Samsung is the primary companion to formally announce a brand new MR headset primarily based on the newly introduced Android XR. The machine, codenamed “Challenge Moohan,” is deliberate for client launch in 2025. We went fingers on with an early model.
Notice: Samsung and Google aren’t but sharing any key particulars for this headset like decision, weight, field-of-view, or value. Throughout my demo I additionally wasn’t allowed to seize images or movies, so we solely have an official picture in the meanwhile.
If I advised you that Challenge Moohan felt like a mashup between Quest and Imaginative and prescient Professional, you’d most likely get the concept it has a number of overlapping capabilities. However I’m not simply making a tough analogy. Simply trying on the headset, it’s clear that it has taken important design cues from Imaginative and prescient Professional. All the things from colours to button placement to calibration steps, make it unmistakably conscious of different merchandise available on the market.
After which on the software program facet, if I had advised you “please make an OS that mashes collectively Horizon OS and VisionOS,” and also you got here again to me with Android XR, I’d say you nailed the project.
It’s truly uncanny simply how a lot Challenge Moohan and Android XR really feel like a riff on the 2 different greatest headset platforms.
However this isn’t a publish to say somebody stole one thing from another person. Tech firms are at all times borrowing good concepts and good designs from one another—typically bettering them alongside the best way. So so long as Android XR and Challenge Moohan bought the good elements of others, and prevented the dangerous elements, that’s a win for builders and customers.
And lots of the good elements do certainly seem like there.
Fingers-on With Samsung Challenge Moohan Android XR Headset
Ranging from the Challenge Moohan {hardware}—it’s a handsome machine, little doubt. It undoubtedly has the ‘goggles’-style look of Imaginative and prescient Professional, however the place Imaginative and prescient Professional has a tender strap (that I discover fairly uncomfortable with no third-party improve), Samsung’s headset has a inflexible strap with tightening dial, and an general ergonomic design that’s fairly near Quest Professional. Meaning an open-peripheral design which is nice for utilizing the headset for AR. Additionally like Quest Professional, the headset has some magnetic snap-on blinders for those who desire a blocked-out peripheral for absolutely immersive experiences.
And although the goggles-look and even lots of the button placements (and shapes) are strikingly just like Imaginative and prescient Professional, Challenge Moohan doesn’t have an exterior show to indicate the person’s eyes. Imaginative and prescient Professional’s exterior ‘EyeSight’ show has been criticized by many, however I keep it’s a fascinating function, and one which I want Challenge Moohan had. Coming from Imaginative and prescient Professional, it’s simply sort of awkward to not have the ability to ‘see’ the particular person carrying the headset, though they will see you.
Samsung has been tight-lipped concerning the headset’s tech particulars, insisting that it’s nonetheless a prototype. Nevertheless, we’ve got discovered the headset is operating a Snapdragon XR2+ Gen 2 processor, a extra highly effective model of the chip in Quest 3 and Quest 3S.
In my hands-on I used to be capable of glean just a few particulars. For one, the headset is utilizing pancake lenses with automated IPD adjustment (due to built-in eye-tracking). The sphere-of-view feels smaller than Quest 3 or Imaginative and prescient Professional, however earlier than I say that definitively, I first have to attempt totally different brow pad choices (confirmed to be included) which might be able to transfer my eyes nearer to the lenses for a wider field-of-view.
From what I bought to attempt nonetheless, the field-of-view did really feel smaller—albeit, sufficient to nonetheless really feel immersive—and so did the candy spot on account of brightness fall-off towards the outer edges of the show. Once more, that is one thing which will enhance if the lenses had been nearer to my eyes, however the vibe I bought for now could be that, from a lens standpoint, Meta’s Quest 3 remains to be main, adopted by Imaginative and prescient Professional, with Challenge Moohan a bit behind.
Though Samsung has confirmed that Challenge Moohan may have its personal controllers, I didn’t get to see or attempt them but. I used to be advised they haven’t determined if the controllers will ship with the headset by default or be offered individually.
So it was all hand-tracking and eye-tracking enter in my time with the headset. Once more, this was a surprisingly comparable mashup of each Horizon OS and VisionOS. You should use raycast cursors like Horizon OS or you need to use eye+pinch inputs like VisionOS. The Samsung headset additionally consists of downward-facing cameras so pinches may be detected when your fingers are comfortably in your lap.
After I truly bought to place the headset on, the very first thing I observed was how sharp my fingers gave the impression to be. From reminiscence, the headset’s passthrough cameras seem to have a sharper picture than Quest 3 and fewer movement blur than Imaginative and prescient Professional (however I solely bought to check in glorious lighting situations). Contemplating although how my fingers appeared sharp however issues additional away appeared much less so, it nearly felt just like the passthrough cameras might need been targeted at roughly arms-length distance.
Proceed on Web page 2: Inside Android XR »
Inside Android XR
Anyway, onto Android XR. As stated, it’s instantly corresponding to a mashup of Horizon OS and VisionOS. You’ll see the identical sort of ‘residence display’ as Imaginative and prescient Professional, with app icons on a clear background. Look and pinch to pick one and also you get a floating panel (or just a few) containing the app. It’s even the identical gesture to open the house display (have a look at your palm and pinch).
The system home windows themselves look nearer to these of Horizon OS than VisionOS, with largely opaque backgrounds and the power to maneuver the window wherever by reaching for an invisible body that wraps across the whole panel.
Along with flat apps, Android XR can do absolutely immersive stuff too. I bought to see a VR model of Google Maps which felt similar to Google Earth VR, permitting me to select wherever on the globe to go to, together with the power to see areas like main cities modeled in 3D, Avenue View imagery, and, newly, volumetric captures of inside areas.
Whereas Avenue View is monoscopic 360 imagery, the volumetric captures are rendered in real-time and absolutely explorable. Google stated this was a gaussian splat answer, although I’m not clear on whether or not it was generated from present inside pictures that’s already obtainable on customary Google Maps, or if it required a model new scan. It wasn’t practically as sharp as you’d count on from a photogrammetry scan, however not dangerous both. Google stated the seize was operating on-device and never streamed, and that sharpness is anticipated to enhance over time.
Google Photographs has additionally been up to date for Android XR, together with the power to robotically convert any present 2D picture or video out of your library into 3D. Within the temporary time I had with it, the conversions regarded actually spectacular; comparable in high quality to the identical function on Imaginative and prescient Professional.
YouTube is one other app Google has up to date to take full benefit of Android XR. Along with watching common flatscreen content material on a big, curved show, you may as well watch the platform’s present library of 180, 360, and 3D content material. Not all of it’s tremendous prime quality, but it surely’s good that it’s not being forgotten—and can absolutely be added to as extra headsets are capable of view this type of media.
Google additionally confirmed me a YouTube video that was initially shot in 2D however robotically transformed to 3D to be seen on the headset. It regarded fairly good, seemingly comparable in high quality to the Google Photographs 3D conversion tech. It wasn’t made clear whether or not that is one thing that YouTube creators would want to opt-in to have generated, or one thing YouTube would simply do robotically. I’m positive there’s extra particulars to come back.
The Stand-out Benefit (for now)
Android XR and Challenge Moohan, each from a {hardware} and software program standpoint, really feel very very like a Google-fied model of what’s already available on the market. However what it clearly does higher than some other headset proper now could be conversational AI.
Google’s AI agent, Gemini (particularly the ‘Challenge Astra‘ variant) may be triggered proper from the house display. Not solely can it hear you, however it may see what you see in each the true world and the digital world—repeatedly. Its ongoing notion of what you’re saying and what you’re seeing makes it really feel smarter, higher built-in, and extra conversational than the AI brokers on up to date headsets.
Sure, Imaginative and prescient Professional has Siri, however Siri can solely hear you and is generally targeted on single-tasks fairly than an ongoing dialog.
And Quest has an experimental Meta AI agent that may hear you and see what you’re seeing—however solely the true world. It has no sense of what digital content material is in entrance of you, which creates a bizarre disconnect. Meta says it will change ultimately, however for now that’s the way it works. And with a view to ‘see’ issues, it’s a must to ask it a query about your setting after which stand nonetheless whereas it makes a ‘shutter’ sound, then begins eager about that picture.
Gemini, then again, will get one thing nearer to a low framerate video feed of what you’re seeing in each the true and digital worlds; which suggests no awkward pauses to be sure you’re trying straight on the factor you requested about as a single image is taken.
Gemini on Android XR additionally has a reminiscence about it, which provides it a lift with regards to contextual understanding. Google says it has a rolling 10-minute reminiscence and retains “key particulars of previous conversations,” which suggests you’ll be able to refer not solely to stuff you talked about lately, but additionally stuff you noticed.
I used to be proven what’s by now turning into a standard AI demo: you’re in a room full of stuff and you may ask questions on it. I attempted to journey the system up with just a few sly questions, and was impressed at its skill to keep away from the diversions.
I used Gemini on Android XR to ask it to translate signal written in Spanish into English. It rapidly gave me a fast translation. Then I requested it to translate one other close by signal into French—understanding full nicely that this signal was already in French. Gemini had no downside with this, and appropriately famous, “this signal is already in French, it says [xyz],” and it even stated the French phrases in a French accent.
I moved on to asking about another objects within the room, and after it had been a couple of minutes since asking concerning the indicators, I requested it “what did that signal say earlier?” It knew what I used to be speaking about and browse the French signal aloud. Then I stated “what concerning the one earlier than that?”….
A couple of years in the past this query—”what concerning the one earlier than that?”—would have been a wildly difficult query for any AI system (and it nonetheless is for a lot of). Answering it appropriately requires a number of ranges of context from our dialog as much as that time, and an understanding of how the factor I had simply requested about pertains to one other factor we had talked about beforehand.
But it surely knew precisely what I meant, and rapidly learn the Spanish signal again to me. Spectacular.
Gemini on Android XR can even do extra than simply reply basic questions. It stays to be seen how deep this will likely be at launch, however Google confirmed me just a few ways in which Gemini can truly management the headset.
For one, asking it to “take me to the Eiffel tower,” pulls up an immersive Google Maps view so I can see it in 3D. And since it may see digital content material in addition to actual, I can proceed having a reasonably pure dialog, with questions like “how tall is it?” or “when was it constructed?”
Gemini can even fetch particular YouTube movies that it thinks are the best reply to your question. So saying one thing like “present a video of the view from the bottom,” whereas trying on the digital Eiffel tower, will pop up a YouTube video to indicate what you requested for.
Ostensibly Gemini on Android XR must also have the ability to do the standard assistant stuff that the majority telephone AI can do (ie: ship textual content messages, compose an e-mail, set reminders), however it is going to be attention-grabbing to see how deep it can go together with XR-specific capabilities.
Gemini on Android XR seems like the most effective model of an AI agent on a headset but (together with what Meta has proper now on their Ray-Ban smartglasses) however Apple and Meta are undoubtedly working towards comparable capabilities. How lengthy Google can keep the lead right here stays to be seen.
Gemini on Challenge Moohan seems like a pleasant value-add when utilizing the headset for spatial productiveness functions, however its true future most likely lies on smaller, on a regular basis wearable smartglasses, which I additionally bought to attempt… however extra on that in one other article.
