Pick the playback hardware by what fails and what you ship. A screen with the player built in is one box on the wall, so a fault means shipping the panel; an external player swaps in minutes from a shelf spare and suits multi-screen fleets; a mini PC with 8 GB of RAM and 128 GB of storage earns its keep when the content is your own software.
A signage screen needs a panel, something that plays the content, and a way to update it. The player can live inside the display, in a small box beside it, or inside a mini PC, and the choice decides what happens on the morning a unit dies: a box swapped from a shelf in about 15 minutes, or a crated panel shipped from a warehouse. The rule up front: one or two screens in one building with rarely-changing content go one-box; fleets that update weekly go on external players with spares on the shelf; screens whose content is software get a mini PC. Panel size is a separate question; the screen size and viewing distance guide covers that math. This article is about where the player should live, and what the choice costs on failure day.
What each option is
Built-in player screen (SoC). The display and the player are one unit: one mount, one power lead, one device to install. On our panel-mount signage displays the player is Android built into the screen, content lives on a microSD card or a USB stick, and the software comes with no subscription, so an update is a walk to the screen with fresh media. Larger commercial SoC platforms update over the network instead. Two limits to know before you plan: the HDMI port on our screens is an output only, so an external box cannot drive them, and the built-in player is designed to come back to its loop after a power cut, which matters on walls nobody stands beside.
Screen plus external player. The panel hangs on the wall and shows; a small dedicated box plays the content and feeds it over HDMI. The player becomes a replaceable part: spares sit on a shelf, one box model can serve many different panels, and a fault at one screen leaves the panel and the mount untouched.
Screen plus mini PC. A small general-purpose computer runs the content: a browser in kiosk mode pointed at a page, a dashboard client, or software you wrote yourself. You own the operating system, Windows or Linux, and everything that follows: patches, security policies, reboots, and a machine that sits at a desktop if the player software stops.
Side by side
| Built-in player (SoC) | Screen plus external player | Screen plus mini PC | |
|---|---|---|---|
| Boxes on the wall | One | Two | Two or more |
| A player fault means | The panel comes off the wall | A box swap, panel stays mounted | Rebuild or swap the PC |
| Replacement path | Crated display, freight, a scheduled visit | A shelf spare, minutes at the screen | A spare PC or a rebuilt image |
| Updates | At the screen by card or stick, or over the network by platform | Over the network from a player dashboard | Any deployment method you run |
| Content flexibility | What the player platform allows | What the player platform allows | Anything the OS runs |
| Security and patching | Firmware, handled by the platform | Firmware, handled by the platform | A general OS you own |
| Wins when | Content changes rarely and one box per wall matters | Content changes weekly and screens span sites | Content is a page, dashboard or custom build |
Read the replacement rows first. Everything else on this table is preference; those rows are downtime.
What decides the choice
How often the content changes. Count updates per screen per month. At one update a month or less, a walk with a card or stick is a chore most owners absorb. Weekly price and special changes turn that walk into a job, and network updates win. The example below puts the walk into hours so you can compare it against staff time.
What you ship on a failure. A built-in player makes the panel the spare part: crating, freight, a scheduled install, and a dark screen until all three happen. An external player makes the spare a box on a shelf. Large fleet operators report more failures from built-in players, and no published failure-rate comparison settles that either way, so decide on the replacement path, which you can control.
Who owns the device. A mini PC is a computer, so it belongs to the team that manages computers. Where that team exists, it can image, patch and monitor the box with the tools it already runs. Where policy blocks unmanaged general-purpose machines, the PC is dead at procurement and a purpose-built player or a one-box screen is what gets signed off. Retail rollouts often pick integrated players for a blunter reason: fewer boxes in the store kit, and nothing gets mislaid between the delivery and the wall.
Whether the player can leave the panel behind. External players decouple the two: match the player to the job screen by screen, hold one spare model that suits many panels, and upgrade players without touching a mount. Where one building serves different jobs on different walls, that freedom is the point.
Where each architecture wins
Built-in player screens win small, quiet rooms. A salon price board, a lobby display, a waiting-room notice: content that changes monthly or less, updates that can be a walk, and an install that stays one box on one power lead. The offline habit helps: the loop lives on the card, and if the content is a table of numbers you can regenerate the whole set without a CMS. The factory data table signage article shows one way to build that.
External players win fleets. Weekly menu boards, seasonal retail windows, notice sets across a chain: content pushed from one dashboard to every site, and a failed unit swapped in minutes by whoever holds the cabinet key. The network becomes a component with this choice: a player that cannot reach its content server keeps playing its cached loop or shows nothing. The article on why an Android signage player keeps losing CMS connection covers that failure; plan the cache for the sites that matter most.
Mini PCs win when the content is software. A live dashboard, a browser page, a build with custom input: content no player platform serves. Specify the box with a sealed case (no moving-air cooling to fail), 8 GB of RAM and 128 GB of storage as the working floor, and mount it behind the panel with room for air to move. Expect about five years of service from a build at that spec; that figure is field planning, so treat it as the age at which a spare becomes worth holding.
Example: 36 menu boards across 12 cafes
A cafe group fits three menu screens per shop across 12 shops: 36 screens. One office owns the content, and prices and specials change weekly. Two shops also want a counter screen showing the live order queue as a web page. The decision, with the arithmetic shown:
Built-in players, card updates. Every weekly change means 36 card swaps. Assume 5 minutes per screen: power down, swap the card, power up, confirm the loop. 36 screens × 5 minutes = 180 minutes, which is 3 hours of walking every week, spread over 12 shops. Run the same fleet at monthly updates and the walk is 3 hours a month, which most owners absorb. At weekly updates it is a recurring part-time job. Decision: card walks do not survive weekly changes at 36 screens.
External players. 36 players, 2 spares on the shelf, about 6 percent of the fleet. A failed player is a 15-minute swap at the screen; the panel stays mounted and the shop keeps trading. A failed built-in player on the same fleet means a replacement panel: assume 3 business days of transit plus a return visit, so that screen is dark for days. Decision: weekly content across sites goes on external players, with 2 spares held.
Mini PCs for the two queue screens. Each runs a browser in kiosk mode on a sealed-case unit with 8 GB of RAM and 128 GB of storage. Storage check: a 1080p loop encoded at 8 Mbps consumes 8 Mbit/s × 3,600 s ÷ 8 = 3.6 GB per hour, so 128 GB holds the OS plus weeks of cached loop video. Decision: the two screens whose content is a web page get mini PCs; the other 36 do not.
End state: 36 menu screens on external players with 2 shelf spares, 2 counter screens on mini PCs, and the card walk kept for whatever genuinely changes monthly.
The decision
Recommendation: choose a built-in player screen when one site holds fewer than about 6 screens and the content changes monthly or less; move to external players with one spare per site once a fleet passes about 20 screens, spans more than one site, or updates weekly; add a mini PC, sealed case with 8 GB of RAM and 128 GB of storage, only for the screens whose content is a browser page, a dashboard or your own software.
Four questions settle the rest at order time:
- Who updates the content, how often, and from where?
- What arrives after a failure, and how long is the screen dark in the meantime?
- Who patches the box, if the box runs a general operating system?
- Where does the player sit, and does that spot get moving air? A box buried behind a panel in still air cooks itself; leave it room.
Questions
How many screens can one player run?
Most dedicated players feed one screen each. Some mini PCs and some player models carry two video outputs and can run two screens from one box, which halves the boxes on a two-screen wall but doubles what a single fault takes down. Check the number of outputs and the resolution each output carries before you plan a two-screen wall on one box.
Do built-in players fail more often than external boxes?
Fleet operators report that they do, and no published failure-rate figures settle it. The decision does not wait on those figures; it turns on the replacement path. A failed external player is a small box from the shelf, a failed built-in player is a crated panel with freight and a scheduled visit. Standardizing on built-in players leaves spare panels as the only hedge, and panels take storage space.
Who should look after a mini PC signage screen?
The team that already patches your other machines. A mini PC runs a general operating system, so it needs OS updates, a reboot schedule outside opening hours and a line in the IT asset list. If nobody in the building owns that routine, the same box becomes the screen that shows a login prompt or an update banner in its first month.
Planning information, not legal or engineering advice. Rules differ by place: check codes and permits with your local authority and your licensed trades.