• Lowpass
  • Posts
  • WiseAR develops camera-free eye tracking for AR

WiseAR develops camera-free eye tracking for AR

Ears and eyes and mouth and ...

Welcome to Lowpass! This week: WiseAR does eye tracking without cameras, and Vizio is in talks with Walmart. Also: Mockumentaries!

Eye tracking for AR could come sooner than you think

I recently began testing Meta’s new Ray-Bans, which now offer access to an AI assistant that uses the device’s camera to look at the world for you. I’ll have more to say about this in the coming weeks, but one of the first things I realized is that talking to an assistant in public can be pretty awkward.

I know, that’s not really news to anyone who has ever summoned Siri or Google Assistant with their earbuds. However, if you see Meta’s Ray-Bans as not just a pair of headphones that happens to live in your glasses, but a precursor to smart glasses with AR displays, then it quickly becomes obvious why voice isn’t going to cut it as an input modality. Can you imagine telling your AI assistant to open up an app, navigate through a complex menu and select a certain item, all while the other people on the bus are staring at you? Yeah, me neither.

That’s why Meta has been working on neural interfaces, as Mark Zuckerberg reminded everyone again this week when he took to Instagram to pan Apple’s Vision Pro headset. Meta’s work focuses primarily on an EMG wristband, capable of detecting mouse click-like interactions and more by monitoring electrical signals traveling to your motor nerves.

There’s a lot to like about that approach – in the future, these kinds of EMG sensors could conceivably be combined with a smart watch or fitness tracker – but it’s not the only possible solution. AT CES in Las Vegas earlier this year, I got a demo from neural interface startup WiseAR that suggests sensors placed in earbuds, or even smart glasses themselves, may be just as useful to control AR interfaces.

Plus, sticking to your face as opposed to your wrist allows WiseAR to do something that’s pretty cool: Eye tracking without a single camera.


Neural interfaces that won’t kill you: Half a decade ago, WiseAR’s co-founders Yacine Achiakh and Alain Sirois were both working in advertising in the Bay Area. After putting in four years working on AI-powered ad tech, they felt like it was time for something else. “We realized that our soul was progressively leaving our body,” joked Achiakh (seen in the photo above wearing one of his company’s prototypes if you’re reading this on the web) during our conversation.

  • At the time, they were having lots of conversations with engineers for Elon Musk’s Neuralink startup. They were fascinated by the technology, but didn’t want to wait decades – and potentially risk people’s lives – to bring neural interfaces to market.

  • That’s why in 2019, they founded WiseAR to explore less invasive approaches. “The goal of our company is to try to bring neural interfaces to the world without having to drill a brain implant in your head,” Achiakh said.

  • WiseAR wasn’t the only company  with that goal, but the duo felt that a lot of their competition were over-promising and under-delivering by painting brain signals as a vague way to improve your overall health and mental wellbeing.

  • “We really wanted to be true to the science,” Achiakh said.

The company also wanted to build hardware for the masses. Wearing an EEG headband to control the force can be a great party trick, but may also get you some weird looks on the bus. That’s why WiseAR settled on earbuds as their initial hardware of choice, and EMG as the signal to monitor for their first prototypes.

  • “We started by building features based on muscular activity, and we looked into what we could provide as value for the user,” Achiakh recalled.

  • Hands- and voice-free control for media playback seemed like an obvious way to improve earbuds. “I live in New York,”  Achiakh said. “In the winter, when I'm on my bike and I have my gloves on, (changing) the music or skipping the ad (is) a bit of a pain.”

  • That’s why WiseAR built a prototype that monitors electrical signals sent to your jaw muscle. Clench your jaw, and you can pause the music, skip a song and more.

  • It’s worth noting that this prototype didn’t really work for me when I tried it in Vegas, perhaps because I was wearing a tight-fitting mask. But I could observe Achiakh control music playback effortlessly with barely visible jaw clenches.

Next up: eye tracking, without a single camera. But WiseAR isn’t stopping at EMG. The company also built a prototype that uses electrooculography, or EOG, which measures the electrical potential between the cornea and retina of your eye.

  • “When you're moving (your eyes to)  the right or the left, it basically creates a perturbation of the electromagnetic field of your head,” Achiakh explained.

  • To demonstrate the technology, WiseAR let me play a simple trivia game, with me answering questions simply by glancing to the left or the right. 

  • One of my sides appeared to be more responsive than the other, but it basically worked, with the headphones picking up my eye movements without any camera tracking my eyes – a pretty surreal experience.

WiseAR’s work has primarily focused on earbuds, as they’re already widely adapted by consumers. However, as smart glasses are gaining traction, and companies get ready to unveil their first AR glasses in the coming years, the startup also wants to make its sensors work on those devices.

  • Initially, this could be as an add-on, perhaps a special pair of earbuds that works in concert with your glasses. However, Achiakh told me that WiseAR is also looking to directly integrate its sensors into the temples of smart and AR glasses.

  • And the company doesn’t stop there. As part of its neural interface research, it is also looking to tap into signals from your auditory cortex. This could eventually allow it to capture spoken words, even if you’re just mouthing them without making a sound.

  • Will that pass the not-looking-weird-in-public test? Honestly, I don’t know. But there are a range of possible applications for this, including in the accessibility field, that are pretty fascinating.

New input modalities require new interfaces. WiseAR’s eye and jaw tracking could still make for some pretty powerful inputs for future AR glasses. Sure, this is pretty low-resolution eye tracking, and nowhere near the precision you’d get from cameras observing your eyes. Then again, higher resolution eye tracking can be pretty hard to get right as well, which is why it can be so hard for some people to select small UI elements in the Vision Pro interface.

If there’s one lesson that AR device makers can learn from the smart TV space, it’s that the combination of simple inputs and easily navigable interfaces is often the best solution. Companies have tried all kinds of things over the years to reinvent the TV remote – magic wands, gesture-tracking via cameras, voice commands – but day-to-day, there’s still nothing that beats the simplicity of a D-PAD remote and a well-structured grid design.

Future AR interfaces may need to be equally pared-down, so all it takes to start an app on your glasses is a simple 👀

This post is available to paying subscribers only. Thanks for your support!
Did this email get forwarded to you? Then sign up for a 7-day free trial here.

The logic behind a possible Walmart-Vizio deal

Earlier this week, the Wall Street Journal reported that Walmart is in talks to acquire smart TV maker Vizio for more than $2 billion. For what it’s worth, I heard some rumblings about this a few days ago as well, so there seems to be something to it … but of course, there’s no guarantee that this deal will actually happen.

Still, there’s some logic behind this, at least from Vizio’s side. The smart TV maker has been looking to execute a playbook similar to Roku’s with low-margin hardware being a loss leader to grow a high-margin ads and services business. But that only works if you actually sell TVs, which Vizio has struggled with in recent years.

  • Ten years ago, Vizio sold 7.1 million TVs. The company managed to hit the same number in 2020, when everyone was spending their stimulus checks, but sales fell to 5.5 million and 5.2 million devices in 2021 and 2022, respectively. (Vizio will report its FY 2023 results later this month.)

  • Revenue numbers are even worse: In 2014, the company generated a total of $3.1 billion. 2022 revenue came in at just $1.9 billion.

  • Ad revenue has been growing steadily, to the tune of $478 million in 2022, but that growth can ultimately only be sustained if the company adds new households.

  • One way Vizio wants to do that is by licensing its TV OS, but that’s a very crowded market.

I did a lengthy interview with Vizio CEO William Wang for The Information late last year, which is worth rereading in light of this week’s news. Two things are especially noteworthy:

  • Wang told me he’s still still “super pumped” to go to work everyday, but some of his answers could also be understood as a signal that he’s ready to move on. “I’m looking forward to getting a next-generation team built to take over that responsibility,” he said.  “That’s something I’m really looking forward to. More so [than] me running day to day.”

  • Perhaps even more striking: Wang told me that he’s thinking about breaking with Vizio’s outsourcing strategy, and vertically integrate some of the company’s manufacturing. “I don’t see vertical deintegration work forever,” he said. “Good entrepreneurs have to adapt.”

  • The latter would be a pretty big departure for Vizio, which to date has relied on contract manufacturers for all of its hardware. It’s also a response to the rise of Chinese manufacturers like Hisense and TCL, which have been able to outdo Vizio with low-cost TVs, thanks to vertical integration.

  • The issue with vertical integration is that it requires capital — a lot more capital that Vizio currently has, even after the Wall Street Journal story pushed the company’s market cap up to $1.9 billion.

The big question: Does the deal make sense for Walmart? We had a pretty spirited discussion about this in the Lowpass Slack this week, with folks reminding me that Walmart is already one of Vizio’s biggest retailers, so the company knows what’s selling. Having their own TV brand could also help Walmart compete more directly with Amazon and Google, and make more money with the TVs it sells

However, Walmart isn’t exactly known for doing services well. Case in point: The company never really knew what to do with Vudu, and eventually sold the platform off to Fandago. Whether acquiring Vizio would help Walmart to get services right the second time around is anything but certain.

Want to get your company in front of an audience of over 18,500 tech and media insiders and decision makers? Then check out these sponsorship opportunities.

What else

The quest to build more accessible music tech. For Fast Company, I wrote about the challenges facing blind and vision impaired musicians and music producers, and efforts to overcome them with new accessibility standards.

Windows Mixed Reality is officially dead. The newest Windows 11 build doesn’t support WMR anymore, and Microsoft is telling anyone who wants to keep using their headset not to upgrade.

The WWE is going short-form on X. Musk’s efforts to build an everything app continues with … five-minute wrestling matches?

Location-based VR is having a moment (again). Sandbox VR CEO Steven Zhao is bragging about growing ticket sales, while The Void’s co-founder Curtis Hickman is once again teasing a return.

Prime Video is downgrading its video quality for most. When Amazon introduced ads to the basic prime video tier, it also quietly took away Dolby Atmos and Dolby Vision.

SiriusXM lays off 160. The latest layoffs follow 475 job reductions last year.

Apple’s tvOS 17.4 beta references a Homepod smart display. Looks like the next Homepod will have a screen.

Mozilla Hubs is shutting down. The metaverse-ish platform is being discontinued as Mozilla is laying off 60 people.

First data on Vision Pro app installs. Good work here from Immersive Wire’s Tom Ffiske, who gathered a bunch of early Vision Pro app install data. The bottom line, in Tom’s own words: “Success is measured in the thousands.”

That’s it

Our family has been catching up on a few classic mockumentaries. So far, we’ve watched “Waiting for Guffman” and “Best in Show.” What should we watch next? Let me know by responding in the comments, or directly replying to this email.

Thanks for reading, have a great weekend everyone!

Photo courtesy of WiseAR.

Subscribe to keep reading

This content is free, but you must be subscribed to Lowpass to continue reading.

Already a subscriber?Sign In.Not now

Join the conversation

or to participate.