Hi, if you want to scan interiors you can use RealityCapture, as it is not limited to 200 images. The images are aligned after capturing, it is not related to loops.
I tried many times on android app all the photos i take turn red and dont get processed , despite i follow the guidelines and take photos that are close together.
In most cases the images are better that video (better resolution, focus, stability, no blur) and basically you should receive a better model from still images than from video. Therefore the images are used in the process.
Hi, there is no such functionality to do that for now. You can try to string it for example. Also, cup is not an ideal object for a photogrammetry scanning (it depends what kind of cup, but if it is shiny and without features, then not).
@@여름숲-k5l So, you are using RealityScan and not RealityCapture (as mentioned in the forum)? Do you have internet connection in the app? Is this happening for each project?
I'm planning to buy a phone, and I want to use this software... for my mom's cooking. So iPhone 12 Pro or Android (OnePlus Nord 3/Zenfone 9) would give better results? NEED HELP CHOOSING
I have a really great idea: Train a generative AI on large amounts of 2d vs 3d scene data. That way you can feed the AI all the 2d photos and videos on the internet and get a digital 3d copy of the planet (Much more detailed than Google Earth). You could even train an AI to convert that into Unreal Engine 5 and have a game/simulator to explore the world. Or GTA 7? The challenge would be to get enough data to train the AI. You could either crowdsource 3d scans of lots of places or you could make synthetic data by using already existing game worlds made in Unreal Engine. Once the AI is trained on enough data, it will make traditional photogrammetry obsolete. Imagine if you gave a 2d image to a human 3d artist. He would be able to use his experience from seeing lots of places in real life and how those relate to 2d photos, to draw a scene in 3d from the image you gave him. This is essentially how a generative AI would do it, just much faster and better (given enough training).
Yeah it doesn't work like that. You can't just take pictures in this app. It does this annoying "capturing" thing and it decides when to take pictures for you. Except the parameters for making it "decide" to take a photo is impossible to decipher. 👎
Really cool app, can't wait to use on Android.
I downloaded and experienced it on my friend's iPhone. It's a really exciting app, I'm looking forward to it coming to Android.
Thanks Epic games for your advanced free technologies ❤
I get it all set up and press the take a photo button and nothing happens?
When Android? ㅠㅠ
today
@@ahahlex thank you bro
When scanning large models, for example interiors, how does the app handle tracking drift? Does it re-align the photos when the loop is closed, or?
Hi, if you want to scan interiors you can use RealityCapture, as it is not limited to 200 images. The images are aligned after capturing, it is not related to loops.
@@CapturingReality is that a seperate app? Can you also do it on android?
@@FireJojoBoy RealityCapture is a separate app, but only for computers. RealityScan can be used on Android and also on iOS.
Is this any better than one of those small lidar scanners ?
I tried many times on android app all the photos i take turn red and dont get processed , despite i follow the guidelines and take photos that are close together.
How small of an object can you scan?
Is there an advantage of doing this with still shots as opose to video camera?
In most cases the images are better that video (better resolution, focus, stability, no blur) and basically you should receive a better model from still images than from video. Therefore the images are used in the process.
i dont belive its free .thank so much .its really useful app.
Hi, i just wanted to integrate this into my company, but wierdly all my photos taken are pitch black.. Anyone with that issue? Any fix or workaround?
How much does the model render cost?
It is free (just your Wi-Fi data are used)
@@CapturingRealityCool. Can I close up shots for better detail of objects or should I only take full body photos?
If I'm trying to say, capture a cup, is there any way to get the bottom of it or am i missing something?
Hi, there is no such functionality to do that for now. You can try to string it for example.
Also, cup is not an ideal object for a photogrammetry scanning (it depends what kind of cup, but if it is shiny and without features, then not).
Every time when I try to capture and try make a model, it says "cameras align failed". what should I do?
How are the images captured? Is there enough overlap between images?
@@CapturingReality I captured 30 pictures and the distance from the object was about 20cm. I think overlap was enough given the distance.
@@CapturingReality I asked for help on the forum and I followed the tips that other user gave to me but it's still same.
@@여름숲-k5l So, you are using RealityScan and not RealityCapture (as mentioned in the forum)?
Do you have internet connection in the app? Is this happening for each project?
@@CapturingReality I used realityscan and I asked forum which is connected to epic games. And also it was connected to internet.
I'm planning to buy a phone, and I want to use this software... for my mom's cooking. So iPhone 12 Pro or Android (OnePlus Nord 3/Zenfone 9) would give better results? NEED HELP CHOOSING
Maybe it more optimize on iphone for now. The 12 pro also have lidar sensor
I cant connect to the server, any ideas?
When will this be available on Android?
It's available now
The lasty thing in the world i would use is Apple, Android mobile devices are the best.
My photos keep getting red no matter how many photos I take
Android, please!
It wont let me take any pictures manually or auto scan
Have you tried the latest RealityScan version? Please let us know if the issue persists at realityscan-feedback@epicgames.com
Gr8 Tips
I love you ❤️
Cool but i cant sign in the app gg
I have a really great idea: Train a generative AI on large amounts of 2d vs 3d scene data. That way you can feed the AI all the 2d photos and videos on the internet and get a digital 3d copy of the planet (Much more detailed than Google Earth). You could even train an AI to convert that into Unreal Engine 5 and have a game/simulator to explore the world. Or GTA 7?
The challenge would be to get enough data to train the AI. You could either crowdsource 3d scans of lots of places or you could make synthetic data by using already existing game worlds made in Unreal Engine. Once the AI is trained on enough data, it will make traditional photogrammetry obsolete.
Imagine if you gave a 2d image to a human 3d artist. He would be able to use his experience from seeing lots of places in real life and how those relate to 2d photos, to draw a scene in 3d from the image you gave him. This is essentially how a generative AI would do it, just much faster and better (given enough training).
Yeah it doesn't work like that. You can't just take pictures in this app. It does this annoying "capturing" thing and it decides when to take pictures for you. Except the parameters for making it "decide" to take a photo is impossible to decipher. 👎
Hi, it is possible to disable the auto capture in the settings.
waaaaw