Small player hardware for an offline image loop

Which player hardware runs an offline image loop?

For stills or one MP4 with no internet, a single-board computer with 512 MB of memory and a quad-core 1 GHz CPU is enough, provided the software is a video looper rather than a browser-based platform.

Two classes of software run on a small board. A video looper does one thing: it plays every video file on the card in order, then starts again. Nothing else is loaded, so the whole job sits comfortably on the small board. A browser-based signage platform is a different animal: it brings up a desktop, starts a browser engine and then renders pages, and that stack is what the smallest board cannot carry. The reference board for the numbers on this page is the smallest common class: four 64-bit cores at 1 GHz, the memory figure named above, a board about 65 x 30 mm. If you are new to the idea of driving a sign from a board this small, running a signage player on a single-board computer covers the basics.

How much memory does signage software need?

A video looper fits on the smallest board. A browser-based signage platform is a job for about 2 GB, four times what the smallest board carries.

The 2 GB figure is the usual floor for a 64-bit Linux desktop driving a browser, because rendering pages is the heaviest work a small board ever does. Two details make the gap worse on the small board. Part of its memory serves the display pipeline, so the memory the software sees is less than the label says. And the memory is fixed: on this class of board it is packaged with the processor, so there is no upgrade six months later. The failure is easy to recognise: the platform installer stalls partway, and after a reboot the screen comes up black. A looper on the same board skips all of that and plays.

What files should an offline loop be?

Export the slideshow as one H.264 MP4 and any board on this page plays it; a folder of JPG stills sized to the screen works too.

The MP4 route is the one that actually ran on the small board: slides exported from a presentation program as video (PowerPoint exports video directly), dropped onto the card, looping. Export slides to video for an offline signage player walks through that conversion. The stills route is lighter still: a folder of JPG images shown one after another, which a TV can do from its USB port with no player at all; play a folder of images on a TV as free signage covers that. Whichever route you take, most playback problems are the file itself, and signage image file troubleshooting for offline displays lists the usual ones. Our small panel-mount displays take JPG, PNG or WebP stills and MP4, M4V or WebM video from a microSD card or USB stick. H.264 MP4 is the one we recommend, because H.265 and VP9 depend on the unit.

One watch-out belongs here. Leave one identical image on the same pixels for months and some panel technologies keep a faint ghost of it, so keep the slides changing, even if the change is only a clock or a rotating offer.

Example: what does a 20-slide loop actually need?

A counter screen at 1920 x 1080 showing 20 slides for 9 seconds each needs one 180-second MP4, about 225 MB at a 10 Mbps encode, and an 8 GB card holds it with room for a dozen seasonal versions.

Here is the arithmetic. The loop length is fixed by the slides: 20 slides x 9 seconds = 180 seconds. File size depends on the encode rate; take 10 megabits per second, a mid-range figure for 1080p30 H.264:

• 10 Mbit/s x 180 s = 1800 Mbit
• 1800 Mbit / 8 = 225 MB

The rate is the input that moves the answer. At 6 Mbit/s the same loop is 135 MB; at 14 Mbit/s it is 315 MB. Either way it is small: a card of 8 GB, about 8000 MB, holds the 225 MB loop with about 97% of the card free, so the same card can carry several versions of the show. Memory is the other check, and the small board's documented figure is 512 MB. One 1080p decode plus a looper fits that, which is why this route ran on the board where the browser platform would not start. Updating means putting a new file on the card and swapping it in; no network, no account, no subscription.

Which board should you buy for a loop?

Buy by the software, not the board. A small looper board is the cheapest way to one offline loop, and the memory floor is the line where browser platforms start to work.

The smallest-board row is the least hardware a working loop can be built from. Every row below it costs more and pays off only when the content outgrows one file.

When should you step up to more memory?

Step up when the content stops being one file: web pages, a dashboard, several screens from one player, or different shows at different times of day all want the browser stack and a faster CPU.

Each trigger has the same root: it needs the browser stack, or more than one output. Pages and dashboards render in a browser, so they carry the browser memory floor with them. Driving several screens from one player adds decoding load on top, and one signage player driving multiple displays explains when that works. When the hardware question gets bigger than a loop, choosing 4K Android signage player hardware covers the stronger end of the market. And if the loop is going up as a shop-floor notice or status board instead of in a shop, the andon board guide maps out where a small screen fits and where it does not.

Scheduling is worth a sentence. On our player the timer covers a daily on time and off time; a timetable of different shows through the day is a job for a full platform, and that is a reason to start at the larger memory size.

What should you check before you buy?

Ask six things: the memory figure, the file formats, how updates arrive, whether the show starts at power-on, sound, and which way the screen hangs.

• Memory: the smallest figure buys a looper and nothing else; the larger one buys the browser world. Decide the software first, then read the memory figure off the spec sheet.
• Formats: H.264 MP4 by name, plus JPG and PNG for stills. Anything the seller cannot confirm, assume it needs testing on your own files.
• Updates: card or USB stick walk-up, or network. Whoever updates the loop should be able to do it without a keyboard and monitor.
• Start-up: the show must come up at power-on with no clicks. Our player does this by design: set up as the home app, it brings the show up full screen at power-on, with no launcher and no keyboard. If the app stops, it is designed to start again within seconds.
• Sound: decide it before you export. An MP4 with no audio track keeps the loop silent on any player and shaves a little off the file size.
• Orientation: a portrait mount needs rotation support in the player; when a signage player lacks portrait rotation support explains the check to make before you buy the mount.

So which player should you actually buy?

Recommendation: for one screen and one looping MP4 offline, buy the smallest board in the table with a video looper, and update by card; move to a board with about 2 GB only when pages or a platform enter the plan.

That is the decision in both directions: the smallest board covers a looper for as long as the show stays one file, and the larger memory covers everything that renders pages. If a separate board, power supply and screen is more wiring than the job deserves, our panel-mount displays are the one-box version of the same answer. Images and video play straight from a microSD card or USB stick with no subscription, and new shows come from the free Studio app on a Windows PC, which builds a slideshow from ready-made slides with your own words typed over them and writes it to the card, checking every file on the way out.

Questions