A light show for every crit: How good are the $39 Pixels “smart” dice?

This site is reader-supported. When you click through links on our site, we may be compensated.

Oh, Kickstarter: the land of wild, wacky promises and broken dreams, where products that could’ve been imagined during a productive shower or a psychedelic trip can become a reality, logistics and physics be damned. As we’ve written and seen, however, it’s a dangerous space for consumers, so much so that Kickstarter warns customers that it’s not technically a “store.” You give Kickstarter money, and it gives you the potential to receive goods or services.

Hence, we prefer to test a mid-Kickstarter product before telling you about it, and that’s the case for Pixels Dice, as seen in the above shiny-and-alluring images. Full of sensors, LEDs, and Bluetooth functionality, these dice sounded like the smartest addition to a tabletop game I’d ever seen when they contended for the 2019 Hackaday Prize. Upon getting my hopes up, I emailed their creator a cold-call request: whenever Pixels Dice actually exist, I want to test their sales pitch.

One very long year later, a package showed up at my door, and it contained two prototype, 20-sided Pixels Dice—currently priced at $39 per die, or $199 for a seven-dice set. Now that the project’s Kickstarter is live, and (as of press time) teetering towards $3 million in sales, I wanted to share my prototype testing experience, along with my somewhat optimistic take on what to expect from the final version, currently estimated to ship in “March 2022.”

Critical hit, now with critical light

As described on their Hackaday project site, Pixels take the board-gaming convention of multisided dice, then add six electronic components: a Bluetooth controller, an array of RBG LEDs, an accelerometer, a battery, a wireless, inductive-charging coil, and onboard memory.

Your imagination might immediately run wild with the sum total of those components, as squished inside gaming dice, and creator Jean Simonet is bullish about their gaming potential in his sales pitch. The obvious biggie is LED light-show possibilities with every roll of the dice, as paired with accurate roll tracking. Roll a 20 (a “crit” in D&D-speak), and your die could explode in a sensational light show. Roll a 1, on the other hand, and your die could light up with the visual equivalent of a sad trombone. Roll anything in between, and each face of the die can light up with its own colors and animations, as chosen by you.

Speaking of: should your dice be synced to a nearby Bluetooth device, your dice rolls could trigger sound effects via a compatible app. Maybe you’d prefer a literal “womp womp” sound, or maybe someone at your table would benefit from the dice-roll number being spoken out loud, or tracked in a D&D-style journal, by a companion app.

Having picked through my share of high-end dice bins at nerdy conventions, I don’t flinch at the idea of spending $39 on a single, blinged-out die. $199 for a full set is another matter, however. And in my testing of Pixels thus far, that’s where I currently draw the line. The prototypes I’ve played with include a mix of strengths and annoyances, tolerable enough for a single-die investment, or maybe even a pair. But I hesitate to dump an entire set’s worth of confidence into a $199 Kickstarter preorder.

Not bad at first LED blush

All of my tests were conducted using Pixels’ nonfinal prototype hardware, which only came in D20 flavors; they’ll eventually come in other popular polyhedral flavors (6-sided, 10-sided, etc.). Anything I describe below could be improved by at least one more year of development, iteration, and testing. Anything could turn out worse in the final product, as well, once the line moves from handcrafted, one-of-a-kind prototypes to products manufactured at scale. For the rest of this article, I will call these prototypes Pixels.

When I unboxed and began rolling Pixels, I skipped syncing to any Bluetooth devices to see how the dice had been set up by Simonet (he personally packaged and shipped these suckers). I learned that each die had its own light-animation template saved onto its memory, and both revolved around a simple ruleset: one general light-show animation for numbers 2-19; a “sad” animation for 1, and a “celebratory” animation for 20. It always recognized a 20 or a 1 precisely; exactly how it measured the other numbers, I couldn’t determine with this template.

Boom: basic sales pitch achieved. If I’d bought these at a store with zero customization options, I’d think that was a fine starting point in terms of unique, high-tech dice. Still, I came to realize the preinstalled animations were not quite up to my tastes. In particular, when Pixels’ light-show animation fills every die’s face, it can be hard to quickly see which number is showing on the top—and you don’t want to be the person at your table making everyone strain their eyes for 2-4 seconds of flashy animations to figure out what you just rolled.

As a result, I’m already keen on recommending Pixels’ opaque-body models (which I’ve tested) over the transparent ones (which I haven’t). These LEDs run pretty bright, and having those lights emerge through cut-out numbers is crucial for readability as it is. I can’t imagine trying to parse a Pixel die roll’s results with more transparent plastic absorbing and showing more obfuscating light.

Case and charging

The Pixels prerelease testing kit, as sent to Ars Technica earlier this year. The white pods hold one D20 die each, and they must be plugged in to charge.
Enlarge / The Pixels prerelease testing kit, as sent to Ars Technica earlier this year. The white pods hold one D20 die each, and they must be plugged in to charge.
Sam Machkovech

Each die comes in its own tiny charging case, essentially a plastic shell to clamp the die into so that its 1 face rests against a charging coil. (The final product will also include larger charging cases that support multiple dice.) Currently, there’s no special animation to indicate that a die has made contact with the wireless coil, but it’s easy to line them up: if the 20 is showing on top, then you’re good to go. Plus, shutting them into the case appears to stop the dice from exploding with dice-roll animations while bouncing around your favorite dice bag.

However, I’ve had these for long enough to leave them clamped in their cases and stowed for weeks at a time, which led to their tiny batteries slowly draining when not in use. Thus, unless Pixels receive a serious redesign in terms of how they identify on-off states, plan to charge yours for at least 30 minutes before any gaming session, lest you wind up in the nightmare scenario of not blowing your friends away with customizable rainbow-gasms whenever your level-7 dwarf ranger shoots an arrow directly into a troll’s eye socket.

Pixels’ charging cases currently include a micro-USB connector, which I eventually came to learn was only for the sake of charging. Since they connect directly to the charging coil, as opposed to a data-transfer interface, Simonet has confirmed that they will not double as a way to connect them to a computer for the sake of updates.

I sure wish they did.

An absolutely alpha app (for now)

The preproduction Pixels come with an app written in Unity and designed for iOS. As a lowly Android user, I asked whether I was out of luck during my testing, at which point Simonet sent me the pre-alpha app’s Windows port, instead.

Sadly, this unoptimized app was the highest-level headache I faced in coming to terms with Pixels’ potential. I’m not here to review this app’s usability, and Simonet was, quite frankly, brave to put this pre-alpha app in my hands as a “good enough for a dev” mess. Simonet’s most notable work experience comes from a wealth of programming work for Bethesda games like Fallout 3 and Elder Scrolls V: Skyrim—but that’s programming, not UI and interface design. At this point, I’m hopeful he funnels some of Pixels’ Kickstarter cash into help on the app-development front.

The bigger issue came from the prototype Bluetooth controller’s flimsy ability to maintain a consistent connection with my PC. Syncing my two Pixels dice to my Windows 10 PC was always a nightmare, and I’m regularly able to connect Bluetooth sources to the same machine with no trouble, whether as portable audio speakers or as wireless gamepads. When pressed on this issue via email, Simonet didn’t mince words: “Remember that there is at least another year of development on the app and firmware side. The only reason it exists as it is right now is because I am not able to demo the dice in person. That being said, I understand your reservations.

“I know you can only take my word for it, but I would never dare ship the dice in this state,” he added.

Fair enough. Until then, however, I cannot test its sync-via-Bluetooth potential, particularly the idea that it would consistently read a number out loud for anyone who’d benefit from such a use case. If that feature doesn’t work 99.99 percent of the time, that use case is tough to recommend.

Template tinkering

With app-related caveats out of the way: Pixels offer a compelling path to light-show customization in the form of PNG artistry. Take a look at the above, zoomed-in array of pixels, and you’ll see exactly 20 rows. Each row is dedicated to one face of a 20-sided die, and Pixels read each face’s row left-to-right as an animation pattern, fed to each LED.

The Pixels app comes with a ton of premade PNGs, all formatted by Simonet to work within LED limitations. For example, certain colors, particularly shades of brown, don’t accurately translate from the traditional RGB spectrum to LED bulbs, so Pixels are hamstrung for some colors by default. Thus, if you’re dreaming of a full slate of customized Pixels dice, each one color-matching your regular group’s “minis,” you’ll have to do some creative formatting of your PNGs.

The thing is, you can upload whatever PNGs you want to your dice, with some caveats. First, they need to fit the above template, mapped to the LED controller’s limitations, or else those blips will result in a blank signal. Second, and more importantly, they have to be teensy-tiny saved files. Prototype Pixels dice each include 12KB of onboard storage, and Simonet says he expects the final models to top out at 16KB, or “not much more than that.”

My Pixels dream is to have one custom animation template play for every roll result, and Simonet says he expects “at least 20 [animation] patterns” will fit on each final shipping die. Trying to split the difference in the meantime and have a series of six brief animations mapped to ranges of roll results has been a struggle, owing to the current, unrefined app. I’ve tried to map certain PNGs to number ranges, like a cold blue animation for 2-5, green for 6-10, and so on, but the current app’s customization menus include assignments like “roll is greater than,” “roll is less than,” “roll is equal to”—and even when I reorder every assignment in every way imaginable, certain ruleset assignments always seem to override others. My back-and-forth with Simonet on this issue never resulted in fully working results; Simonet boiled this down to “incomplete” hardware.

Optimistic about PNGs

But when it works, granular PNG control on Pixels is pretty sweet. In particular, I like being able to map dim or entirely black portions on certain die faces. Having seen how Pixels look in action, my preference is to have the top-most face remain brightly lit. After a brief just-landed animation across the die’s full body, I’d like for the closest surrounding faces to drop 75-100 percent in luminance, while the lower faces continue animating. This way, the number is clearly readable, while its relative value (blue for lower, orange for higher) is perceptible wherever friends sit at a gaming table.

Simonet’s current pack of default templates isn’t very imaginative, though I’d imagine the community’s involvement with PNG generation could change that dramatically. I’m hopeful that by the time these ship, the way Pixels reads these PNGs is a bit clearer, as well. You can currently tell the app whether to read an uploaded color pattern in absolute or relative fashion; in the former case, the number 20 will always light up red, while in the latter, whichever face lands on top turns red. The problem here is that it’s unclear exactly how the PNG shakes out for the rest of the die’s faces, which makes mapping an exact animation a trial-and-error process. (For me, that’s stymied in the prototype phase by wonky Bluetooth connections, making it laborious to upload and test new experiments.)

I’m hopeful Pixels’ retail version will include a PNG template tool to help obsessive nerds (as in, the kinds who’d pay $40 per die) lock down animations on a number-by-number basis. (Ordering the PNG’s rows in numerical order doesn’t make logical sense, for example, since a D20’s numbers are distributed in alternating fashion.)

The pre-alpha app hints to some flexibility for animation patterns, at least, especially if you would rather pick a single color, a gradient, or other function-based color waves instead of fiddling with the granularity of hand-coded PNGs. Those options may prove particularly useful if you want to work within the confines of Pixels’ tiny storage.

Define “very tough”

In terms of build quality, I have two opaque dice in hand, and these appear to be the Kickstarter campaign’s “onyx black” and “hematite gray” models. Both are made from resin, and both feel fine enough in terms of heft at roughly 8 grams in weight, similar to other plastic and resin dice options. When I cup a Pixels die in my hand and roll it around, then do the same with one of my old Chessex, I perceive a slight difference but nothing like the difference between plastic and aluminum dice. The only exception comes when I take a Pixels die fresh out of its charger and feel its warmth, as the resin definitely holds fresh charging heat for a few minutes.

I’m not sure if that’s why one of my two Pixels dice has a slight chip after two months of on-and-off testing, but sure enough, the 20 face on my black model is currently damaged, albeit slightly. I only noticed this while putting final touches on this article, so I’m pretty sure the chipping happened during my testing at some point, even though I’d never dropped these from a great height.

Simonet’s Kickstarter campaign pledges the following:

Pixels are really strong and don’t expect you to be careful handling them; all the electronics are encased in very tough resin. Your Pixels will last for many years!

But a slight chip in plastic dice, which you can discard in favor of a $1-4 replacement, is one thing; running into the same issue with a $40-a-pop series of dice is another. Even minute shape alterations can forever alter a die’s rolling balance. I hesitate to assume the worst for the retail dice’s build quality, but I’m hopeful the Pixels team tunes everything it can about packing so many materials (and their inherent heat) into their high-tech dice.

Recommended, with reservations

A Pixels die prototype... in the nude!
Enlarge / A Pixels die prototype… in the nude!
Jean Simonet

Having played around with prerelease Pixels for this long, I’m convinced enough to throw $40 of my own cash at a single die, before the current Kickstarter campaign concludes on Thursday, April 8. The prototype dice’s current amount of customization options, wonky as they are, have me comfortable with Pixels’ potential to eventually dazzle my table cohorts without blinding them.

I’ve been frank about my current Pixels issues, so I want to make clear: I’m grateful to the folks at Pixels for offering unrestricted hands-on time with an unfinished project, especially since I wouldn’t otherwise be able to test their sales pitch. I appreciate that kind of transparency in a prerelease Kickstarter, and that gives me faith that the final product will aspire to greatness, even if that comes at the cost of, say, delays down the line. (It’ll be a while before my in-person D&D group resumes a regular campaign, anyway.)

And we at Ars Technica also appreciate Pixels’ commitment to the open source movement. Simonet has already published Pixels’ API on Github and reminds buyers that every single kit technically counts as a “dev kit” (though joining the official “Dev Kit” tier on Kickstarter means getting a D20 ahead of other buyers for the sake of whatever Pixels-compatible apps you want to work on ASAP).

But Simonet and his team have their work cut out for them before I’d recommend Pixels’ full ecosystem of customizable, Bluetooth-synced bluster to an average user or a hobby shop’s impulse-buy aisle. We’ll keep an eye on the project and aspire to post a “final” review at some point, however long that takes.

    <em>Listing image by Sam Machkovech</em>
                                                  </div>

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.