Still in that uncanny valley for me. Zuckerberg almost looks human in that render, it's freaking me out...
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Starts Improving Vision Pro Personas in visionOS 1.1 Beta
- Thread starter MacRumors
- Start date
- Sort by reaction score
Still in that uncanny valley for me. Zuckerberg almost looks human in that render, it's freaking me out...
Yeah, I saw that too:
"In this interview, it's unclear whether the avatars were being decoded and rendered onboard, or whether a PC was doing this work. A USB-C cable is connected to both headsets."
"Regardless of which specific Quest headsets the shipping Codec Avatars feature is intended for, the quality almost certainly won't be the same as Meta's most impressive demos, such as the Lex Fridman interview with Mark Zuckerberg. Those demos, and that interview, used Codec Avatars that had been generated using a specialized capture rig with over 100 cameras."
$3500 still seem expensive? 😆
So in less than a week they got to feed their algorithms the personas of Vision Pro users and get useful data out of it?
Meta's Codec Avatars are incredible but:
- Don't ship on a consumer product
- Run on high end PC's (based on Mark/Lex's commentary from the same interview where the screenshot is from)
- Require a high end setup to capture and hours of post-processing (Link)
A more apt comparison is this:
But again:
- doesn't ship on a consumer product
- Takes 15 minutes to process
Last edited:
I've gotta be the only person in the universe who thinks Apple Vision personas are kind of cool, like a futuristic dream.
Im pretty sure the teams developing these different features are not the same.
It's used in other video calling apps too.
Zoom, Webex etc if you use those at work. Most people don't turn on their cameras anyway but for companies with stricter policies I imagine this technically could be useful.
I think the way it works is visionOS inputs the Persona wherever a "front camera" would be used. E.g. filming yourself on Snapchat.
...that are still cheaper combined than tne AVP. And I thought Apple Silicon was so advanced, with its many unified-memory GPU cores and neural engine?
Hello, fellow alive human! I do like it here on the Earth planet too. I will now show you my teeth in order to convey my trustworthiness.
Last edited:
You think that these things are sitting in boxes?
How would there not be usable data?It constantly amazes me how made-up stuff by other companies that isn’t available (and may never ship) beats what Apple is shipping right now. 😆
I am sure the avatar could be much more sophisticated if the Vision Pro took more images from more perspectives. Prioritizing convenience often is the wrong way. Many users would not mind a 15 minute setup process, if the resulting avatar looked more realistic. That 40 second or so setup obviously is not good enough.
This is the weirdest feature/hack/workaround I think I’ve seen in a long time, or ever, from Apple. It’s offputting enough that I would just avoid FaceTiming someone using this. Why do I want to look at an animation? The whole point of FaceTime is you’re seeing the person you are talking to. That’s a no for me dawg, I’d tell my caller to take the goggles off and talk to me on their iPhone, iPad or Mac. Even a conversation over my Ring doorbell would be visually superior to this.
You think they're not working on that? The team creating personas isn't the same working on the hardware to get a wider FOV.
And now we run into the opposite problem: It's too HD. That detail is really distracting, that fake greasiness is trying to look realistic but in an awful way, and there's that weird uncanny dead look in those eyes.
I hate to say it, but that ugly cartoon version of Zuck has more the right idea. A good artistic style ages more gracefully than trying to make the most realistic model possible. The problem is they need to find an artistic style that isn't absolutely atrocious, blinding, or corporate and soulless
If only Apple knew this, they’d be soooo much better. Give them time to bring out version 1.2.
If you look at the 1st iteration and now version 1.1, the difference is quite large. It’ll get better and it’s actually only in Beta.
I get the idea behind shipping the product before ironing out a lot of these features. You get lots of attention from the initial release and then drip feed updates to keep content creators pushing out "check out this latest update" to stay being the main product people are talking about. Seems to work well.
Just read The NY Times review on the AVP. In a nutshell, the first generation headset lacks polish and purpose.