WINK Wall is a mesh video wall system. Each display is driven by its own single-board computer running an image we supply, so a wall grows one screen at a time instead of being sized around a single controller.
Feeds come from WINK Media Router, with WINK Forge handling transcoding where a camera's native stream isn't suited to direct display.
Get Started Read the Technical Brief
A video wall built from a mesh of inexpensive nodes rather than a single wall controller. Every display gets its own small computer, a Raspberry Pi or a Pi-class clone, mounted behind the screen and connected over the network.
You pick the node specification based on what the screen needs to show. Standard 1080p and 2K displays run on lower-cost hardware; 4K displays use a higher-specification board capable of decoding and rendering at that resolution. Both run the same WINK Wall software.
For each specification we provide a matching ISO image. Write it to the node's memory card, put the card in the board, and power it on. The node configures itself: it comes up on the network, joins the wall mesh, finds the controller, and picks up the layout and stream assignments that belong to that screen position. There is no per-node setup to perform by hand and no operating system to build.
Because the nodes form a mesh rather than hanging off one machine, the wall has no single point of display failure. A dead node affects one screen, and replacing it is a matter of swapping in a card with the same image, because the configuration lives centrally rather than on the device.
Pick the hardware tier per screen. A 2K-class board drives 1080p and 2K displays; a 4K-class board drives 3840x2160 panels. Mixed walls are fine, since the tier is a per-screen decision.
We supply the image matched to that tier. Write it to a memory card with any standard imaging tool. Nothing is compiled, patched, or hand-configured on site.
On first power-up the node brings up the network, announces itself to the mesh, and registers with the controller. Assign it a screen position and it takes the configuration for that position.
The node pulls its assigned feeds directly from WINK Media Router and renders them to the panel. Layout changes are pushed centrally and take effect without touching the hardware.
| 2K-Class Node | |
|---|---|
| Target displays | 1080p and 2K panels |
| Hardware | Raspberry Pi or Pi-class single-board computer |
| Image | WINK-supplied ISO, written to memory card |
| Typical use | Operator positions, secondary walls, dense multi-camera layouts |
| 4K-Class Node | |
|---|---|
| Target displays | 3840x2160 panels |
| Hardware | Higher-specification Pi-class board with 4K decode |
| Image | Separate WINK-supplied ISO for the 4K tier |
| Typical use | Primary wall panels, single-scene focus displays, large-format screens |
| Display Support | |
|---|---|
| Resolution | Up to 4K (3840x2160) |
| Interface | HDMI 2.0 |
| Refresh Rate | 60Hz |
| Color Depth | 10-bit |
| Network | |
|---|---|
| PoE Standard | 802.3at (PoE+) |
| Network Speed | 1 Gigabit |
| Protocol | RTSP/RTP |
| Configuration | QR Code / Mobile App |
The conventional video wall is organized around a controller: one appliance with a stack of output cards driving every panel over long display runs. It is expensive, it is a single point of failure for the whole room, and its capacity is fixed on the day it is bought. Adding a tenth screen to a nine output controller is not an incremental purchase.
Two things made it practical. Cheap ARM boards acquired hardware decoders good enough for the resolutions operations centers actually use, and a camera stream at 1080p and a few megabits is not a demanding decode target. More importantly the streams arrive from Media Router already normalized, so a node is not negotiating with forty camera firmwares. It is pulling one consistent feed.
On a controller based wall a controller fault is a dark room. On a mesh, a node fault is one dark screen out of twelve, and the operator can move that screen's content to a neighbor from the controller while a replacement is fetched. Because configuration belongs to the screen position rather than the hardware, the replacement is a spare card with the same image.
A node fails while the room is running an incident. The wall degrades to eleven screens instead of going dark, the affected content is reassigned in seconds, and a spare card from the shelf restores the twelfth without anyone touching a layout. The cabling story is the same discipline: one Ethernet drop per screen carrying data and power, no long display runs and no extenders to fail.
WINK Wall reaches End of Life on 31 October 2026, with security patches to 31 December 2031, and spare nodes can be re-imaged from the same ISOs throughout. For new build-outs the visualization role has moved to WINK Crossroad. Contact us and we will map an existing layout onto the replacement path.
Talk to us about existing WINK Wall deployments, spare node imaging, or migrating the visualization layer to WINK Crossroad.
Request Information