuring the software, and maintaining and upgrading in the future.
